Hi there, everyone. I'm Deb Wasser. Good morning, and welcome to Etsy. I'm the Head of Investor Relations, and with me is Gabriel Ratcliff, our Senior Manager of Investor Relations. We are so excited to have you all here in Brooklyn today for our first ever Investor Day. We're also happy to welcome all of you listening on our webcast. For those of you in the room, please make sure to mute your phones. If you need to exit during the presentation, please use the door in the rear. We appreciate that very much. Thank you. Sorry. Before we begin, let me just remind everyone that today's presentation contains forward-looking information and non-GAAP financial measures about our outlook, our market opportunity, and anticipated financial performance.
We encourage you to read our cautionary statement regarding forward-looking statements, as well as our reconciliation to non-GAAP measures, which are available in our slides today, as well as on our webcast website and in our SEC filings.
Our agenda today will include presentations from our senior leadership, followed by Q&A. Please hold all your questions until the end of the meeting. Speakers will include our CEO, Josh Silverman, our Senior VP of Product, Kruti Patel Goyal, our CTO, Mike Fisher, Raina Moskowitz, our Senior VP of People, Strategy and Services, and last but not least, Rachel Glaser, our CFO. We'll have a 15-minute break at about 10:30 A.M., and we'll have lunch and a seller showcase. We hope that you all can join us for that.
Good morning, everyone. Thank you for making it out this morning to spend some time with us. I really appreciate it. We get about 15 minutes once a quarter to tell our story, so it's really nice to get a little bit more time to give you a fulsome view of what we're all about. I really appreciate it. I want to start by recognizing a couple of people who are in the room. We have Jill Simeone, our General Counsel, and very key member of our executive team here with us. We also have two members of our board of directors. We have Melissa Reiff and Jonathan Klein. Melissa is the CEO of The Container Store Group, and joined our board in 2015. Jonathan Klein is co-founder and chairman of Getty Images, and joined our board in 2011.
Thank you both for being here. There's three things that I hope you will leave feeling and believing. One, that Etsy has a very large addressable market opportunity. Two, that we are in the early stages of unpacking that opportunity, and three, that we've got unique and defensible rights to win. Let's dive in. I'm going to start with the very basics of the Etsy business model. Etsy is a two-sided marketplace. We've got about 2 million sellers selling about 60 million items to about 39 million active buyers. Buyers come because of all of the sellers and all of the unique items that they have for sale. Sellers come because of all the buyers, and that's a really critical point. We are not a tools business. The number one thing we give sellers is buyers.
We're not just a place to put up a shop on the website, we're a place to actually come and sell goods. We're also not a traditional e-tailer. We don't live or die by picking the right few items that are right for this season. Retail, real estate, and managing logistics are not core parts of our business. In fact, they're not part of our business at all. Rachel's going to talk a little bit more, later about, in more detail about the beauty of the two-sided marketplace business model, but it is indeed an amazing business model. If it's such an amazing business model, why aren't there more of them? Well, in fact, it's lightning in a bottle to be able to get a two-sided marketplace to work at scale.
I've spent a little bit of time in Silicon Valley with some venture firms, you get pitched five times a day on a two-sided business place pitch. They'll all come and tell you, "It's going to be amazing once it's at scale," and they're always right. The challenge is, roughly 0% of the time can you get a two-sided marketplace to scale. In any given year, there are literally thousands of startups that are funded with the idea that they're going to become two-sided marketplaces, and maybe one per year does. If you think about it, you probably can't count on more than two hands the number of really large, at-scale, two-sided marketplaces that there are. Why? Because the day you open, sellers come, but there's no buyers, so they go away. The day you open, buyers come, but there's nothing to buy, so they go away.
While that sounds like a really basic problem, it's actually incredibly hard to solve that chicken and egg. Very few companies do. Once you have, it's a really magical thing. A couple of things in particular I love about the two-sided marketplace business model. One, it gets better as it gets bigger. Very few businesses get better as they get bigger. Two-sided marketplaces do. Second, there's very deep moats. We're going to spend a lot of time today talking about both of those. Before we do, let's pause for a sec and talk about our particular marketplace, and in particular, what we have for sale in our particular marketplace. What do you think about when you think of Etsy? Maybe you think about a cat vest or an apple cozy. Maybe you think about fingerless gloves. How about a storage case for your iPhone?
How about a hat that looks like my cat? Do any of these sound familiar? Maybe some of us have bought them. It is, in fact, true. All of these wonderful items are for sale on Etsy. You can buy any of these things on Etsy. But they together make up a tiny fraction of what we actually sell. Here are some other things you actually can buy that make up the bulk of our sales on Etsy. The number 1 category on Etsy is home furnishings. We have beautiful decorative elements for any size home, apartment, or office. They fit any style and any budget. In fact, we have 400 different types of home accessories for sale at this moment on Etsy. It's not just the accessories. We sell beautiful sofas, desks, bookshelf, lighting.
Maybe you live in New York City and you have an unusually small space you're trying to furnish. Maybe you have a weird sized dining room and you need a dining table that's made just for you, or lighting fixtures that really speak to how your sense of style you want to present. We sell over 1,000 pieces of furniture each day. What you know when you buy from Etsy is no one's going to walk into your apartment and say, "Oh, yeah, I just saw that on Main Street." Oh, yeah, I just saw that on some other e-commerce site. You know it's going to be unique, and you know it's going to come with a story. Our second biggest category is clothing. Not just any clothing. We sell everything from everyday to vintage couture.
What it all has in common is it will uniquely express your sense of style. You'll show up at the party knowing that no one else is wearing the same dress. Maybe that's why we sell 8 pieces of clothing and 1 pair of shoes every minute of every day. In fact, three-quarters of these items come from non-U.S. sellers. Etsy is a trendsetter. You might have noticed Etsy being profiled or with clothing being worn by Beyoncé and Jessica Simpson and many other trendsetters out there. Our third largest category is jewelry. We sell everything from costume jewelry to fine jewelry. Speaking of fine jewelry, what more important occasion do you have in your life than your wedding? Which is maybe why we sell so many engagement rings and wedding bands. In fact, we sell 1 engagement ring or wedding band every 90 seconds on Etsy.
Speaking of weddings, we don't just sell the wedding bands and the engagement rings. We sell dresses. We sell decorations. We sell everything you need for the wedding. In fact, last year, we sold over 1 million bridesmaid and groomsmen gifts on Etsy. For many people, what's the next major life event after weddings? Maybe babies. We sell everything for babies and children. We sell toys, we sell baby furniture, we sell baby clothing. When you show up at the baby shower, you're not going to have the same gift as everyone else. When you walk into your child's room, you're knowing that it's going to express their sense of style, their interests. Of course, there are so many other life events for which Etsy can provide the perfect addition. Birthdays, anniversaries, holidays.
We haven't even talked about so many others of our smaller categories, things like bath and beauty, that are fantastic and growing really fast on Etsy right now, where you can find things for every sensibility and every skin tone. In fact, if you own our stock or you're thinking about owning our stock, I challenge you, go back and look at the last 90 days of purchases you've made online. Look at everything you've bought online. Then spend a little bit of time and see how many of those things you could have bought on Etsy. I promise you'll be very pleased with what a high percentage of things you've bought online you could have bought on Etsy. Those items are going to be more unique. They're going to be more tailored to your taste, and they're going to be sold at a fair price.
Just looking at some of my recent purchases over the past few weeks. The desk in the upper right, that's the desk in my office. For me, it has some meaning. In a world where automation is changing the nature of work, that desk comes from a General Electric factory in DeKalb, Illinois. There's an Etsy seller that cuts down that factory and turns it into furniture. I think it's not only very beautiful, but I think it has some meaning that motivates me every day. We spend a lot of time in the kitchen in my family, and we could have bought any old rolling pin, but that particular rolling pin made Valentine's Day into a more fun family baking project. We've got a new six-month-old puppy. She's adorable. Her name is Bonnie.
Not only is she super cute, but I think she looks even better in that custom knit sweater made just for her. We sell lots of beautiful things for lots of different occasions. What does it all add up to? How big is our TAM? This is a conversation I've had with probably almost every single person in this room. Let me start with there is no market for handmade because handmade is not a market. Nobody wakes up in the morning and says, "I want to buy something handmade. What do you got?" People wake up in the morning with a very specific, "I want to buy furniture. I want to buy clothing. I want to buy jewelry." We compete in each of these markets. How big are these markets? Well, Wayfair sells only home furnishings, and they say that their market is over $400 billion. ASOS sells only clothing and only in some countries. They say that their market is about $300 billion.
Farfetch sells only luxury goods, and they say that their market is also about $300 billion. Blue Nile sells only jewelry and gems. They say that their market is over $400 billion. We participate in every single one of these markets, and we have something really unique and different and special to offer. How big is our TAM? Rachel will talk a little bit more about our most recent definition of our TAM. The truth is, I think it's a little bit hard to define. What I can say is we are nowhere near the point where our TAM is a limiter to our growth. We've got a big market opportunity. What's our right to win? Let's spend a few minutes on that.
We as a team have spent a lot of time on that. I think when you're defining strategy, the most important starting place is to look outside your walls. What are the trends that are happening in the industry, and how do those affect us? The old saw of retail is you've got selection, price, convenience, pick two. You can only ever be good at two at a time. The internet has changed that, and it has allowed for the arrival of massive scaled e-tailers who do all three exceptionally well. They have a nearly unlimited selection of things for sale, and they buy them 10,000 units at a time, so they get price discounts that they can pass along to the consumer. They store them in warehouses ready to ship in an hour or a day.
As a result, they're gaining scale and starting to consume all of our consumption. Almost everything we buy is being bought from one or two massive logistics providers that get things to us cheap and fast. As they do that, they commoditize everything that there is for sale. Even the mighty Nike, who spends billions of dollars promoting their brand and hiring some of the most legendary athletes of our time, has recently capitulated and allowed their shoes to be sold in some of these mass e-tailers. Now you're just another size 11. I don't envy anyone who's selling products that carry a barcode and can easily be compared with other products, because if it can be commoditized, it will. As you're on these sites every day, buying all of those commodities of life, you need diapers, you need socks, you need batteries.
What are the things that those mass e-tailers, those mass logistics providers, can't do? If I'm there every day buying all of the commodities of life, it can't be special. When the world has a default, you yearn for the antidote to that default, when you want something to feel special, when you want it to feel meaningful, when you want it to be fun, when you want it to express your sense of identity, show that you have care for someone else, or when you want it to feel human or personal. Test yourself. I wanted to get you something really special, so I went to Walmart.com. I had a half an hour, and I just wanted to have fun, browse around, get inspired, so I went to Amazon. There's a very special role for a site like Etsy in a world that's becoming more and more commoditized.
What are the particular pillars that will underpin our competitive advantage, our right to win, in this new future? We believe that there are four. Each of them is powerful individually, but collectively, we believe that they're unbeatable. The foundation is our collection of unique items, and I've spent a bit of time talking about that already, so I won't belabor the point. What good is that unique collection of items if it's overwhelming, if you can't find what you want? Best-in-class search and discovery becomes essential. We don't need to be better than Google at search. We don't need to be better even than Amazon at product search. We just need to be the best in the world at search and discovery for the ultra long tail of items, for things that don't map to a catalog, to things that are each unique and ephemeral.
When they're sold, they're sold. That's where we need to be and are investing to be the best in the world. Let's take an example, a simple search for coasters. A search for coasters today returns over 200,000 results on Etsy. That's amazing. It's also an interesting technical challenge, and we've made great progress in search, and that's been an important part of the financial results that we've delivered over the past couple of years. You will now notice that all of the items on this page are actually coasters. That's progress. By the way, it's harder than it looks. Mike is going to spend a little bit of time talking about that. You will also notice that it is a relatively random assortment of what you get from 200,000 coasters, right? It almost feels like you could be throwing darts.
There's not a point of view about quality, and it's not a curated set of coasters that particularly speak to my aesthetic, to my sense of style. We are investing to build a search and discovery experience that brings up the best of Etsy, those things that really are high quality and high quality according to me that speak to my aesthetic, that are things that I'm likely to like. In this world, we need to solve for the tyranny of choice. We need to do a better job of personalizing, and we need to meet buyers where they are. Sometimes they're very high in their purchase funnel. They're just looking for something interesting to decorate their house. Sometimes they're very deep. They know they're looking for a coaster, and they want a bamboo coaster.
We need to be able to understand that and tailor our shopping experience to meet the buyers where they are. We're investing to build really best-in-world search and discovery, underpinned by having it be fun and inspiring, having a strong point of view on quality, and a much more personalized buying experience. Having that collection of items and finding what you want still doesn't do justice to Etsy. Etsy is a uniquely human shopping experience. We have sellers who literally raise the sheep, shear the wool, darn the wool, and knit it into a sweater. You take that same sweater and you put it up on Amazon, and it's just another yellow medium sweater. You need the opportunity to tell the story. What is it about this particular item, how it was made, why it was made? What's the seller's story? That's when things really come to life.
That's when the buyer fully appreciates what they're getting, and the seller is fully appreciated. We've made strides in making Etsy a more human experience. Kruti is going to tell you more about the opportunity that we have to really bring that to life, because it is indeed the personal touch on Etsy that makes such a difference. Each one of you, I'm sure, has gone and talked to some buyers. Hopefully, you're all buyers yourself. I'd be surprised if you've had very many conversations where the buyer didn't tell you about the personal handwritten note they got from the seller, and what a big difference that meant to them. We hear it over and over again. Etsy is building the superhuman buying experience.
How do we take the experience you get when you go to that artisan's workshop and see them make their craft, when you go to that vintage seller's shop and meet directly with the owner and discuss with her the particular items you're considering? How do we take that offline experience and bring that experience online? That's where we're investing. That's where we're going. That's a huge opportunity for us. Putting all of those together, the last fundamental pillar that's critical for us is trust. If you're buying a Gillette Mach3 razor, you can buy it from any of 100 sellers, and you really don't care. You know it's a Gillette Mach3 razor.
When you're buying a dress made bespoke from a maker, when you're buying a dining room table designed just for your room or a piece of jewelry that's important to you, buying it from someone you've never met is a trust hurdle. For each of these makers individually to invest to build that level of trust is both incredibly time-consuming and incredibly expensive. They can put up their own web store on the internet and be yet another tiny voice in a sea of shouting and a sea of sameness. It's Etsy's job to fill the trust gap to connect the buyers to the sellers. That means creating a set of consistent policies so that the brand of Etsy is the foundation that the buyer trusts, and we lend that trust to each of our sellers so that they can all benefit from the collective whole.
In order to do that, there's several things we need to do. We need to build a strong brand promise that we consistently deliver on, enhanced trust signals, and best-in-class member support. We've been investing and making significant progress on this over the past couple of years. We have lots of opportunity to continue to do better. Kruti's going to talk more about some of the investments we're making in this area. To conclude this section, when I roll the tape forward in my mind, I don't see a world 10 years from now where there are 2 million places to buy things online. I just don't.
While that may sound obvious, folks who are investing in e-commerce builders to allow everyone to put up their own shop and infrastructure players to make all of the payments and everything else seamless, the fundamental thesis must be that 10 and 20 years from now, there are going to be literally millions of places to buy things online. I think there's going to be a very small number of places to buy things online. I think there's going to be one or two massive e-tailers with incredible logistics facilities that sell all of the commodities of life. When the world has a default, it yearns for the antidote. I think there will be a very small handful of players who have earned a right to exist and do something different in that world.
I think Etsy is incredibly well-positioned to be one of those few people thriving in that environment. That's the vision we're building. Don't just hear it from me. I think it's really important. In our culture, we always bring the sellers' voices into the room. We thought it'd be nice at this moment to show a little video where our sellers can tell their stories and bring their voices into the room.
The online market today, you don't really know where your stuff's coming from anymore. For me, it was really important to keep it handmade and sourced in-house. We make every product to order. Being in a community where people value a handmade product and know the blood, sweat, and tears that goes into everything that's made, they know where that product's coming from. The money that they spend on Etsy is actually going back to a person, to a family.
Opening up a shop on Etsy completely changed my life. One of the things that I appreciate the most about Etsy is that it lets me do what I love and what I want to do from anywhere in the world. I can be in Knoxville, Tennessee, where I choose to be because I'm able to sell to people in Paris, in London, in L.A., in New York, right from here.
The feedback that I've gotten from customers is they don't have to search through the halls of Walmart and see the same old, same old thing. They know that they come to me and find something completely different. Etsy makes it easy for me to have communication with my customer. It's a trust factor. I feel like they trust Etsy, they're going to trust me and vice versa.
Etsy has built this community of people who have gone from being really passionate about something and having it as a hobby, to doing it full time and making a career out of it.
Spoiler alert, the packages that you have there, you saw Matthew from the Pretentious Beer Glass Company, Each of you have some of Matthew's work there in your gift bags. I hope you enjoy. Next, I'm going to introduce Kruti Patel Goyal. Kruti Patel Goyal is the SVP of product at Etsy. She's been with Etsy for 8 years. She started her career at Morgan Stanley and then was at General Atlantic before coming to Etsy. She's had, I think, about every job you can have at Etsy. No one understands the secret sauce of Etsy better than Kruti. She's also a terrific leader, really analytical, and deeply understands our customers. As a result, I think she's incredibly well-positioned to take the product forward into the future. I'm delighted to announce Kruti.
Thank you, Josh, and hi, everyone. My role as head of product is to obsess over our customers. It's to understand their challenges, their needs, and how to delight them. While we've shown over the past couple of years that through greater product focus, we can drive the business forward, we still have a long way to go to deliver that truly delightful experience that brings to life the true promise of Etsy. I want to talk to you first about what it is that makes Etsy so special for our customers. In a word, people. Etsy is powered by people, and people are empowered by Etsy. Across the internet, e-commerce is getting more and more impersonal as it gets more automated. You can buy an item with one click.
It's picked off a warehouse shelf and packaged by a robot and delivered to your doorstep without you ever having to talk to another human being. While that's really great for some things like when you need to reorder toilet paper or buy a vacuum cleaner, for things that matter and things that you care about, you want and deserve a different experience. Etsy is different. It's a place where you buy things with meaning, things that you care about. You might even say things that spark joy. Why? There's a story behind every item, there's a human being behind every item. On Etsy, we connect shoppers with unique and distinctive items and the people behind them. What makes Etsy so special is our sellers and how much they care. You can see that in the handwritten notes that they include in their packages.
You can see it in the attention to detail and the care that they put into making their items, and you can see it in the personalized service they provide. That's also the part of the experience that our buyers often tell us they remember most about shopping on Etsy. It's a totally different experience. Items that are made just for you, that are hand packaged, and that are delivered directly from a seller to your home. It's a personal buying experience, one that we think a lot more people are in the market for these days, particularly at a time when people want to surround themselves with things that they care about rather than more disposable stuff. Our opportunity and our challenge over the next several years are to evolve our product to bring this truly people-first shopping experience to our customers.
I want to talk to you about our vision for a few key areas that we're investing in, particularly centered around the four parts of our right to win that Josh shared earlier. Best-in-class search and discovery, human connections, a trusted brand, and our collection of unique items. We looked at each of these areas through the lens of our customers to figure out what customer experience we need to create to really deliver on each of these. Let's start with best-in-class search and discovery. Right now, we have over 60 million unique listings on Etsy.
These are non-standard SKUs, which means that we can't tap into a catalog of standardized metadata that make it easier to match an item with what a buyer is looking for. What that means is that even though we have many amazing, beautiful, incredibly and high-quality crafted items on Etsy, our buyers have to do more work than they should to find them. As a result, search can feel really overwhelming. Our challenge is to take this vast sea of listings and make it feel manageable and curated just for you. To give you a sense of scale, right now, almost 60% of our searches yield over 800 search results, and a full third return over 10,000 results.
If you remember that search for coasters that Josh showed earlier that had over 204,000 relevant search results, just to put that in perspective, if you went to Michigan Stadium, you could put a set of coasters in every seat twice and still have 4,000 left over. That's a lot of coasters to evaluate. Buyers tend to purchase items from the first page of search results and over half the time from the first three rows alone. This makes it even more important that we surface the very best items for our buyers at the top of our search results. We need to narrow down that 200,000 to the top 20 for you. Our vision is to make it easy for buyers to find the items that perfectly match their taste. Today, we're just getting started.
We see a big opportunity to improve conversion rate and customer satisfaction by better surfacing all the gems across that vast set of search results. Here's how we get there. The first step is finding the most relevant search results for any given search term. As Josh mentioned, and Mike will talk about later, we've made a lot of progress here through our investment in context-specific ranking. Simply what this means is that when you search for a dress, we should only show you dresses rather than all the other items that we have that are related to dresses, like personalized dress hangers, dress patterns, or artwork with drawings of dresses. This is where we are now, where the search results we show you are more relevant to the search term. You're only seeing dresses here. There's still a lot more that we need to do.
Next, we have to identify among that set of relevant search results what the best quality dresses are. To do that, we're going to layer in more signals of quality, things like photo quality, customer reviews, and information from those reviews about craftsmanship. Then to get you even better search results, we have to identify the best of those dresses for you. To do that, we'll leverage machine learning models to map item style to individuals' taste, as well as an understanding of our buyer data across the marketplace to uncover patterns across groups of users that share similar interests and intents to help us better predict what any individual user wants. This is going to help us deliver a more personalized set of search results.
To see how this could work, let's take a look at how our experience could evolve for search and discovery over the next few years. First, let's meet two of our hypothetical shoppers, Janine on the left and Jackie on the right. They're both great Etsy customers, so we know a little bit about them already. They're both over 30. They love Etsy and come here regularly, shopping multiple times a year. Jackie and Janine are what we call habitual buyers, and Josh will tell you more later about what we know about our habitual buyers. Given that Janine and Jackie come here regularly, we know a little bit about their tastes, but this time, they're shopping on Etsy for something new, something they haven't shopped for with us before, which is a dress for a special occasion.
Janine's past activity on Etsy gives us a set of useful signals to work with, things like what she's viewed before, what she's favorited, and her past purchases. As you can see here, she frequently shops on Etsy for jewelry. What we know about her taste is that she likes bold patterns, bright colors, and she gravitates towards asymmetrical shapes and interesting materials. Jackie, on the other hand, tends to shop on Etsy for home decor. Her past purchases point to a style we call organic modern, and her favorites tell us she's into vintage items as well. She gravitates toward more muted, lighter colors and traditional shapes.
now they're on Etsy looking for something that they haven't shopped with us before, which is dresses, we should know enough about them to better predict the right results for them from the other shopping missions they've been on before with us. Right now, because they're both in the same market, the search results that they get are the same. In the future, what begin as two similar searches should quickly deliver two completely set of results, each tailor-made to the buyer's specific tastes. We expect that users as different as Jackie and Janine will find the high-quality, personalized items that they're looking for on page one of their search results. Ultimately, they should be able to quickly find the perfect dresses for each of them.
because of that, not only do we expect them to purchase more, we expect them to come back more frequently because they know that Etsy understands their unique taste. In summary, our plan is to continue to invest in improving relevance and then to methodically add on additional dimensions of quality and personalization into our ranking algorithms to make Etsy feel smaller, more curated, and more personal for every shopper. Connecting you to an item is only the beginning. We've sparked your interest and shown you something that resonates with you. what happens next on Etsy is the real magic.
that's because unlike other shopping destinations, behind a picture of the item is another human being that has their own story and that's excited to answer your questions and make an item just perfect for you. This is where Etsy already stands out, because you can find a seller and talk to them directly. In the example that we just talked about, Jackie could reach out to the seller in Lithuania and have that dress that she found made perfectly for her in a different length, like the one that you see on the right. This is just a fraction of what can be done on Etsy, because directly connecting with sellers opens up a world of customization options. Customization is really a spectrum.
You can choose from among a small set of structured variations to personalizing an item with your name or your monogram, to collaborating to create something entirely original. On most other sites, what you see is all that you can get, in some instances, you can choose between a small set of options, potentially a different color or a different material. that's where it ends. On Etsy, that's just the beginning, because buyers on Etsy have the potential to create something entirely new by collaborating with our sellers. even though most things on Etsy can be customized, only about 30% of items are marked as customizable, and those items have an almost 30% higher order value. We believe there's a huge opportunity to increase the awareness and understanding of customization options on Etsy. Communication is a really big part of the story as well.
For 8% of our transactions, buyers and sellers engage in a conversation through a tool that we call Convos before they make a purchase. Those inquiries convert at 12 times the normal rate and 200% of our typical AOV. When a buyer connects with a seller, they're much more likely to make a purchase and to spend more with us. While we know this isn't a purely causal effect, because when buyers reach out to sellers, they tend to have more purchase intent already, we believe that increasing the visibility and the relevance of this functionality will drive more engagement, more conversion, and higher AOVs over time. These data points show that there's a really big opportunity ahead of us to bring human connection to the forefront of our buying experience.
Our vision for the future is to make it easy for any buyer to connect with a seller to get the perfect item just for them. This is what the promise of a people-centric marketplace looks like. Let me walk you through another example of how our products could evolve in the future to really elevate this experience. Imagine another hypothetical shopper named Andrea, who just bought her first home and is looking for the perfect light fixture to put above her dining room table. Because of our improvements in search and discovery, she's able to quickly find Kyle, whose shop, Illuminate Vintage, has mid-century modern lighting that's just her style. She's able to quickly learn more about Kyle, his story, and his production process. The fact that he comes from a long line of lighting designers is a story that really resonates with her.
She likes everything about the shop and about Kyle, she's ready to buy something. There's one problem. She needs a light fixture that's bigger than anything that he currently sells in his shop. Anywhere else, she'd just have to keep looking. On Etsy, she can reach out to Kyle to design the perfect piece for her home. She could share her color palette, her room dimensions, and her inspiration directly through our app, he could share back different materials that he has available and shapes that he could make. Through easy-to-use collaboration tools, they can quickly agree on a direction. He can share a couple of sketches, she can approve what she likes. What comes next is the actual production of the light fixture. Typically, this is the most opaque part of the buying process.
You order something, then one day it shows up on your doorstep. What we envision is bringing to life a much more transparent process where you can see what's happening every step of the way. As Kyle makes the light, he can share updates and photos from the Etsy app, Andrea can follow along as her lights take shape from the design all the way up until it's packaged up and she receives her shipping notification from Etsy. Our product vision is to make Etsy synonymous with buying the perfect thing, whether it already exists or it only exists in your head.
That means that when you think about the spectrum of customization that we talked about, you can create everything from a necklace with a different gemstone to match your child's birthstone, to a blazer with a custom lining that reflects your unique style, to designing a dining table that matches the unique dimensions of your New York City apartment. To do this, we have to do three things. We need to elevate the visibility and the range of customization options and communication tools across the marketplace, introduce powerful tools for collaboration between our buyers and our sellers, and make the end-to-end process transparent. Let's talk about trust. Trust is critical for any e-commerce site, but even more so for a person-to-person marketplace, because the trust that we establish between buyers and Etsy has to extend to all of the sellers across our marketplace.
This is really powerful, not only for our buyers, but also for our sellers. Our over 2 million sellers individually set their policies and shipping prices across the over 60 million items on our marketplace. That means that every time a buyer purchases on Etsy, her experience is a little bit different each time. Essentially, it means that we're setting an infinite number of expectations for our customers right now. That creates a lot of friction. We want buyers to know every time they come to Etsy, no matter which seller they purchase from, that they're going to have a great and consistent experience. We believe that trust begins with setting the right expectations, and it ends with delivering a consistent experience to our buyers. Setting expectations.
Our buyers should never be surprised by what they receive, how much shipping costs, or how long an item will take to arrive. As early as possible in the purchase journey, we want to make sure that buyers have the right information to confidently make a purchase. That's everything from having the right product information available, to having information about availability and popularity, to information about return policies. We also believe that special takes time, and next-day delivery isn't always the right expectation. While there are millions of items on Etsy that are available for immediate shipping, most of our items are made to order. They're not sitting on a shelf in a warehouse. We want to make sure that buyers understand when things are made to order and when they'll be ready to ship.
Then we have to deliver with low-cost shipping and a clear timeline for delivery so that you know exactly what you're getting and when, every time you make a purchase. Let's talk more about shipping. We know that shipping is still an important point of friction in the Etsy buying experience. For buyers, there are a lot of factors that go into their shipping expectations, from speed, to price, to transparency, to returns. What we've learned through our customer research is that the factor that matters most to them is shipping price. Of course, everyone likes free, but our buyers do appreciate that Etsy's different. They recognize that our sellers are small businesses and that oftentimes items are coming across borders to them. They don't expect shipping to always be free, but they do expect shipping always to be fairly priced.
We really leaned into this learning over the last year. We made significant progress over the holidays. We invested in seller education to help our sellers understand their buyers' expectations. We provided them tools to make it easy for them to adjust their shipping prices and to meet their buyers' expectations. The result is that we changed shipping prices on Etsy. We really moved the needle on the number of items that shipped free by the end of the year. We know that there's still a lot more to do. It used to be that you came across free and fair price shipping on Etsy only occasionally. Over this past holiday, global visitors saw free and fair shipping the majority of the time. That still means that a lot of the time, our buyers were seeing shipping prices that were too high.
We're continuing to invest in this, so in the future, we're sure that we can meet buyers' expectations for reasonable shipping costs 100% of the time. Our vision is that buyers are never surprised by shipping on Etsy because they know it's always going to be reasonably priced, and that this removes a really important point of friction for our buyers. This brings us back to the foundation of our rights to win, our collection of unique inventory. This area is really all about our sellers and making sure that Etsy remains their first-choice place to sell. Investing in all the areas that we just talked about will continue to make Etsy the best place to buy unique and special goods, and that, in turn, will continue to make Etsy the best place to sell those goods, because Etsy is where our sellers' customers are.
Of our over 2 million active sellers, 80% are one-person shops. Our sellers spend almost half their time on marketing and business management. We know time is their most valuable resource. The number one thing that they tell us they need help with is marketing. This isn't surprising. Online marketing is complicated. There are multiple channels to learn and to optimize, from social to SEO. We believe that our sellers shouldn't have to become online marketing experts to grow their businesses. They should be able to spend more of their time creating, innovating, and providing great customer service.
We believe that we can not only help our sellers save time, but we can help them market more efficiently and effectively than they could on their own by leveraging our breadth of data about purchasing behavior across Etsy and our understanding of each of these marketing channels. Our vision is to create a simple end-to-end marketing solution for our sellers that makes it easier for them to run their businesses. What that could look like in the future is as simple as our sellers starting by telling us their goals. In this scenario, our seller can set several types of goals, ranging from setting sales to visits to conversion rate growth. She sets a goal to increase her sales by 10% this month.
Based on her goal, we can create a simple step-by-step plan across marketing channels that's effortless for her to put into action. In this example, we're recommending a range of actions from increasing her monthly ad budget, to creating a targeted sale for the upcoming Mother's Day holiday, to scheduling social posts, highlighting what we know are her top-selling items to help her drive more traffic to her Etsy shop. She could choose to accept individual actions or click to accept all the recommendations and get started right away. In this future state, what used to take her hours and hours can be handled in a matter of taps. Our sellers can't learn all of that from reading a book, and they shouldn't have to.
That's a look at a few of the high-impact areas our product teams will be working on over the next few years. As a product organization, we're energized and we're confident in tackling these challenges. That's because over the past year, we've realigned our approach to product development to be faster, more focused, and more customer-centric than ever before. One of the changes that we've made is organizing our teams into cross-functional product squads that are oriented around solving our biggest customer problems. Each of these teams has a KPI, a customer KPI, and business goals that they're focused on. We've also invested in tools and processes for them to bring rapid experimentation into our core process.
We're measuring our progress towards our KPIs by looking at experiment velocity, hit rate, and win size across our entire product portfolio every week, every month, and every quarter, that we're learning and iterating continuously. This rigor and accountability has yielded tangible improvements that are already moving our business forward. We're really excited about the foundation that we've laid over the past year because it paves the way for the next chapter of Etsy's story, where we really lean into the human connections and the personalized experiences that make Etsy really stand out and unique in a world of increasing automation and sameness. Before I turn it over to Mike, I'd like to share one more video with you so you can hear from some of our most passionate buyers in their own words.
Hi, I'm Rachel, Etsy helps customize my home.
Hi, I'm Nick. I've been shopping on Etsy for about three years. My first purchase was my wedding ring.
My name's Natasha, and I shop quite a lot on Etsy. It's a little bit of an addiction for me. I think my favorite thing about Etsy is that when I purchase something, the likelihood that I'm going to leave with something that only I have is pretty high.
One of my favorite pieces in my house are nautical charts that I have hanging in my dining room. I didn't see anything I was looking for on the normal retail sites, and when I looked on Etsy, I found exactly what I was looking for within the first page or two of search results, and I could customize the size that I needed.
When shopping on Etsy, I tend to be able to find stores that are interesting to me, but also may have something that I didn't even know what I was looking for.
I go on Etsy, it's always an exploration. This dress, I purchased from a seller in Bulgaria. I really wanted to create this robe that was voluminous and that had these colors that evoked the sunset. We went back and forth until we finally came to a design that we were both very happy with, and she produced in something like three days, and it was beyond what I was expecting once I received it. You'd be hard-pressed to find another e-commerce website where you can have that sort of experience.
I really like decorating with ginger jars right now. I looked on Etsy, and I did find one that was the perfect size, and it actually came with a very sweet note from the owners. It was handwritten. It was darling. It was very personal. I often shop at big retail stores, such as Restoration Hardware or Pottery Barn. I don't often get handwritten notes from those retailers.
There's a great human connection. I've had times where it's been coming up to a wedding anniversary, and I've left it a little bit too late, and having direct contact with someone, they've bent over backwards to help me get something express delivered. Having the backup of Etsy as a brand really gives me confidence in working with these small businesses. I might not necessarily work with someone that makes earrings in Lithuania if it wasn't backed up by the Etsy brand.
Shopping on Etsy, you value the items in a different way. They have a story, and they are a part of your life in a different way than just being a thing that you purchase without much thought.
I do like that I'm supporting an individual. They might have children, and it's going towards baseball lessons or something like that, and you just don't know where the money's necessarily going when you shop from larger retailers.
Great. Kruti has painted what I hope you'd find to be an ambitious and inspiring product vision. Bringing that product vision to life takes world-class technology, and we're fortunate here at Etsy to have a really world-class technology team. Next, I'd like to introduce Mike Fisher. Mike and I met about 15 years ago. In 2003, he joined PayPal to be the vice president of engineering, helping PayPal to scale through hypergrowth at the same time that I had joined eBay. Mike is a graduate of West Point and began his career as a helicopter pilot in the Army. He has a PhD from Case Western, where he taught artificial intelligence, and he's helped to scale some of the most exciting hypergrowth companies in recent memory.
Leading a world-class engineering team requires that you have both deep technical chops, the credibility of the team, and great leadership skills, and Mike has both. I'm delighted to introduce Mike Fisher.
Thanks, Josh. Terrific. Thanks, Josh. Hello, everyone. I'm real excited to be here. I'm going to walk you through where we are today and where we're going in the future. If you only take one thing away from my presentation, it really should be that Etsy is a technical innovator. We need to be a technical innovator because we're much more than just a marketplace. Everyone knows Etsy is the marketplace for creative goods that connects buyers and sellers from around the world. That's just the tip of the iceberg. That's what you see. Etsy's really much more than that. What you don't see is that we're really a big data and machine learning company. The Etsy marketplace generates over a billion events a day. These are buyers' clicks, add to carts, favorites, all of these create data.
This data is processed in real time and in batch with our big data systems. The process includes cleaning, transforming, ultimately turning that data into information. There's many consumers of that, one of which is our machine learning algorithms. They use this data to improve their predictive models. These insights and improved models are then placed back into the marketplace to improve our searches and our recommendations and our personalization. We have unique big data challenges that are specific to our platform. I'm going to tell you why that is in a few minutes. Our mission, our culture, and these interesting challenges are what attract the best-in-class engineering talent. We're extremely agile. This allows us to move very quickly. The results are that we're building a world-class shopping experience. Let's talk about the people who power this tech innovation.
Engineering is Etsy's largest department. Over half the company is part of our technology teams. This includes engineers, data scientists, product managers, designers, all working together. We hire great tech talent. We're doing it in a way that ensures we meet our diversity goals, because diverse workforces make us better. Raina's going to talk more about that soon. Many of our candidates have offers from the likes of Google and Facebook and Spotify. We even have engineers that leave for these big companies, and they come back. They come back because we have an inclusive and innovative culture. They come back because we're nimble and fast. They can have a tangible impact on the business outcome. They're not that proverbial cog in the big wheel. Our engineers get asked to speak at all types of top conferences around the world.
QCon in San Francisco, Knowledge Discovery and Data Mining in London. We publish papers in peer-review journals. We are considered thought leaders in the tech industry. We have a brand that's known as Code as Craft. We sell out our speaker series in a matter of minutes. Just last week, we had over 300 people just in this facility watching one of our speakers. We have tens of thousands of people who read our blogs. They want to hear what we have to say about technology and putting that to work. We have open sourced over 60 projects that other technology companies are using. What drives our engineers is that they're solving some of these complex and interesting challenges in tech. One of the most complex challenges for us is creating best-in-class search.
Why is search so difficult for Etsy compared to all the other e-tailers? Well, many of the things that make Etsy so special makes it extremely complex. There are no SKUs. There's no catalog. Items are one of a kind. They're often customized. That means when they're gone, they're really gone. We don't get a chance to sell the same shirt 500 or 5,000 times. Sellers label and tag the items how they want to, how they want to describe their items. Combine this with over 60 million items in inventory, makes it for an interesting tech challenge. Let's talk about how search works at Etsy today. We build our search indexes only using title and tags. Descriptions currently are too unstructured. We allow the sellers to put whatever they want in that.
They can tell their story, how they built it, where they sourced it, anything they want. Image quality is currently too varied. It could be super high-quality images to not so high quality. Until recently, image recognition was extremely difficult. After we retrieve all the items based on the title and the tags, we then use predictive conversion rate. Three items that we currently use. You can take a look at this listing from one of our sellers, and you can see that the title is pretty stuffed with words, adding to the difficulty. Of course, sellers try to game the system. An item is listed around the holidays, often sellers will put that into their titles because they know people search for Valentine's Day gifts.
You can see this seller did it, even though the pillow really doesn't have much to do with the Valentine's Day holiday. The first big challenge is determining what is actually relevant. If you ever watch the show "Silicon Valley," this is the hot dog, not hot dog problem. To solve this problem, we also turn to machine learning, and we're making great progress. Before we talk about that, let's look at how search used to look on Etsy. Say a buyer was looking for a cool, unique desk, and her favorite color was blue. She typed into the search bar at etsy.com, "Blue desk." 23,000 results come back. She would see things like a blue desk calendar or a blue desk lamp or a blue desk clock. Oftentimes, the most relevant items, the actual blue desk, might be back on page five or 10.
That can be a very frustrating shopping experience. Fast-forward to today. We take search terms and the displayed results from everybody's product searches for the last several weeks, and we analyze them based on dozens of features, including how the buyers interact with them. Do they click on them? Do they browse them? Do they add to cart? Do they favorite them? Do they eventually purchase them? Then we build machine learning models that can re-rank the top 800 search results in real time. The first page shows the most relevant items. Considering, as Kruti mentioned, that the first page of search results generates 83% of the search purchases, it's critically important to get those first few listings right. Moving a relevant listing from page five up to page one is important for our business.
Today, when a buyer does that exact same search for a blue desk, she sees much more blue desk. Now, you'll see we're not perfect. There's a blue desk lampshade up in the top right, but we're much, much better than we have been. Overall, these search improvements have unlocked nearly $260 million of annualized GMS over 2017 and 2018. As I mentioned, we still have work to do. For the basic search result I mentioned, we only use conversion rate, titles and tags. There's so much more that we can use just for that basic search result. In the future, we can leverage machine learning to help us understand the information that's in images.
We can also use it to help put structure around those descriptions, so we can take that unstructured data and descriptions and turn it into useful information, all to improve that basic search. For re-ranking, we're already using machine learning, but there are still dozens of other factors that we can add. We can start looking at the seller ratings, shipping times, price, LTV, all of this. We currently only rank the 800 top results, but we have 23,000 back from our search queries. We can use a lot more of that data. Relevance is all relative, right? The results for that same search term should look different depending on who's searching for it. How do we get this personalized? In the past, if you're in the U.S. and you were throwing a football-themed birthday party for your child, you'd be pleased with this.
If you're in the rest of the world, you might not be so pleased with these search results for football, right? In Q4 of 2018, we started taking your location into effect or into account into your search results. Location is still the only lever that we're pulling for personalization. Again, there's so much more that we can optimize for. In the future, we can use many, many more contextual cues, holidays, local events, geographical trends, weather, all of this, your style, all of this that we've really just untouched. There are countless ways that we can continue to optimize search. Search itself is just one of many, many areas that we can apply machine learning. Last year, we began to use some of our machine learning models on Pro List, and the result was that we generated $46 million in annualized revenue.
We are truly at the early stages of what machine learning can do at Etsy. There are so many other areas. We can expand into categories. We can help buyers find out that there's other categories that might be interesting to them. We can help highlight that best quality items. All of these we can apply machine learning to and improve across the marketplace. Our machine learning innovations take an immense amount of compute power. We need to be able to process that, all that data, and turn it into useful information. Our data centers had finite compute power, so we could only use relatively simple models in the past to do this. Not only do we need massive amounts of compute power, but we need them in elastic manner.
The graph you see are some of the models being trained, and you see how sort of spiky they are. It's a very bursty nature of training, and so we need that elasticity. When we need the demand, we need to have it on hand. When running in our data centers, we often had to budget more than a year out, often order six months out to have the hardware built for us, receive it, rack it, and eventually get to use it. If we ordered too little, we couldn't run and train all of that machine learning models or process that data that we wanted to. If we ordered too much, it was just wasted. It sat there idle. Still consuming energy, but not being used. This is one of the reasons that we've moved to the cloud.
For Etsy, migrating to the cloud has given us this immense amount of compute power that we need, and it's given it to us in an elastic manner. We can spin up literally hundreds of servers in a matter of minutes now. What used to take six months or longer, minutes. It allows us to utilize that compute power efficiently. When we don't need it, we can turn it down or let someone else use it in the cloud. All of this increases our ability to make more sales and happier customers. Within nine months of beginning our migration, and this really speaks to the level of technical expertise that Etsy has, within nine months of beginning the migration, the Etsy marketplace and the mobile apps were already migrated to the cloud. Oftentimes, our peers in the technology industries would take years to accomplish this.
We're now more than halfway done with this migration, and we're on track to complete early 2020. In the past, we've had to support infrastructure like data centers and like a lot of other homegrown technology, whereas most of our engineers had to be focused on that to support it. Today, we have more engineers than ever focused on etsy.com, working on building features that customers see and utilize. Our goal over the next few years is to even continue to increase that. So we can eventually have more than 70% of our engineers working on product features and having to support less and less of the infrastructure, leveraging trusted third-party providers like Google and others to take up some of that infrastructure for us. Having and maintaining process and tools that allow the engineers to work efficiently is vitally important.
Companies that follow non-agile practices like traditional waterfall development methodologies might release code or make a release once a year. Even your agile companies, right, that follow these agile methodologies, they typically push code, push a release every two weeks. We are what's called a continuous integration, continuous deployment shop. Being able to push code very rapidly means we can learn very quickly, and we can make adjustments to ensure that we're maximizing the impact of our features. Kruti touched on this earlier, but in the average quarter, we run hundreds of experiments. About 30% have successful outcomes initially. The point is that we aren't successful, but we're nimble and we're fast, and we can iterate until we have a successful positive outcome. If we didn't experiment to improve the marketplace, everybody loses.
We learn from everything we do, and we use that knowledge to make meaningful improvements on our product features. Let's take a look at an example of how we experiment. When a buyer finds what they want, they might not actually go through with the purchase. They're not ready. They get distracted or something. When they come back, because that item is unique, it might be gone. We're encouraging them to complete the transaction right there with something we call nudges. This is a listing without a nudge. By adding a prompt that alerts the buyers that others might have it in their cart, we're aiming to create that sense of urgency and help them understand that this might be gone when they come back. How do we know what the best nudge is?
There are many variations of nudges, right? We can have alarm bell that says, "Don't miss out," or a shopping cart that says, "Other people want this," an hourglass, "Almost gone." We test all of them, and we figure out what helps the buyers complete the transaction so they don't lose out on that unique item. Of course, it was the hourglass. These same principles of agility and accountability that we use in our experiments are baked into all of our work. Our teams take their work extremely seriously. We keep scorecards. We measure the GMS and KPIs. We review our metrics, and we measure success in what we do. We've developed competency matrix for our engineers' career growth that helps them understand what the expectations are for delivery, domain expertise, communication, and problem-solving, and many others.
In conclusion, we have really large, unique technical challenges. The mission, the culture, and these challenges are what attract really strong engineers. Making fundamental improvements and changes like moving to the cloud empower that world-class engineering team. Improving search is just one of many examples of things coming together. There are many more opportunities across the entire marketplace. Thank you very much, and we'll take a break until 10:45 A.M.
Great. The next speaker I'd like to introduce is Raina Moskowitz. Raina spent about 13 years at American Express at a variety of roles from strategy, products, marketing, and operations. Most recently, she had built and led an organization that grew to a $200 million revenue organization, deepening American Express's relationships with its own customers. Sounds like a familiar challenge. She joined us in early 2018 and has really done a fantastic job of helping us to crystallize our strategy, as you're seeing laid out here today, significantly up-level the quality of our member support, and also really do so much to improve our people operations. People are the heart and soul of Etsy. Doing a better job recruiting great talent faster, driving engagement, training and learning, and retention of that talent. She's done a fantastic job leading that team and making really great progress.
Before I hand it over to Raina, I want to take a quick second to talk about our guiding principles. We've spent some time thinking about how we show up, and in particular, how do we show up on our best day. We've crystallized those into five core guiding principles. We use these as a rubric when we're interviewing employees, and we use this as a rubric when we're thinking about things like promotions, raises. We really operationalize these values, and I'm going to show you a very brief video of some of our employees talking about the employee experience here at Etsy.
Etsy's mission is to keep commerce human. To me, that means two very different things. One is the sellers and buyers that we support every day. The other side of it, which is near and dear to my heart, is keeping commerce human by creating opportunities for the 800 people that work here at Etsy. Employees use the guiding principles to make sure that we stay focused on the things that matter, and in remembering that first and foremost, what we're doing is creating opportunities for people.
We commit to our craft" is something that I think about a lot because I think that's one of the aspects of Etsy culture that I really appreciate, is that craft is not limited to creating goods, but it's also creating skill sets that enable you to be a better professional. Something that was fascinating to me when I first joined Etsy was how we minimize waste.
There are also a bunch of little things that Etsy does, even here in the office, that work towards reducing our carbon footprint. I can see all around us in everything that we do, even in the small decisions that we're making from a facilities perspective, that Etsy is a company that cares.
From more the operational and how we conduct business point of view, I would say one of the things that we're trying to do is making sure that we're conscious of the meetings that we're having. I think thoughtfulness of time is really what allows people to grow in their own individual ways, but also be present for the moments where they need to be present.
For the first time versus anywhere else that I've ever worked, I feel like I can bring my whole self to work because the things that are different about my perspective and my experience are actually the things that bring the most value if I bring them to the table.
You want to embrace ideas that challenge yours because those are going to end up in a better product and a better outcome. Etsy, I've seen do the work. The current leadership really believes in this. It's reflected in their priorities.
It's important to dig deep because if you're not digging deep, you're not solving the right problems. Someone could say, "This is happening, and I don't like it." It's just like, "Why do you not like it?" It's about understanding the core problem that they're experiencing.
Our model lets us continue to scale this digital marketplace without kind of inevitably falling into that pit of becoming a purely transactional, faceless website. This notion of digging deeper is really about how do we keep that in front of people.
Etsy is a company that really cares about management as a discipline. There is a larger responsibility here, and leading with optimism is a huge part of that.
You should feel empowered to have a voice that contributes directly to your end goals, and I think that that's something that Etsy caters to beautifully, and that, to me, is something that is so unique.
If you're a person that wants to work somewhere that feels real, feels like you can make a difference, feels like you can speak up, then this is the place for you.
Please welcome Raina Moskowitz to talk about our human capital strategy.
All right. Josh, thank you for the intro, and thanks again to everyone for spending time with us today. I'm so glad that you had a chance to see that video and hear a little bit from the team, because we get asked all the time about our culture. What's changed, what's driving the strong performance, and how are we attracting and retaining top talent? We know that human capital is an important component of your investment lens. Today, I'll talk about the journey that we've been on over the past two years, and I'll focus on our talent and engagement metrics, as well as our economic, social, and environmental impact. Everything you've heard so far today is about how we'll be able to achieve our vision to be the world's best marketplace for people who value craftsmanship, self-expression, and the search for special.
While our product experience will continue to evolve, our mission remains constant. We're here to keep commerce human. This matters because in the competitive tech talent market that we're in, people want to work for companies with a meaningful mission. They want to make a difference in the world and have a positive impact on their customers each and every day. Our mission-driven culture is what allows us to compete with the likes of Google and Facebook in markets like New York and San Francisco, and you heard Mike speak about our amazing engineering culture as well. Our talented team drives strong business performance and ultimately enables us to deliver even greater economic, social, and environmental impact. It's a virtuous cycle, which in turn allows us to attract the best talent. I want to acknowledge that building a great culture isn't a new concept for Etsy.
Etsy has always been mission-driven, and it's always been a great place to work. What is new about our culture? Over the past two years, we've raised the bar. We've held ourselves more accountable to even more ambitious goals on behalf of our customers and on behalf of our stakeholders, and we've made the transition to become a high-performance culture. To do this, we've evolved our people and talent processes to drive greater alignment and accountability at all levels of the company and across all functions. We've introduced quarterly goal setting, establishing target and stretch goals at every level of the company, from our frontline service center to our CEO. We've expanded variable compensation more broadly from 9% to 100% of our employee population so that individual and company success are even more aligned.
We've set expectations for our managers about what great leadership looks like at Etsy, and we're investing in training and career development for Etsy's future leaders as well. We've also evolved how we operate. We've instilled a metrics-driven mindset, creating accountability meetings and dashboards so that Josh and every level of the company can understand how we're progressing against our goals on a daily, weekly, or monthly basis. You heard both Mike and Kruti talk about this as well. We're organized in dedicated squads of product managers, designers, and engineers who are focused on improving our most important customer experiences. We've invested in the training and the tools for prototyping and rapid learning so that we can continue to increase our experiment velocity and hit rate. All of these foundational processes and investments have translated to greater efficiency and productivity of our team.
The teams are clear on their mission and on what success looks like. As a result, we've nearly doubled revenue per employee over the past two years. It has been a major cultural transition, and you can see when looking at our voluntary attrition data, that many people opted out back in mid-2017, and that's okay. We hired nearly 300 employees last year. For the past 12 months, Etsy's voluntary attrition of 14.5% has been below the tech industry benchmark of 15.9%. The team who's at Etsy today is highly engaged and excited about our future. That shows up in our employee engagement results. We've seen impressive increases in employee engagement based on our most recent 2018 survey. Confidence in the executive team increased to 84%, and belief in Etsy's success over the next three years increased to 85%.
Both of these metrics are above the tech industry benchmark, we're confident that this upward trend will continue. What really allows us to attract top talent and keep them engaged is that our business model and our mission are 100% aligned. Etsy gives creative entrepreneurs around the world the opportunity to earn a living while doing what they love. We make money when our sellers make money, and 39 million buyers around the world find that special item that they can't find anywhere else. At a time when people are questioning the role of technology in their lives and in society, we know at Etsy, the mission is our day job, and that is why our team is excited to show up at work each and every day.
Of Etsy's 2.1 million active sellers around the world, 87% are women, and 49% stated that they started their shop to meet a financial challenge. We empower entrepreneurship and provide the tools they need to grow a thriving creative business. The story's bigger than just the sellers themselves. Etsy sellers and their shops have a halo effect on their communities and our economy. In 2018, we started officially tracking the economic impact of our Etsy sellers in the U.S. For every dollar an Etsy seller earns, she generates $2.33 for the U.S. economy. Overall, that means U.S. Etsy sellers contributed $5.37 billion to the national economy in 2018. They created 1.52 million jobs in the U.S., which is basically enough to employ the entire city of Philadelphia. In order to serve our diverse set of buyers and sellers, we've built a diverse and inclusive workplace.
Again, this is not a new concept for Etsy. The team has been focused for many years on recruitment and talent development efforts. We're leading the way in gender diversity throughout the company. As of December 31st of last year, 56% of global employees identified as female or non-binary, including 52% of leaders throughout our company. We also stand out as one of the very few public companies whose board and executive teams are both at least 50% female. At a time when gender diversity in technology roles is being scrutinized, we're proud that 38.4% of our tech roles are female, nearly double the average of other tech companies. At Etsy, it's not just about the numbers. You heard in the video earlier that one of our guiding principles is to embrace diverse perspectives.
We know that having a range of voices at the table leads to better decisions and outsized business performance. While we're proud of the progress we've made, our work isn't done. In addition to continuing to strive for gender parity, we're also focused on being the employer of choice and increasing our racial diversity at the company as well. In addition to our economic and social impact, as a global marketplace, Etsy has always been conscious of our ecological impact, and we know that this is something that's important to many of you as well. In 2016, we committed to powering our operations with 100% renewable electricity by 2020. Last year, we signed a power purchase agreement that enabled the development of a new solar farm in Virginia.
Our migration to the Google Cloud Platform and installation of solar panels in our offices place us firmly on track to meet our goal. While all of these achievements are major strides towards addressing our environmental footprint within our direct control, we believe we need to hold ourselves to an even higher standard. 98% of Etsy's total emissions stem from shipping. While we don't actively control the process of shipping between buyers and sellers, we are in a position to do something about the environmental impact. Last week, Etsy became the first global e-commerce company to offset 100% of carbon emissions generated by shipping. We're doing this by investing in environmentally friendly projects, like projecting over 10 million trees that would've otherwise been cut down and sponsoring a wind farm in India. We're doing it for less than $0.01 per package.
When we verified the cost with our partner, 3Degrees, the conversation amongst the team changed from, "Should we do this?" Why haven't we done this yet to why haven't others done this already? We're leading from the front, and we hope to mobilize the industry. To start, we picked up the tab for a day to offset all e-commerce shipping emissions last Thursday in the U.S. In addition to being good for the planet, it's also good for driving loyalty. Customers are shopping their values. 88% state that they're rewarding companies that address social and environmental issues. Our own Etsy customers agree. In a recent survey, 90% of our buyers stated that environmental sustainability matters to them, we know it matters to you as well.
With our recently filed 10-K, we believe we're the first e-commerce company to fully integrate our economic, social, and environmental reporting metrics alongside our financial metrics. We're very proud to lead in this important area. At Etsy, we've long held the conviction that we can be both a great business and a great corporate citizen. In fact, we believe we can't be one without the other. When you spend any amount of time with the team in our building or at any of our buildings, around the world, you know that this is true. It's deep in our DNA. That's why we continue to attract the best talent and drive strong results. Don't just take my word for it. In a recent employee engagement survey, an Etsy engineer provided this quote, I think it sums it up pretty much perfectly.
Etsy is the happiest, most diverse, most doing things that matter company I've worked for. It makes coming to work a joy." Thank you to that engineer for that awesome quote. Thanks to you for being here today. With that, I'll turn it over to Josh to talk about our levers of growth.
Thank you, Raina. We're going to look forward at levers of growth, I do want to take a second, to build on what Raina said. There's a narrative out there that somehow the strong financial performance that we've had over the past couple of years has come at expense of being a good citizen, nothing could be further from the truth. The goal of a company's got to be to grow the pie for everyone so the pie gets better, I do believe we've done that. In this world of fractured politics, if our corporate leaders can't step up, if our businesses can't step up and be both good citizens and good businesses, then where are we? It's the exact same focus and discipline and accountability that we've brought to being a good business that we've also applied to citizenship.
We've looked at, if we want to be a good citizenship, what does that mean? Let's define it. We came to three areas. It's about economic empowerment, which is, in fact, our day job measured by GMS. We believe we achieve that better with a more diverse workforce, we believe we can do that with low and getting to zero carbon footprint on our planet. Not just saying it, but actually publishing metrics and holding ourselves publicly accountable has been a critical part of our journey over the past two years. In the 10-K that we're filing, you will see all of the metrics around economic empowerment, diversity, and environmental footprint published as part of our 10-K integrated into our financial results. They come together, they're part and parcel of the same thing.
You will see us making progress and holding ourselves accountable to make progress. We passionately believe here at Etsy that we can be both a great citizen and a great business at the same time, that those two things are, in fact, self-reinforcing. Let me turn it over to levers of growth. We're going to talk a little marketing now. I know this is front and center on many people's minds. Let me start with buyers. We start from a strong foundation. With active buyers, these are people who've bought on Etsy at least once over the past 12 months. Active buyers have grown at a rate of 18%, they've had great experiences. The moment of truth when you're a buyer is when the package actually arrives, you open the box. By and large, they're delighted by what they see.
The net promoter scores of recent purchasers are consistently over 70%. These are very, very strong scores in e-commerce. Etsy's a trendsetter. In fact, we have over 26 million mentions on Instagram right now, four times as many as Amazon, nine times as many as eBay. We punch way above our weight being at the center of the conversation. People aren't just talking about us, they want to hear from us. We have two and a half million followers on Twitter, almost as many as Amazon. Again, punching way above our weight. This shows up on the site. We have over 100 million unique visits each month to Etsy, and 85% of them come for free. This means we're not dependent on upstream sources of traffic like virtually every other e-commerce site in the world.
When you have a terrific experience, you can certainly throw fuel on the fire through smart marketing. Let's take a look at some of our brand health metrics. The top of the funnel, prompted and unprompted awareness. Generally speaking, given that we are a still very small $4 billion e-commerce site, it's pretty good. You don't see a lot of $4 billion e-commerce sites with unprompted awareness of about 40%. The bottom of the funnel is also not too shabby. The percentage of people saying that they're loyal and more than loyal, that they would actually go out and advocate, they'd tell their friends, "You should be shopping on Etsy." What's unusual is the concave nature of this shape. It's very unusual to see more people say they advocate than say they have a purchase intent. Probably most of you have experienced this.
You've talked to buyers and they say, "Oh, I love Etsy." You say, "When's the last time you shopped on Etsy?" "Oh, come to think of it's been a while." "Why haven't you shopped there more often?" "You know, to tell you the truth, I don't really know why I haven't been there more often." How many of us have had that conversation? I have it every single day. We're going to dive into that a little bit. We're going to talk about some of what's underpinned that, and how we're going to unlock that opportunity. Let me start by talking about how we're going to grow active buyers, then we're going to talk about frequency, then we're going to talk about average order value. These are three powerful levers of growth, and we think we have meaningful opportunities to grow each of these three.
Collectively, they can be extremely powerful. Starting with active buyers. We don't need to serve everyone. We just need to serve people who are likely to love us. We survey the U.S. and global consumer base about once a quarter. We survey about 3,500 consumers, and we ask them a set of questions about attitudes and behaviors. About 40% of them turn out to be what we call expressives or seekers. Expressives are people who see what they buy as a form of self-expression. They want to be someone who wears something no one else has seen or who gives the most thoughtful gift. They're seekers, meaning they enjoy the treasure hunt. They like to spend time shopping to find something unusual. These are expressives and seekers. It's about 40% of consumers. They spend a little more, disproportionately more than the average consumer.
Prior to 2013, we did exactly zero marketing in this company. While we're a 13-year-old company, we're five years old when it comes to marketing. The marketing that we've done so far has been exclusively at the very, very bottom of the funnel, almost exclusively on Google Product Listing Ads. In fact, even today, over three-quarters of Etsy's marketing dollars of the media spend that we put against our advertising goes to Google. Let's take a look at what that shopping experience looks like. You're on Google. You type in a keyword, let's say wedding dress. You're taken to a Google search result page with stuff related to wedding dresses, and lo and behold, there is a Product Listing Ad for an Etsy wedding dress. You click on the link, and it takes you to a page that has exactly one wedding dress on it.
Now, maybe we got it exactly right, Mike and Kruti talked about the problem of search, and we just nailed it, maybe. Maybe we didn't nail it perfectly, and in fact, that page is a dead end. Where do you go from there? Not only have we not shown you the other robust selection of wedding dresses we have for sale, but you've left with the feeling that Etsy only sells wedding dresses. If we're lucky, you think maybe Etsy only sells wedding-related stuff. You certainly have no sense of the breadth and depth of offer that we have on Etsy. That happens again and again, hundreds of thousands of times a week.
In fact, from paid marketing, visitors are six times more likely to land on a single listing page than they are on anything higher in the funnel, the homepage, or a curated selection of items or anything that's giving you a sense of breadth and depth in selection. The obvious opportunity here is for Etsy to move higher in the funnel and get a chance to tell our story. The story we're telling you here today about all the things you can find on Etsy, we need that opportunity to tell that story to consumers. We're optimistic about the fact that that storytelling can really do a lot to move people and do that storytelling in an integrated way, with high funnel moving to mid funnel, and finally getting to low funnel. We talked about some television advertising that we did in the fourth quarter.
We're pleased with how that is going, I thought I'd take a quick minute to show you one of the ads that performed particularly well.
It all centers around my dog. His name is Chaos.
I'm bohemian boho.
He's like, "Oh, the pen you got me, I love that pen.
I just typed in red ballet shoes.
I want to find pieces that represent who I am. That's what I get from Etsy.
He handed me a box. It was a necklace with my dog's name on it.
I get a lot of compliments on these.
Something handmade and personalized on Etsy comes with love.
Go to etsy.com for anything you want.
That was one of our better-performing ads. Those are not paid actors, they're real people. They're showing you the real things that they bought on Etsy. You'll notice that there's a diversity of products and diversity of occasions from a diversity of people. That kind of storytelling, we think, is very powerful, and we're optimistic about how that can do in the future. Let me pull back now and talk a little bit about our marketing philosophy, and Rachel is going to talk about this in more detail in her section. I think of marketing in two buckets. You've got your proven channels and your immature channels. With your proven channels, Google is the one proven channel we have today.
It's where we spend most of our money, and we know that we can get a positive ROI on at least some amount of spending. The first dollar you spend in a proven channel is going to get a very, very high ROI. Then you're going to spend more and more and more, and the marginal return on the next dollar spent is going to decline until the marginal return on the next dollar is something less than whatever return threshold we've set. The trick is, first, figuring out the shape of that curve so that you can spend up until you hit that return threshold, but no more. I want to at least have our team achieve that, in a perfect world, we'd get to the point where the marginal return on the next dollar spent is just below the threshold.
We don't manage to a fixed budget, you only get to spend this much money. We manage to a return threshold, get to the point where the marginal return on the next dollar is lower than our threshold. That's why when you look at when we publish results and you see payback periods and you say, "Gosh, the payback period feels really fast. Why aren't you spending more?" It's because we're not thinking about the overall return. We're thinking about the marginal return on the next dollar. The work of the team in a proven channel is to shift that curve up. There's new people entering the market. They're bidding against you. That's creating competitive pressure, but there's a lot we can do. What's the quality of our data feed? How smart is our bidding strategy? How good is our segmentation? What's the quality of our landing pages?
There's lots we can do to shift that curve up so that we can either spend more money at the same return or get a higher return for the amount of money that we're currently spending. That's the work in a proven channel. We've got new channels we're testing. These are immature channels, for those, we set a budget. We say we're willing to spend $X million to try to create new channels. In those channels, what we're looking for is something that has the potential to really scale. Something where as we unlock it and make it work, we can actually spend meaningfully against it to acquire new customers and reengage our existing customers. Television is one example of a new channel we're investing against. In the early days, it's often the case that you're not ROI positive the day you launch.
Sometimes you get lucky. Mostly you're testing what's the creative strategy in front of what audience? What are the channels that work? What time of day? All of these things are things that we test. What we like about television is not only does it get top of funnel and give us a chance to tell the story, but it's a pretty big channel. You can imagine the shape of that ROI curve might be relatively flat for quite some time. You can spend and spend before you start to run out of audience. Facebook, Instagram, Pinterest, these are all examples of highly scaled channels that we ought to be able to crack in time and will allow us to scale our business even faster.
We're making investments in each of these that we consider to be research and development until the point that we prove them, we figure out how to crack the code, and then they become proven channels that are not managed to a budget, they're managed to an ROI threshold. Within each of these, we're just starting to optimize the funnel. Again, we're only five years old when it comes to marketing, and we have not invested significantly in marketing technology, building that MarTech stack that gives us robust analytics, robust segmentation, great ability to automate landing pages, and on and on and on. I'm very optimistic about our potential in the future to really scale our marketing and drive new active buyers into the business. Once we've brought those active buyers into the platform, of course, we want more frequency, and we've talked quite a bit together about this.
We have 39 million active buyers with a very high net promoter score, and why don't they shop there more often? That's a great question. One of the first things most businesses would do is look at a best customer analysis. Well, who are your best customers and what behavior are they modeling? We are only now beginning a best customer analysis. I'm excited about the opportunity that that has for us and for people who ask, "What inning are you in?" I'd say, "We're only starting best customer analysis right now." There's so much potential here. We've defined this best customer as what we call a habitual buyer. There's about 2 million habitual buyers. They come on at least six separate purchase days per year, and they spend about at least $200 on the site. That's a habitual buyer.
We've got about 2 million of them right now. When you dive into who are the habitual buyers, I take heart from the fact that they're very diverse. Different parts of the world and the country and different ages and income levels, and it looks a lot like the snapshot of America or the geographies that we're in from a demographic perspective. Psychographically, they share a few characteristics. They express themselves through the things that they buy. Thoughtful gift-giving is important to them, and they appreciate the ability to make something just to their liking. There's a lot of people out here who look like our best buyers but are not yet our best buyers. That's our opportunity. We dive in and we examine what are their behaviors and what can we learn from them. For search and discovery, they've become search pros.
I gave you a challenge earlier in the morning. I said, "Go back and look at the last 90 days of things you've bought online." If I could pair you with a search pro, with a habitual buyer to sit next to you, I'm confident you would be delighted with the quantity and quality of things they would find relative to things you'd bought in other places. How do we take their experience and build it in as the core default search experience that everyone is experiencing so you don't need to be a pro? That's exactly the work that Kruti and Mike have talked about. By learning how our habitual buyers navigate and search the site, that's driving and informing the product roadmap behind the search and discovery work that Mike and Kruti talked to you about. Human connections.
The habitual buyers understand the power of things like convos and personalization and customization. We're using that to inform the roadmap that Kruti talked about of how do we make convos and personalization and customization really core to the whole experience. Then trust. Habitual buyers understand things like how shipping and returns work. How do we make that an expectation that every buyer can have when they come to the site? There's so much power in taking the experience our habitual buyers have and making that the default experience for everyone, and that's driving and informing the product and marketing roadmap that we have. We're also taking those habitual buyers and saying, "Who are the lookalikes to those buyers?" As we market ourselves out in the world, maybe we don't need to market to everyone. Maybe we can disproportionately market to people who look like our habitual buyers.
That's very helpful in terms of driving efficiency. As I think about frequency, one of the big pillars of strength is there are so many occasions for which Etsy is relevant. There are millions of people who come to us reliably every year on Valentine's Day. There are millions of people who come to us reliably every year on Mother's Day or Halloween. By the way, if you haven't tried Etsy for Halloween, it's unbelievably good. Holiday season is great on Etsy. Each of these cohorts tend to think we are only for Valentine's Day or only for Mother's Day or only for Halloween. We haven't even started with the life cycle events.
How about weddings or furnishing a new home or going back to school or having a baby or getting that special dress for that special event or all of the other life cycle events that happen throughout the year? What gives me great confidence is we don't need to invent any new purchase occasions. We already deliver extremely well on so many relevant purchase occasions. Our opportunity is just to string them together so people who've experienced one of these great purchase occasions realize that we can do so much more for them. We're just getting started in building the capabilities to do this, segmenting our buyers, speaking to them at the right time, in the right place, and delivering them a personalized customer experience. The last lever of growth I'd like to talk to you about today is average order value.
It's frankly, something we haven't spent much time on over the past two years, an area that we think is really rich going forward. We're a long way from table stakes. We're one of the last e-commerce sites you can go to that doesn't show you people who bought this also buy that. It's a really big opportunity to grow cart size, we haven't yet begun. How about thinking about how you merchandise high-quality inventory? We talked a lot about quality and raising the bar on perceptions of quality. Well, here's an Etsy experience. When's the last time you walked into a jewelry store and saw a $14 pair of earrings sitting right next to a $4,000 pair of earrings? Never happens. Never. 0% of jewelry stores do that. It happens hundreds of times a day on Etsy.
The chance to visually merchandise our inventory and curate collections that make sense is really going to help to allow us to sell even more diverse and high-quality items. Trust isn't built overnight. It takes time. We're so good at accessorizing the home or the person or the event. Consistently delivering a great experience on these lower stakes purchases gives us the opportunity to win in the higher stakes purchases. Getting comfortable investing in items that you'll live with for years to come. Inside the company, we call this the cushions to the couch strategy. To sum up, we're just getting started as we drive our levers of growth, we believe there's rich opportunity in driving more active buyers, having them come back more often, and for even higher AOV items.
With that, I want to turn it over to Rachel Glaser, our Chief Financial Officer, to take us through some of the financial results and the outlook going forward. Rachel has been my partner here pretty much since the day I started. I think she started a week or two after me. She began her career at Walt Disney. She went to Yahoo and then was the CFO at Move Group and Leaf. She's got a rich and diverse set of experiences as a CFO of both companies that are consumer-facing and media, as well as two-sided marketplaces. She's done a terrific job of bringing real rigor and discipline to the finance organization, most importantly, embedding the finance organization as a real partner of all of our operating teams.
You've heard each of us talk about accountability and metrics and using that as a core part of how we operate every single day. Rachel has done a great job bringing her team together with the product and eng teams to really bring that to life. With that, I'll turn it over to Rachel Glaser.
Thank you. Thank you for that really nice introduction, Josh. It is so nice to see so many of you here today at our home base. We've seen you all over the country and all over the world in some cases, and here you are at Etsy. Thank you for coming. After listening to my colleagues talk about our mission, our product strategy, our marketing opportunity, our strength in technology, and our awesome culture, I am very excited to say we believe we're poised for some long-term future growth. I'm going to talk to you about how that translates into the economics of our business. I'm going to talk about four areas.
First, I'm going to show you our current GMS growth trajectory over the past seven quarters or so and how we're bending the curve. Second, I want to talk about why we love our marketplace model and the discipline around our investment and resource allocation. Then I'll talk about our market size a little bit more. Josh touched on it, but why we believe it's really still a very early stage opportunity. Very last, I will give you our long-term financial outlook. Bending the curve. We've got a lot of focused strategies to drive improvements in our core marketplace, and these have enabled Etsy to bend the curve on growth. Looking at this first chart, you can see we've had incredibly strong top-line performance. We did $3.9 billion in GMS last year, up from $3.3 billion in the year before.
That's over 20% growth on a constant currency basis, and that's actually 40% growth since 2016. When you look at it quarterly, you can see that we're not only growing, but we're accelerating our growth. On a currency neutral basis, last quarter was the fifth consecutive quarter of growth and the highest Q4 we've had since we've been a public company. Q2 of 2017 was an important turning point for us. That's when new management took their places and we reorganized to increase velocity, focusing on far fewer, more impactful initiatives to re-accelerate our growth. We focused on improving the Etsy platform for sellers and buyers by rationalizing all of our initiatives down to the fewest things that had the highest value, and all of them focused on etsy.com's core marketplace.
Many of these were customer-facing product and marketing initiatives. These were to drive visit growth, conversion rate growth in frequency, and increases in average order value. Let's take a look at some of those. Here you see we've had robust buyer growth, as Josh has said, since 2015, which has grown an average of 18%. Active buyers have grown to over 2.1 million, which is an average of 11%. Our ratio of buyer to seller has also been improving. Today, we're at 19:1, up from 17.5:1 in 2015. This is still quite low. We feel we are not supply constrained. We have plenty of supply to support much more demand, and a lot of our initiatives are focused on growing demand because we believe that's what sellers really want, are more buyers.
We said before that 60% of our buyers only come to Etsy one time a year. As Josh put that calendar out there, you can see that there's at least something special happening at least one time a month. I get frustrated a little bit looking at that chart because I can think of even more things than we were able to fit on that one page. There's many more occasions to come to Etsy. We've really been pushing hard on repeat frequency, and we're starting to see the green shoots of growth there. Here's one way to look at it. GMS per active buyer has been trending up the past five quarters. Here we're showing that it's up 2.2% in Q4, which is the fifth quarter of accelerating growth in GMS per active buyer.
We define an active buyer as somebody who's come to Etsy one time in the last 12 months. We also give you a number on our repeat buyers. Over 80% of our buyers are repeat buyers, but that's somebody who has come to Etsy one time ever, meaning they might not have been here in three years. Our focus is really getting our buyers to come back more often, and the data shown here suggests we're headed in the right direction. Here's another way to look at it. This is showing you 3 categories of buyers. Josh talked about the habitual buyers. Those are buyers that come to us six or more times in 12 months and spend over $200 a year. They are actually the fastest growing segment we have. They grew 22% last year. The next layer down are multiple buyers.
They come between two and six times in a year, and they were growing nearly 21% last year. The last category are the buyers that come one or more times per year. Essentially, when you talk about the search pros that Josh talked about, those are buyers that are finding workarounds to find the best product. Those are our habitual buyers who've made tricks to get in there and figure out how to find this stuff. Imagine if our product gets better and better, and search gets easier and easier, how many more of the green multiple buyers we can convert to habitual buyers, and how many of the one time a year people can convert and become high potential to become habitual or multiple purchase buyers. One more way to look at it.
We're looking at improvements in our retention rate, and we're examining our cohort curves. Like most companies, there is decay in cohorts over time. The highest retention rates are for the oldest cohorts, and they each have decreased a little bit each year. You can see here, the oldest cohort is 2014, shown as the orange line, and the more recent cohort is 2017, shown in the green line. What we're seeing now, and what we're really pleased about, is that the retention rate for the most recent cohort is at about the level of the cohort from the year before. We're starting to see these trend up, as you can see, for each of the cohorts in 2017, 2016, and 2015. I've just shown you three different ways that are all telling the same story.
GMS for trailing 12-month active buyers are accelerating every quarter. Habitual and multiple purchase buyers are growing faster than overall buyers. Our retention rates are improving and trending up. We often get asked questions about and get compared to other companies, particularly traditional e-commerce players and companies with SaaS models that provide tools and services but do not have a marketplace. Etsy's model is a robust two-sided marketplace, and I've just shown you a bunch of data about the growth on the top line. Another reason we love our model is that it isn't capital intensive, and we have a high flow-through of incremental revenue to bottom line, enabling us to deliver profitable growth. Traditional e-commerce, on the other hand, has lower margins. It typically requires a lot of infrastructure, and companies have to plan inventory and make guesses on trends and design.
The goods have to be sourced and manufactured. There is warehousing, picking, and packing, and of course, the logistics associated with that. Those logistics include taking receipt of the returns and often bearing the cost of the return shipping. All that inventory is carried on the balance sheet. All those wrong guesses mean that they have to be marked down and sold at reduced margins or jobbed out. While there's some attractive features of a SaaS business model, there are some disadvantages too. For instance, there's very low barriers to entry, and that leads to many providers all competing on price and promotion, which drives ARPU down. Downward price pressure drives churn. In a SaaS model, churn directly leads to loss of revenue.
Unlike in a marketplace model like ours, in which we have significant supply to fill the demand, even if a seller leaves the platform. In a marketplace model, if you lose the seller, you rarely lose the sale. We love our model. Sellers create. We build the best experiences so buyers can find their way to the product. Our two million sellers share in the marketing and promotion of their shops and inventory. The seller handles the order management, and we support them with customer care that we have been investing in to ensure it is world-class. There is so much to like about a two-sided marketplace model. They get better as they get bigger, as Josh has said. They scale. We have over two million sellers who help us market the products for sale in our marketplace.
That's nearly the same size as Walmart's employee base, but we don't carry any of those operating costs on our P&L. Our sellers handle all aspects of listing their inventory and distributing it. We don't carry the inventory. We're not supply constrained. We don't own the fulfillment apparatus, and our business model is therefore not capital intensive. As I said, we have very high flow-through of incremental revenue, and all of this results in a financial model that we love. We also have a lot of levers of growth. Since 2015, our take rate has improved every year. We aim to provide a fair exchange of value in which our goals are aligned with our sellers. Let me give you two examples. One is Etsy Payments, which we launched in 2013. For one thing, we saw that Etsy Payments was giving a much more consistent experience for buyers.
Buyers really liked it. The second thing, as a result, we saw that sellers who were using Etsy Payments had a higher conversion rate on checkout. In 2017, we mandated Etsy Payments. To be an Etsy seller, you need to use Etsy Payments platform. In that case, we feel it was a fair exchange of value. We make money from the use of Etsy Payments, but their conversion rate is going to increase. Another example is the pricing change we implemented last summer. In that case, a couple of things happened. The pricing change increased the rate on transaction fee from 3.5% to 5%, but they only pay a transaction fee if they've made a sale. If they win, they make money. If they make money, so do we. Our values are very aligned there. Secondly, it really changed our business model.
It created more margin for us. With that increased margin, it enables us to invest more back in the marketplace. One thing we committed to when we made that pricing increase was to invest more in marketing. We were enabled by a higher LTV from that increased pricing, and we could continue to invest in marketing at the same or higher ROI. Let's look at marketing for a minute. Our performance marketing channels have become a much more meaningful share of our GMS in the past few years as we've started to scale our investment. Last year, GMS from paid marketing channels was over $680 million, which was actually a 44% CAGR from 2015. We actually spent $112 million last year, which was a 59% increase. Now, that large increase, in part, was driven by our decision to reinvest in marketing associated with our pricing change.
Nevertheless, we think we're still in very early innings of realizing the potential of marketing for Etsy. Our investments, as Josh pointed out, are focused primarily in a very small number of low-funnel Google channels, almost entirely products like SEM and PLAs. We like those products because they're very easy to measure. We can clearly see the traffic coming from them and the conversion rate coming from them, and we can measure the ROI. We're continuing to evolve and improve our marketing measurement and discipline. While we've always kept a close eye on the return of our portfolio, we are optimizing to better target high-value buyers so that we neither overspend nor underspend to acquire buyers. For example, some categories have higher GMS, like furniture and jewelry. Some channels are more incremental than others.
We might be attracting buyers who might otherwise never have shopped on Etsy, but meanwhile, we might have acquired buyers that are cannibalizing GMS from returning purchasers. Today, we're not actively segmenting and targeting those subcategories in a refined way. One other key point to keep in mind here is that over 80% of our traffic still comes to us for free. Half of that is coming to us directly by typing in www.etsy.com. The other half is coming from organic search results and shops and listings that sellers or buyers might be posting on social media. A lot of our traffic is coming to us on their own, and I'm just talking about the piece that is an opportunity for us to apply dollars to spend to acquire new buyers coming to Etsy.
To just drill down on it a little, there are four ways in which we intend to continue to optimize. The first is on LTV. Josh talked about three levers of growing LTV. That's active buyers, frequency, and increasing average order value. I'll add one more, which is take rate. All four of those ways are ways in which we can grow LTV, and as LTV increases, we're able to spend more dollars at the same ROI, or we're actually able to increase the ROI. Second, we're working on targeting and re-targeting. As I said, all buyers are not created equal, and they don't all have the same LTV, and we're not doing a very refined job of segmenting those buyers and those categories that those buyers are in. Third, we're testing and optimizing. Josh showed you that lovely television campaign.
That actually started with two separate campaigns. We tested each in three local markets. We picked the winner, and then from the winner, the one that he showed you, there were multiple variations of that same campaign with different buyers and different products featured. From that, as we ran the ad nationally in November and December, we were able to pick the winner of those variations and put that into more of the rotation. We were also able to see which cable networks were performing the best and put the highest-performing cable networks more into the rotation, and we improved the ROI over time once that campaign was running. We're taking all of those learnings from Q4, and we're creating a new campaign that will begin airing in Q2.
We're very excited about continuing to improve the ROI on television and get that channel to be very ROI positive for us as well. Lastly, on our mature channels, like brand SEM and PLAs, we require not only a positive ROI on that portfolio of spend, but on every marginal dollar of spend, we require that to be ROI positive as well. That's what we call MROI. We're continuing to refine and move more of our less mature channels into that model as well. Flashing forward, where you see back in 2016, we were spending only on these Google products, very low funnel. In 2018, we started to throw some television and digital video into the mix.
Going forward, we think that upper funnel initiatives, as we get them to work in an ROI positive way for us, will be much more a significant part of the overall marketing spend. We also spend on product investments just like we spend on a dollar of marketing. In 2019, we expect product development employees to comprise over almost half of our entire employee population. When I say product development employees, I'm including the squad. That's product engineering, product managers, product designers, data science, and product analytics. As Kruti said, we're evaluating how much incremental GMS they can unlock. We're focused on increasing the wins, increasing the velocity of things we get into experiment, and increasing the size of the win. They don't all succeed. We have some fails, we have some wins.
Much like a film studio, we have product that ends up on the cutting room floor. The metric I am showing here includes the total incremental GMS with a total development investment in the denominator, not just the product development investment that wins. You can see that this investment is getting more and more productive. It grew significantly in 2018, and we expect it to continue to grow again in 2019. Putting that all together, this has directly impacted our bottom line, with adjusted EBITDA hitting 23.1% for the full year 2018. Free cash flow generation has increased meaningfully. Last year, we generated $178 million of free cash flow, and at the end of 2018, we had cash and cash equivalents of about $624 million. What do we do with all that cash? We invest for growth and value creation.
First, we invest organically, and that is a lot of what Kruti and Mike were talking about. We invest in marketing and product development. We also invest for efficiencies in platforms that help us scale. Google Cloud would be a good example of that. Secondly, we evaluate third-party opportunities, and we use a very disciplined approach to thinking about M&A. For example, we could invest in new geographies, which we did last summer when we did a transaction with DaWanda, the German marketplace. We could invest in adjacent marketplaces. These are marketplaces that have similar business model characteristics to us but might have different products and different buyer segments in completely different ways. We could invest in new technologies. Blackbird is a good example of a technology we invested in several years ago to strengthen our machine learning capabilities.
We could invest in vertical extensions so that when you come to Etsy for wedding, you are getting the full experience. If you are coming to Etsy for jewelry, you can search by cut and clarity and size. Those are a number of ways that we might think about M&A. The third category of use of cash is return of capital scenarios. In 2018, we repurchased about $135 million of our stock. That is about 5 million shares, at an average price of $30.28, which I think is a pretty good return on investment. All of this is why we believe we are in early innings of Etsy's growth opportunity. We know many of you ask us about the TAM, and here is the TAM work we did in 2017.
We looked at our top six categories, which were clothing and accessories, home and living, jewelry, craft supplies, art and collectibles, and party goods. We looked at our top six geographies, which is the U.S., the U.K., Canada, Australia, Germany, and France. In those top six and top six, it resulted in a $1.3 trillion TAM. We took just the online portion of that, and that is about $155 billion core market TAM opportunity. In 2017, with the GMS, we had approximately 2% share of the $155 billion opportunity. We did more work to update that TAM. First of all, the market grew. Secondly, we added all our relevant categories, not just the top six. We added things like beauty and personal care, toys and games, and collectibles.
That's a $1.7 trillion TAM and $249 billion if you just take the online component of it. In 2018, we had $3.9 billion in GMS. That means we're still less than 2% market share of that opportunity. Still huge opportunity. We're a tiny size of it. $249 billion is a really big number. I get a lot of questions from all of you, sometimes skeptical. Are all consumers in this segment really in the market for something special? We did some research to test that question. Last June, we did market research to answer this question. We surveyed 3,500 people, and we found that 40% of them are both likely or very likely to be shopping for something special or unique and are classified as expressives or seekers. What are expressives and seekers?
An expressive is somebody who takes time to find an item that is just right and expresses herself that way through all aspects of her life. A seeker is somebody who loves shopping and enjoys the hunt and searches for a good deal and is persistent and perseveres to find just that perfect item she had in mind. Applying that 40% to the TAM, focuses the market size down to $100 billion. This is a much smaller, more defined, and targeted market opportunity, but still, we're just a tiny fraction at about 4% of that opportunity. Taking this math forward, we expect these markets to continue to grow, especially as offline moves to online. By 2023, that smaller, more targeted size will grow to $170 billion.
To reiterate what Josh said earlier this morning, any way you slice it, we believe Etsy has a very large market opportunity, and demand is not something we're worried about. Where does all this add up? Many of you have asked us to provide a view on our long-term growth for Etsy, and today we're going to put a stake in the ground. We made you wait for it, though. Etsy has been growing faster than the overall market. In 2018, we grew GMS 20.8%, compared to 16.9% growth in online sales for our top 6 categories. Over the next five years, we expect to continue to grow faster than the overall market.
Active buyer growth has averaged 18% over the past four years, and frequency has been improving, as I showed you, in three different ways, with the habitual buyer segment growing fastest, GMS for trailing 12 months accelerating, and retention rates by cohorts improving. We expect GMS to grow at approximately 16%-20% over the next five years. Like any company, we're going to ride the tide of favorable and unfavorable macroeconomics. When we have robust consumer confidence, we could be above this range. When conditions are less favorable, we could be below this range. We've used external data from our TAM analysis and significant internal work on our strategy to land on the number 16%-20%, and we believe, on average and over time, this gets us to about $9 billion at the midpoint by 2023.
For revenue, we expect that growth will be slightly faster than GMS over the next 5 years. Let's reflect on the 16%-20% for a moment. Our analysis of the addressable market shows that it will grow approximately 12 points in the next 5 years. This shows that we will grow approximately 4-8 points faster than the overall market during this time period, consistent with how we performed last year. Looking at that on a share basis, we believe that we will still only be about 5% of the total market share by 2023, up from 4% in 2018, using that smaller size market that's defined for special. We think we can continue to scale. We think about our overall cost base in 3 categories. First, those that directly vary with revenue, like performance marketing and credit card transaction fees.
Here, we'll continue to find efficiencies like we did when we implemented the single ledger initiative. By and large, we would expect that these are going to grow in line with revenue growth. Second, we have costs that are semi-variable. This would include product development resources and member support. It isn't a one-to-one ratio, but we directly measure the top-line return from the dollars we invest there. We would expect this category of costs to grow slightly slower than revenue. Last, all of the support and back-office costs, which would be inclusive of G&A as well as cloud expenses, people and workplace costs, and other management overhead. We are driving for efficiencies here, and we expect these costs to grow at a slower rate than revenue and to be a contributor to margin expansion.
Putting all that together, we expect that the margins will be at or above 30% in the next 5 years. Just in case you're wondering why there's two different colors of blue on that bar, that's just because it's a range. It's going to be 30% or higher by 2023. 2018 was a solid year for us. We did a lot of things that made us a little bit better, and we're going to continue to do that in 2019. This is the guidance we gave for 2019 GMS in the range of 17%-20%. Let me take a minute to reiterate the notes that we called out on our call just two weeks ago about Q1. We said we had an elevated period of marketing spend during Q4, and in Q1, we are pulling back from that a little bit.
I like the term we're resting and testing. We want to really understand what we learn from our television test. We're not running television right now in Q1, and we're looking at some of the incrementality of the spending that we did on our more mature marketing channels. We pulled back from marketing a little bit, and we also said we saw a small bit of macro. We're still very comfortable with the 17%-20% GMS range that we gave you. With that said, GMS could be between the midpoint and the low end of the full-year range, driven by factors we chose to do ourselves and the macro factors that we discussed. With that said, for the first quarter. With that said, let me just make sure that's very clear. I'm talking about the first quarter only.
The guidance for the full year is 17%-20%, that's GMS guidance. We've also put out guidance for EBITDA. Pullback on marketing means less expenses in Q1. In Q1, EBITDA actually could be toward the high end of the range. Think about those factors, and that's the guidance we're giving you about 2019, which we're very excited about. Then, as I said, our five-year targets are 16%-20% on average and over time for our five-year time horizon. We think we're going to grow faster than the overall market. We'll grow slightly faster than GMS for revenue, and then margins of about 30% or higher. To sum it all up, the key headlines for today and everything that you've heard from all of my colleagues, we are an early-stage opportunity in a large and expanding market.
We have a unique and defensible right to win. There are numerous untapped opportunities in product development and marketing. We consider it still very early innings on this very ambitious roadmap we have. We have a world-class technology and engineering team to fuel the growth and a disciplined leadership team driving a high-performance culture, which all results in a significant runway for sustainable long-term growth. Thank you. I will now invite the entire management team back up on the stage, and we will open it up for questions.
Thanks, everyone. While we get everyone settled up here, we have till about 12:30 P.M. for questions. I'll let it go probably a little bit longer than that, but not by much more because we want to make sure everyone sticks around for lunch. When we come to give you a microphone, there'll be three of us. I'll be in the middle, Gabe will be on the right, and Mahmood will be over on my left. Please make sure you state your name clearly and your company, so we have it for the webcast. Also, if you could just limit to one question. If we end up getting to as many people as we can, we'll be happy to come back and get another one from you. Okay, first question over here.
Hi, Scott Devitt with Stifel. Question relates to GMS growth, the five-year forecast that you have in place. The vast majority during the turnaround period of growth has come through buyer growth. Frequency's increased slightly, but it's mostly been buyer growth. I'm interested in terms of the inputs in that five-year GMS growth, how you think about buyer growth relative to frequency relative to AOV. Thanks.
I'm happy to start. The truth is, we can get there multiple ways, and that's part of what makes me feel really good about it. I believe that we have substantial runway to grow the number of active buyers. I believe we have substantial runway to grow frequency, and I believe we have substantial runway to grow AOV. We're putting investments against all three of those levers, and I believe that all three of them will bear fruit. Relative to each other, how much fruit they're going to bear is something we're going to test and learn our way into. You don't need to believe that we knock it out of the park on all three to believe the guidance.
We can get there in multiple ways. We just need to make some progress against each of those three in order to get to the guidance that we've laid out.
Hi, thanks. This is Ed Yruma from KeyBanc. One of the positive surprise in the fourth quarter was the growth in international. That was obviously very nice. It created the GMS growth. I guess, as you think about international, talk about the role of DaWanda. Are there markets you're particularly interested in, and do you think that this will continue to power growth? Thank you.
Yeah. I'd start by saying that Etsy is a global business by nature. We've talked about our six core markets, but I want to say that there's sales and purchases happening in almost every market around the world.
The day a seller lists her item, it's automatically machine translated into 8 languages, 9 if you count the U.K. English is different from U.S. English.
My friends across the pond will have a point of view on that. It's automatically machine translated, everything is global. We're seeing robust growth in our international markets, which is terrific. One of the great things about a marketplace is it just makes us better. Part of what makes Etsy such a cool shopping experience is that you can go buy amazing fabrics from Lithuania, for example, was one of the sellers I think we had highlighted. That's part of the strength. Now, we've talked about our 6 core markets, what we mean by the 6 core markets is that's where we're trying to build two-sided domestic vibrancy. Everything I talked about in a two-sided marketplace being hard is also true in geographies.
The day you open up in France and a seller shows up, there's no buyers, the day you show up as a buyer, there's nothing for sale. One of the great advantages we have as a global marketplace is as we operate in France, from day one, a seller coming in France had her listings machine translated and available to be bought around the world, so we can start to develop supply. The day a French buyer shows up, she sees a lot of abundance from all around the world and actually has something to buy. Now, over time, as we bring more and more French supply and more and more French demand, we knit those 2 together, and we can actually create a domestic business.
What's exciting about that is while it's cool and interesting to buy things from around the world, when you're trying to build frequency, domestic actually matters. On average, more people prefer to buy locally when they can. What we now have seen is that both the U.K. and Germany both have more than 50% of sales be domestic. We think that's a really exciting milestone that can really help to fuel the growth. The DaWanda deal that we did, partnership agreement that we did, definitely helped to bring more abundance in Germany and helped to bring more buyers in Germany and got us to that 50% mark, which is exciting. We are disproportionately investing in France and in Canada and in Australia to help them to get there as well. That's what we mean by core market.
I do want to point out, we still have a lot of activity happening in Israel and in Turkey and in Eastern Europe and in lots of other parts of the world. I'll take one other second on international to talk about India. We also are making some investments in India. Why India? First we saw that when a buyer buys something from an Indian seller, their lifetime value goes up. It appears that Indian sellers are bringing unique value to the marketplace that buyers are really delighted by. No surprise, India has a great history of artisanship and craftsmanship. It's also true that in the developing world, you can actually afford to pay people on mopeds to go village to village and educate sellers, which is exactly what we're doing in India.
We have a small team of very talented people that are going to different workshops in different villages and teaching them how to get up and running on Etsy. We're trying to learn how to scale that because that might be something that actually could be expansible to other parts of the world as well, not just India. It also, by the way, allows us to vet the selling practices in those markets to make sure that the sellers are consistent with the values that we believe are very important for our marketplace, particularly around things like workplace safety. We're excited about the runway for international and the opportunity it presents.
Hey, Josh. Tom Forte from Davidson. Rachel had a lot of excellent charts, and she showed a prolonged period of high sales growth with margin expansion, which means a ton of free cash flow. How should we think about your intent to use the free cash flow, either for additional buybacks or potentially M&A? Thank you.
We, am I on? We showed you three ways in which we plan to use our cash. One is to continue to invest it organically, as we have been doing in marketing and product development, lots of things that we think can help us scale. Google Cloud is a good example. We'll be doing a lot of things like that, like investment in Zendesk to make our member support group world-class and to make them efficient and automate as much as we can. There's things that need to happen in the finance and HR worlds to make those more efficient. Those are good investments in organic opportunities. There's certainly, we think about M&A. We always look at build by partner.
The examples we've got to point at now are more tuck-in. We look at things large and small and ask ourselves, "What's the rationale for investing in this thing, and can we do it faster by buying it, or can we build it ourselves?" There's a long list of things that we look at all the time, including commercial partnerships, by the way. They don't all have to be acquisition ideas. Then there's return of capital. We've done a nice buyback in the last year. We think that's a very high return on investment. There may be a time where we think there's something that has higher return on investment, so we would pause that and look to do other things. We look at the whole size of the opportunity, and we use a very disciplined approach to evaluating our choices there.
Thank you.
Great. Thanks. Kunal Madhukar, Deutsche Bank. Maybe a multi-part question related to the expressives and the seekers. For this 40% of the population, one, what are they buying right now if they're not buying on Etsy? Second is, how do you get them to buy on Etsy? What is it that you need to change? The third is, when you showed that concave curve, how was that concave curve different for these guys, or was that concave curve different for these guys versus the rest of the population? Thanks.
Great questions. I'll start by saying we don't have all the answers to all of that. We are just diving in now, and we're a 13-year-old company. That may surprise you, but it excites me because where we are as a marketing organization is less mature than 13 years. I can tell you that our habitual buyers fall into two camps. Many of them are also single category, so they love us for home furnishings. I've met some habitual buyers who are all about vintage couture. When you ask where are they shopping, they're in couture shops. Actually, there's a surprising amount of couture activity on Etsy, not what you think of when you hear the word crafty. It's a big category for us. It really depends on the category of interest. Some of them also shop all across Etsy for all kinds of things.
They're really the ones who've really unlocked the magic of Etsy across the way. Some of the highlights, when we showed some of the TAM, and we talked about Wayfair and ASOS and Blue Nile, these are examples of the kinds of places where you would see our shoppers shop. What does their curve look like? Well, obviously their intent to purchase is much, much higher. Most of our habitual shoppers are on our site every single month, and actually a pretty high percentage of them are on the site every single week. It turns out that for a fair number of Etsy shoppers, even though they might only buy once a year, they're actually on the site once a month. Our opportunity is there. We've just got to do a better job surfacing the right inventory to drive conversion.
Some of the things we talk about around fulfillment, too, they're there, do they have the confidence it's going to arrive on time, do they feel good about the shipping price, those kinds of things are also important to kind of get them over the line. The third part of your question was?
Yeah, no, it was basically just about the concave curve and what you really need to do.
Great.
Thanks.
Thank you.
Great. Heath Terry, Goldman Sachs. Josh, Rachel, as you looked at the longer-term guidance, that 30% plus EBITDA margin that you're targeting, you guys have shown that you've got the ability to invest incremental profitability into faster growth while driving returns at it. Is there something particular about that number that you chose to target it versus investing more of that profitability back into the business to grow above the high end of that 20% growth that you put up for GMS?
There's no magic to 30%. It's what other marketplaces do, and I guess what I would take from that is there's no reason why we shouldn't be able to show the kinds of margins you see in other scaled marketplace business. To your point, we're a growth company. We're investing for growth, and when we see an opportunity to drive profitable growth, we will take it. When I talked about the marketing philosophy and that we don't hold proven channels to a budget, we hold them to an ROI threshold, that's the kind of discipline that we apply. Margins might go up and down. We're not trying to manage margins to a number. We're trying to drive profitable growth. Several of you have asked me, could you grow faster with lower margins, and incontrovertibly, 100% the answer is yes.
Absolutely, we could grow faster with lower margins. That's not really the question. The question is, could we have profitable growth with lower margins? We could make 2019 a bigger top line with lower margins, if it wasn't profitable growth, we've got to grow over that in 2020, then we've got to grow over it in 2021. By the way, we're all shareholders of this company, and I don't think any of us like to spend money that we don't think is delivering a good return. We're a growth company. We're very much a growth company, we're a disciplined growth company. For example, I've seen a lot of marketing organizations look at the portfolio return on investment as an investment, across all of the money we spent in marketing, on average, it delivered this. Well, that's not really the question.
The question is the last dollar you spent, how good was the return there? I think having that kind of sharp discipline will allow us to continue to drive growth sustainably into the future and do it with a good margin structure. Eventually, the margin structure will be how much can we do to unlock more and more high-potential growth channels. I'm optimistic that we can continue to find really good ways to continue to invest, and that'll show up as we do that in the top line, then the bottom line takes care of itself.
Can I just add one quick thing to that? We also said that revenue will grow slightly faster than GMS during that time frame, that growth is going to come from further adoption of our existing services, maybe a new service or two. We're not sure, that's one point, it doesn't necessarily assume that there's substantial increase in our take rate. There may not be substantial increase in our take rate, there could be if we think there's an opportunity to align our value, there's a value exchange with the sellers. What you're looking at is a model that resulted in a 30% or higher margin based on what we know today.
Obviously, there's a lot of cash that will be thrown out of that model, those investments that we make with that cash could end up investing in things that contribute even more to the bottom line than what we're currently anticipating.
Thanks.
Shweta.
Shweta Khajuria, RBC Capital Markets. How do you think about opportunity in Promoted Listings? More inventory, better product and algorithm. Are you at a point where you have sort of benefited already, where's the upside on that? Second is on product. You've been working on search and discovery for quite some time. We've seen benefits of it. Where are you right now? How much upside is there? Is it in relevance, in better personalization, or better quality of image recognition? Where do you think there's most opportunity, where have you seen most results?
Those are both great for Kruti and Mike. You want to start with Promoted Listings?
Sure. I think we've talked about this. We see continued opportunity in Promoted Listings. There are really three levers there. How many sellers are using Promoted Listings, the inventory that we have, and how well those ads are working. Effectiveness and in turn, budget utilization. We've made a lot of investment over the last couple of years in expanding the inventory for our Promoted Listings. All sellers aren't created equal, Promoted Listings work better for our better sellers. What we're really focused on going forward is improving the return and improving the budget utilization for ProList.
Can I add on ProList before you get to the search? The same thing with ProList, we've just started to apply our machine learning models to that, and we saw incremental revenue because of it, and I think it's still very early days with that.
Yeah.
I think that's the other lever that we absolutely plan on continuing to work with.
The second question was search generally, and what inning are you in of search?
Yeah. I'll start, and maybe you can take it from there. We're still really early. We have invested a lot in search over the last couple of years. We talked about context-specific ranking and how that's really benefited us. Again, understanding what it is that you're looking for and showing you just those things versus all of the other items that are related that have those same search terms in them. Where we see continued opportunity is by improving what we show you in terms of both quality and personalization of those search results, and we're really early in that, and there's a lot more work that we have to do. We think we can leverage our machine learning models and our data science team to be able to do that.
Yeah. The way I broke down search today was that there's this basic search, there's re-ranking, and then there's personalization, and I think we're very early innings on all of those. As I mentioned, in basic search, we're still only using the listing's label and the tags and then predicted conversion rate. There's many more things for just even the basic search. The re-ranking, we're already using machine learning, but we can continue to improve those models over time. Then personalization, we're using location. There's so many more other factors that we could use on personalization. Early innings on all three of those with search.
Brian Nowak, Morgan Stanley. I have two. Just appreciate all the color on the paid versus organic traffic in GMS. As you laid out the GMS forecast, I know it's very early in marketing, how are you thinking about that changing in the next five years? What % of the business do you think eventually becomes paid or maybe contribution from performance channels that you talked about? Then the second one on the search, understanding it's very early there, but can you just sort of talk to what data sets you capture from your buyers now, what % are logged in? To what extent is data something you still have to work on to really improve the overall personalization of search?
Great. Maybe I'll take the first, and Mike, you want to take the second? In terms of what % of customers or traffic is going to be paid versus not in the future, again, I think we've got multiple levers to get to the 16%-20% GMS growth that we showed you, but there's a real virtuous cycle. What you've seen happen over the past two years is primarily we've made conversion rate improve. That's the metric we've moved the most. When conversion rate goes up, the value of a visit goes up. The second thing we did is we did change take rate. That also made the value of a visit go up. Those two things allow us to spend more to get a customer to come in.
Frequency is the other lever where we're seeing early green shoots, and Rachel showed you three different data points to support the fact that we are in fact seeing frequency improve, although plenty of runway to do better. You're really cooking with gas because now your lifetime value has really expanded very significantly. I think there's a real virtuous cycle between the operational improvements that we make and our ability to invest for growth. Ultimately, it's pretty simple. If a customer comes and has a good experience, they come back more often. Bless you. They come back more often, and their lifetime value goes up. Everything that Mike and Kruti talked about around getting a better product search experience, getting more trust and safety, all of these things actually build more customer loyalty, which drives our lifetime value.
The other thing that's going to muddy the waters a little bit is we talked a lot about investing more in higher funnel. It's just harder for us to attribute. The world of digital video is actually fast evolving, and I'd love for us to get to a world where we know that someone is in the market for a wedding, and we're able to specifically target wedding-related content to them, high funnel, mid funnel, and low funnel. Once they've had a great experience with that, we understand the next life event and the next occasion, and we can move with them, and the world of video is moving in such a way that even in digital video, even in things like TV ads, you could start to talk in a more segmented way to your customer base.
It will be harder to track on a kind of one-to-one basis how much is coming from paid. Again, I feel like there's a big opportunity in organic for us to do better. By the way, even things like SEO, we don't talk a lot about that, but there's a big opportunity for us to do better in SEO. We have a terrific relationship with Pinterest, and many of our sellers promote on Pinterest, and yet we don't do a great job of converting that traffic. That's a big opportunity for us. Our sellers, we could give them more to empower them to promote their listings and their shop on their own. That shows up as free traffic. There's a lot of opportunity for us to really drive the organic or the unpaid as well as driving the paid.
On the search data, I think I'd start by saying that we'd love the buyers to come to us and tell us their style and what they're looking for and exactly, and help us with that, and maybe that's a path we go down. We already have an awful lot of data that, whether you're signed in or not, we have other buyers who come to the site and look for similar or maybe the exact same items. So we don't necessarily need tons and tons of data about an individual buyer because we have massive amounts of data about all the buyers, and we can use that to predict what you will be interested in as well. If you look at machine learning and sort of where the state of the art is today, it's not a new concept.
It's been around for many, many, many decades. What is enabling it today is the data and the processing power. We have the data. We have lots of data. In fact, we only use a couple of weeks' worth in a lot of our models because of the processing power. Getting to something like Google Cloud allows us to unleash the processing power. Now you combine the data that we have with this massive amount of processing power, and it really is exciting to see what we can do with that.
Bradford Seagraves with Davenport. I was curious on the product and the technology side, you mentioned a couple examples, such as on the search, doing a curated sort of search list, as well as when you buy this, other people have bought these few items. When would you expect to roll those examples out?
The way to think about our product development process is that we're really focused on customers' problems that we're trying to solve, and then we take this process of rapid experimentation and iteration to get to the ultimate endpoint. What we know is that we want to provide better recommendations to help our buyers make a purchase, find what they're looking for. We know that we want their search results to feel more personalized and curated. What we're going to do is try a number of different ways to get there until we get to the solution that we think is ultimately the best for them. What you'll see is ongoing improvements across all of these areas.
The way that we organize that at the top of the house is we pick at the beginning of a year a finite set of customer problems that we want to solve, and we set a target on how much incremental GMS we're going to unlock. We assign squads to a customer problem and a target. One customer problem might be expressed as, it's really hard for me to continue my journey or to pick up on a new mission. I experience lots of dead ends at Etsy. That's a customer problem, right? We talked a little bit about, I go to one listing, and it's a dead end. You give that problem to a squad. It's got designers, analysts, and engineers.
They're really smart, they know that they have to unlock this much GMS by solving this problem over this date. They test a bunch of different things, we tend not to dictate, like the way to do it is you have to show people who buy this also buy that. They spend a lot of time on the web. They do benchmark. We do like to learn from other people, how do other people solve this problem, we don't dictate to them. We say, "Solve the problem of dead ends that customers can have satisfying and complete journeys, and do it in such a way that you unlock X amount of GMS per quarter." They try different ideas, and they're testing weekly.
One of the great things about our infrastructure is a designer comes up with an idea, we can rapid prototype it in front of a few customers in a workshop, then code it and get it up on the site. Within days, we know, did it perform better than the default? If it performs better than the default, it becomes the default, and we measure what lift it created. That becomes what we call banked GMS or incremental GMS, that's added to their scorecard. If it didn't work, no problem, roll it back. That's fine. Keep trying and keep learning. Squads have some weeks that go great and some weeks that don't go great. Some squads have a great year, and other squads don't have as good a year. We have a portfolio amongst them such that we get relatively consistent delivery.
Kruti and myself and the whole executive team can look every month at, per squad, how are they doing at delivering against the target in solving the customer problem and doing it in a way that unlocks GMS. It's that kind of discipline that both empowers the team. When Mike and Raina talked about people like being at Etsy because they feel like they're not a cog in the wheel, they actually make a big difference for customers. They're really empowered. When we talked about accountability and focus, it's also really clear to them what success looks like. That combination, I think, is working really well for us.
Hey, guys. Ygal Arounian from Wedbush. Kind of addressed both of the things I want to ask about partially, so I'll ask them a little bit differently. Rachel, you talked about take rate, just to be clear, in your forward guidance, there's not necessarily an increase in take rate that's being built into it. There might be, but there might not be. How do you see sellers' willingness to accept another increase in price if you're, like you did the first time, giving them kind of benefits, showing them the benefits of what you could do with that? We get a lot of investor questions around future price increases. Then on search and discovery, talked a lot about search within Etsy. I'm curious about discovery outside of the platform.
Things like image search or being able to use your phone, take a picture of something, and have that kind of populate searches within Etsy. Some of the key drivers, maybe social commerce and taggable posts on Instagram. How do you see those kind of driving traffic to the site going forward?
I can do a quick answer to the first question, which is, the answer is no. There's nothing baked into that forward forecast that's really inorganic. There's no explicit take rate increase assumed in our forward forecast. No explicit M&A assumed in that forward forecast. It's what we think we can do with the resources and how they'll grow each year.
The second question was about search and discovery and-
One of the things that Josh referred to a little bit earlier is the power of our sellers and their marketing and their promotion off Etsy. We find that that is one of the best ways to get our sellers' products in front of the right people. Leveraging all of those things that you're talking about and helping our sellers leverage those, is an important way to make sure that we're out there on all the channels that their buyers are out there. Last year, we introduced the first version of a social promotion tool that makes it easy for our sellers to promote their items or milestones that they've achieved on their social networks to promote their shops. As we think about helping our sellers do end-to-end marketing, it's really about tying all of that together.
Both the promotional free marketing as well as the paid marketing to understand how that all looks together for them, and to really leverage the power of 2 million active sellers in a way that we could never do on our own.
Actually building on that. We've talked a bit about the fact that sellers have a real appetite to promote their shops, as you would expect, both through things like their Saturday fairs. You go to a Saturday fair somewhere, you're very likely to encounter a lot of Etsy sellers, but also on places like Pinterest and Instagram and others. Many of them have an appetite to invest behind their growth. Promoted Listings is one way that they can spend money to become more prominent and invest. We've started to create channels where they can invest in growth off of Etsy. Google Shopping is where they give us a budget, and we will buy traffic on their behalf.
We have tools to be able to do that, in a way that's much easier than if they had to go buy those books that Kruti showed you and try to study up for themselves. Another part of the story that I don't think we've been great at telling is it's also more cost-effective when we do it for them. If you set up your own website and then you go to Google and you buy traffic from Google, you have a very low Quality Score, a Quality Score. You have to pay more to get prominent. When you're under the Etsy umbrella, there's a lot of power that comes with that actually makes it cheaper to buy a visit. I don't think we've even begun to tell that story yet to sellers.
Building the tools and the capabilities, to explain to sellers how we can help them leverage their marketing dollars more effectively than they could do on their own, is, I think, a big opportunity.
It is Laura Champine from Loop, and my question is for Mike, and it is a when and how question about personalized search. I think you mentioned in your presentation that you are only personalizing for location right now. When will you add factors? What factors will you prioritize? If random descriptions that sellers have autonomy to put in act as a gating factor, how will you address that issue?
I think like Kruti answered the question before, that we do not think about when or dictating what to the teams, that we empower the teams by giving them problems to solve and then goals and KPIs that would indicate when they have met success. I think that is the exact same thing we are doing here with search, is we have identified problems that we see the buyers experiencing on the marketplace, and that we are trying to solve for them this year and let the teams iterate.
That is actually the power of our process, is that we do not dictate. We do not have a person or a group of people who think they have to know everything. In fact, we know we do not. That when we have an idea about something that we think will work, we know it does not, some % of the time.
Because we experiment, we can do that very rapidly. Whether that experimentation is early prototyping or actual coding experiments, it is all about learning quickly. That allows us to make sure that we are actually pushing for the right thing, as opposed to just saying, "We are going to add this particular feature on this date," and not actually know if it works and it solves the real buyer's problems. That is the answer, is that we are iterating on this and continue to expect that it will continue to improve over time.
We test fast. It is not like things here take quarters and years to happen. Several people have commented on new management came in in May, within a matter of months, we were actually already starting to inflect the curve and how on earth is it possible. The prior leadership teams have done an amazing job of building an architecture for rapid iteration. Etsy is still a place where you can go from idea to test live and market in a couple of weeks. First, that makes it super fun. In terms of attracting great talent, there are not a lot of places at our scale where that continues to be true. It also means that longer-term product roadmaps that are five and six months out, unfortunately, it can feel a little unsatisfying, the answer is it depends.
We're going to test some features this week, and if they work, we're going to go much deeper into them. If they don't, no problem, we're going to go in a different direction. That's where learning to run an agile company is a bit of a learning for me, too, where I tend not to look at a product roadmap that has a specific feature that's going to be launched nine months from now. Instead, what we say is we're going to have to solve this customer problem, and we will know it because we will achieve the following KPI nine months from now. It's a little bit different way of managing, but it's fun for everyone when we get used to it.
Thank you, Deb. Mark Moore, from Thorncreek Capital. The question is actually for Kruti, and then if I may, a small clarification from Rachel on the financials. Kruti. There's been a ton of success when I think about experience on Etsy. We actually went to an Etsy first model at the house, and now have the world's greatest ping pong table, which is pretty sweet. One aspect that we've struggled with is the issue of when, which is to push past what and love to when, whether that be emotional purchase or rush, two-day type purchases. I wonder if you guys have a thought about today-only stuff, or can be delivered in two days, or some kind of a scroll news feed type of experience to be able to tease out that opportunity.
If I may add, I'll ask Rachel also, which is, Rachel, for the first time, you mentioned the GMS impact of product marketing spend, which, if I get it right, is a little over $100 million over the $650 million or whatever it is, which is pretty similar to the actual take rate. I'm just kind of wondering how you think about the IRR of that type of project.
Wondering if I'll get started.
Sure.
You touched on a lot of different things there. A little bit about shipping speed and a little bit about discovery. Starting with the first, this is one of the benefits of having such a broad variety of things on Etsy. We have millions of items available that are ready to ship, and millions of other items that are made to order. Depending on what your need is and what your intent is on Etsy, we're really confident that you can find it.
What our job is to make sure that we're surfacing the information that you need to know earlier on in your purchase journey, that you know if things are available and ready to ship, if that's what you need, a gift in a rush, or if it's something that a seller is able to customize from scratch, that you know if you have time, you can purchase that. We've already started making some improvements there. What you saw over the last holiday season ago is we started off showing filters that help you understand whether you can choose whether something's ready to ship or ready to ship in less than three days. There's a lot more that we need to do there, and it happens throughout the buyer journey.
We know that buyers have a variety of needs, and we want to make sure that we're able to serve those through the product experience because we know the inventory on the site already meets those needs. In terms of discovery, there's a lot of work that we're doing there as well. What you may have seen if you use our mobile app is that we've already introduced some new ways of discovering fresh new inventory on that platform. Again, we're using machine learning to really help us aggregate data sets about what is popular and what is popular for someone like you to make sure that we are showing you fresh, interesting, and inspiring content at the moments when you're looking for it.
This is another thing that's really important is sometimes you're in the market for something that you know you want. Sometimes you're looking to just be inspired and browse. We need to really quickly understand which of those paths to take you down so that we're showing you the right content for that particular session that you're with Etsy.
On the $682 million of GMS from paid marketing, we said we spent about $112 million on performance marketing in 2018. That's about $6 to $7 of GMS per dollar of spend. It's not bad. What Josh and I both talked a lot about was continuing to optimize from there. We have included in that number are things that are less mature. Paid social might be an example where we're working those up to be ROI positive, they aren't there yet. Another example would be better segmenting the PLA and SEM buys that we're spending money on because some of that buy is for very valuable buyers, they might be the habitual buyers, or very valuable categories that come with much higher GMS in the category. A button is not worth the same as a sofa purchaser.
That's some of the refinement we're doing is to continue to improve that ROI in that number you've done as a portfolio look, we look at it much more on a individual marginal look.
The other thing I'd add is it's an average. Averages can be very confusing. GMS per dollar spent is one way of thinking about things, of course, there's an annuity stream, right? You've either brought in a new customer who has an annuity stream, or you've reengaged an existing customer. Those two have different values to them. By the way, not all customers are created equal. Did you bring in someone who looks like a habitual buyer into a very high-value category, or someone who's a super pragmatist into a low-value category? What we are really working on is getting ever more nuanced about thinking about CAC to LTV, getting really smart about that.
I'd add that I really appreciate that you've moved your whole family to an Etsy first model, and I encourage the rest of you to do so. I look forward to comparing recent purchases.
We have one over here.
Ron Buchbinder with IFS Securities. You have been very focused on ROI. How do you balance existing initiatives, like in search and marketing, versus new initiatives maybe like seller financing or augmented reality that you really wouldn't know the current ROI on?
I'm going to start. We start with customer need. There's an endless number of things that could be good ideas. If the bar is it a good idea, we're dead in the water. We start with is, what does success look like over a certain time frame, let's say three years? Then we ask ourselves, what are the fewest things we need to do well to achieve that goal? If we can fall in love with those few things where we're going to really move the needle and get really passionate about them, it becomes much easier to say no to the very long list of things that are truly good ideas. That's part of the very painful process we went through last spring and summer. It's not last spring and summer anymore. Now we're talking 2017. It's been almost two years.
We weren't working on bad ideas. We were just working on too many good ideas. Having the discipline to say what are the fewest things we need to do really well, and then doing those things with excellence has really helped us to raise our game.
The other thing that I'd add is we really think about our product development as a portfolio. For every single one of these initiatives, or these customer problems that we're trying to solve. We talk a lot about what are the bigger bets that we need to make, what are the foundational items that we need to build, that won't necessarily have an immediate GMS or financial impact, and what are the things that we have the most confidence in are going to move the ball forward in a short period of time. We look at those holistically. It's really important to make sure that we're moving the ball forward while we're working on some of these bigger ideas that aren't going to bear fruit right away.
I think Josh has introduced this framework that we call foundational, evolutionary, and revolutionary ideas. We look at that at the top level across the entire portfolio, but actually within each problem space that we're looking to solve, we're doing that as well.
We have time for one more. I saw someone wants a second, unless there's any other takers, I'm going to let, Mark, you said you want a second?
Yes. Sorry. Thank you. I guess the question is for Josh. Josh, I suppose more so in Europe than here in the U.S., the issue of privacy and data, and I'm wondering where you guys have come out. I think the most acute opportunity is probably the purchasing of third-party data to be able to match up to your own databases, to be able to accelerate the personalization aspect. Just how Etsy's going to approach this issue and as you guys grow. Thanks.
First, it's such an important issue, and I really appreciate you raising it. It's something that we spend a lot of time on and think a lot about inside the company. A couple things I would say. One, is it's something we've taken seriously for a very long time and we've cared about for a very long time. It's not new to us. Two, our business model is very privacy compliant. We do not depend on buying data or selling data as a fundamental part of our business model, and we feel really great about that. We can serve our customers really well with the data they've given us to be used only for the purpose that they've given it to us for. We feel good about our activities with regard to GDPR in Europe. GDPR in Europe was an important step forward.
We think it's probably a sign of things to come, the whole company came together to really make that a positive initiative, we feel very good about the progress we've made, the results we've seen in GDPR compliance in Europe, we feel well prepared, therefore, as a company. It's a really important topic. I think it's sort of finally getting its day in the sun, which many of us think is important and the right thing. We take it super seriously, as a leadership team.
Okay. With that, we are out of questions. I'd like to thank everyone for coming today. That actually officially ends the webcast today. I'll turn it back over to Josh.
Great.