Let's get going. Afternoon session. I'm Brad Erickson. I cover internet here at RBC. Very pleased today to be joined by members of the management team from GoDaddy, straight from Phoenix, Arizona, in the Bay Area.
Thanks.
Obviously. CFO Mark McCaffrey on the left and Chief Business Officer Gourav Pani.
Thank you.
Thanks for being here. Nice to have you guys.
Thank you for having us.
Yeah, of course. Of course.
Great New York weather.
Of course. Yeah, no, it's balmy, right? It's just like Phoenix or the Bay Area. Cool. I got my usual laundry list of questions, and we want to talk product today. Really, really excited to have Gourav here. Not that you're not great, Mark.
I get it. I get it. He's the popular one.
Yes, yes. No, it's really nice. Certainly a lot to talk about in AI. You're kind of the hot commodity. Anyone who has questions, feel free to raise your hand, and we'll obviously try and elevate those as we go through. I guess to start, just high level, just kind of walk us through the pillars. You guys have a lot of announcements, frankly, along with AI. You've got the product family within Airo. You've got Airo.ai. I want to talk about ANS a little bit. On that three pillars of AI, kind of walk us through the high level strategy, and then we'll kind of get into it from there.
Certainly. If you look at the three pillars, the first pillar is our AI pillar that is Airo, and that is consumer facing. We started off with a generative AI approach in 2023, and now we have transformed it into an agentic approach. In that process, we've launched Airo.ai as a brand new interface to acquire customers and actually simplify their onboarding and growth through agentic technology. That is thing number one. Pillar number two is what we are doing internally within the company. Obviously, we have spent a lot of time educating people on the power of using AI. The output there is, if you look, for example, any new platform that we've built in the last year, about 90% of that code is written by AI. Based on, obviously, we're an almost 30-year-old company, and we have a lot of legacy code.
If we look across the board at all the code written at GoDaddy, in October of this year, 46% of all code was written by AI. That includes legacy platforms over 13 years old that are being completely tested by AI. AI is finding new patterns to validate test strategies that we've never done with people. That's pillar number two. Pillar number three is what we are giving back to the tech platform and the community through ANS, which is Agent Name Service. It is how we champion the open internet, what DNS did for websites, ANS is going to do for agents. That's been our three-pronged approach.
Got it. Yeah, let's start with the consumer side of it, obviously. Airo's been out for a couple of years now, right? It's obviously morphed into the agentic landscape and all. To date, what are you, I guess, with your SMB customers, and especially you're obviously always testing with everybody, where are you seeing kind of the most value add and addressing something that was most underserved with those customers?
I would say that the most underserved segment is about to get supported within the next two weeks. What we are going to release to market is a Co-Founder Agent. The purpose of the Co-Founder Agent is to allow anybody to come and ask a question around a venture or business idea that they have. The Co-Founder Agent goes and does market analysis, looks at the competitive lens for that business in the geography where the person wants to start their business, gives them pros and cons of starting the business, risks, pricing strategies, et cetera, and creates an entire business plan for the individual.
What's interesting about that is for the aspects that GoDaddy can enable off the business plan, like building an entire online presence, building a brand book, et cetera, the agent will ask the person, "Would you like me to take care of some of these things for you?" If you answer yes, that agent now delegates responsibility to a Website Agent, a Domain Agent, a Logo Agent, a bunch of aspects for commerce, et cetera, and just builds the business right there organically for the customer. I feel that level of simplification where the customer actually doesn't need to even know what to call their brand, just have an idea in their mind, and everything gets created for them. I believe that is the most underserved aspect, and we're about to serve it.
Yeah, it's great because it's actually meeting now customers at a funnel level that we hadn't met them before. Before they even think about the idea, they just think about what they may want to do. And then the ideation becomes live almost instantaneously. You talk about the health of the front of the funnel. This is really potentially a game changer for us. Yeah, yeah, absolutely. Maybe help us distinguish between that and Airo, right? Because, I mean, Airo was a big leap forward, right? Just in terms of automation and reducing some of that friction. This sounds like yet again, but where does Airo, kind of as we think about what Airo was maybe a year ago or 18 months ago, where does that sit in the roadmap relative to this co-founder product?
Airo, obviously, was an experience that helped customers attach to their core business. It helped us with getting customers to purchase more units, spend more dollars with us, and it also increased the retention rate. That was the core contribution of Airo. Airo still continues to exist, but what we've built with the agentic experience is to take Airo to a whole different level. We're still using elements of Airo, but this whole agentic experience is bundled under Airo Plus. Customers don't actually have to understand each aspect of it as an individual SKU and figure out what they need to purchase. All they get is Airo Plus, and with it comes an assortment of agents that help them solve their jobs to be done. It's not just about taking care of a blank slate problem, which is what Airo solved for.
Just do not ever start with a blank screen. Have something to go off of. This actually predicts what will set them up for success and builds out the entire experience. Another thing that makes it different is, unlike most other products in the market that are reliant heavily on just third-party LLMs, we have built a next best action model using our own data that actually gives customers an insight into what they should be doing next so that they do not have to think about 20 things to do. It is just the one thing they have to do next. As they go through the one step and the one step and the one step, we believe that type of an experience will take them further in their journey.
Yeah. When you think about the customer acquisition of these products, how do you think about this relative to the main channel? What will the kind of usage characteristics look like? Think customer LTV. Talk through that.
It's hard to tell what the LTV would be because we have to still collect more information on the types of customers we get. What we have seen organically, the customers that are coming in are very different from the customers we have acquired before. godaddy.com will continue to acquire the customers that we have traditionally acquired. Airo.ai seems to be acquiring a different cohort of customers. I think that makes it highly incremental. For instance, when we look at the prompts that are coming into the App Builder product, which is one of the capabilities of Airo.ai, that audience is very sophisticated. We have also seen interesting things like.
What does that mean?
It is actually a more technical audience. It is not an audience that is a microbusiness audience that needs a drag-and-drop product. They are more than willing to engage with it. For instance, we saw someone solve the problem of building an entire website in Arabic with right-to-left text. We did not even know that was possible. Lo and behold, someone was using it. Now we are optimizing for it to figure out how to build it the right way. We have an employee who speaks Egyptian Arabic, and she was pressure testing it. She said it actually met her bar of what it would take to build a commerce site in Arabic for the Egyptian market. That is not a priority market for us. The possibilities are endless. It just opens up a lot of doors.
Got it. Okay. So, I mean, I guess then the question becomes, I mean, market size, right? We know how many SMBs are in the world. We know how many domains you have, quarter of the world. What do we think about the size of this opportunity if it's such a distinct customer?
Yeah. Early stage, and we just launched this, and we're starting to see the interactions which Gourav's talked to, which is great. You used the word. This is all incremental to what we're doing on godaddy.com. If you think about Airo, it launched two years ago, and the results it's now showing within our cohort. We talk about that cohort a lot, the customer paying $500 to us every year. You see that the average order size is going up. You're seeing the attaches coming faster. You're seeing the retention almost near perfection at that level. Now you think about this being launched into that environment and the incrementality that will provide on a long-term basis. Obviously, we'll talk about it as we get more information, but it only helps that analysis we always talked about.
If you assume a domain is 1x and a customer goes from a domain to a website to an email to commerce, that becomes 83x in that LTV equation. This is going to be incremental to that 83x.
Yeah, yeah. I guess just where are we in this rollout? I get it. It's very new, and I'm asking you all sorts of leading questions.
Oh, it's exciting.
Yeah, but talk us through what you can in terms of the roadmap. I don't want to steal any thunder from the dinner in two weeks.
Yeah, we have the dinner in two weeks. Thanks for promoting it right now. I'll let you talk about the product a little bit, but we couldn't be more excited. We've launched it. We haven't put money behind driving traffic there, but yet we're seeing people come to it on their own and discover it. When we talk about the roadmap here, I think the only thing that is slowing us down is we have to focus on quality to make sure we're not overwhelming our customers. Gourav, you can talk about that. This is all out there. These agents are out there. They're in market. They are working. We are doing this within our existing model today where we're obviously generating a lot of free cash flow and growing the business, but it's all within the parameters.
Yet we're able to introduce these agents into market now.
I mean, you said it's an incremental customer, kind of a distinct customer. You're not putting any marketing force behind it. Where are these folks coming from?
It's tough to tell. I think the two areas where we have mentioned Airo.ai are social and through PR. For whatever reason, we believe one of those two or both of them are working. Obviously, the number one priority for us right now is not about just getting traffic. It's about creating a seamless experience for our customers. Specifically, when we went live and announced it in our earnings statement on October 30th, we had six agents live, five core functional agents and one orchestrator. Within the 2 1/2 weeks since then, we have launched five more agents. In the next two weeks, we'll launch 11 more agents. Roughly in two weeks' time, so effectively a month from when we first went live, we'd have 22 agents. Every agent is essentially a job to be done for our customers.
It is an atomic unit of work that gets done end-to-end for small businesses. Now, that number 22 actually could be double. There is a reason why it is not double. It is not because we cannot build features fast enough. It is because we have raised the bar internally to say, "There is no feature we are going to ship that does not have a seamless experience for customers." We cannot ship our org chart. We cannot ship products. We should be shipping entire experiences for customers. When we have done user studies, that is the feedback we get from customers. Looks really polished. It is fantastic. I wish it did not make me think when I went from thing A to thing B. Customer expectations on the cognitive load being lower is a priority for us.
Once we have that dialed in, and obviously, we have it dialed in for 22 agents, I think by investor dinner, you'll see a much bigger number than that. Once we get there, then we can actually think about investing behind marketing.
Got it. Funnel-wise, obviously, you always, I mean, you have a structural advantage just objectively because of owning a quarter of the world's domains. I mean, is that where you go after first, or do you go broader than this? No one's actually said the word vibe coding in the room quite yet. I think we're not sure if it's a good word or a bad word still or whatever. Yeah, how do you think about that as you kind of bring it to market?
I mean, we will solve problems for existing customers. For example, there are customers who have dated websites, and we can improve those using agents. Agents are able to take in hints from an existing asset and create a whole new variant of it. That is one option. The other one is obviously improving their social media presence and marketing, email marketing, blogging, all of those capabilities. The Blog Agent is actually live already. There is a lot of good value that we can create. It is vital that we do not create what we call internally as AI slop. AI slop is just like churning out generic information that is thrown out by an LLM. What we have to do is to enrich that information with hints from our own data set.
Our data that we have collected over almost three decades actually helps us fine-tune the content that we will show to customers based on verticals. That is the main strategy that we're going to go after, is how do we think about verticalizing, how do we think about expanding these capabilities, but not going so fast that we have a drop-off in quality.
Talk about just the ability to customize the LLM because obviously, I think that's like a really important point here. You've spun this up fairly quickly, right? We were only talking about, I mean, I remember talking to Aman just probably two years ago about initial usage of ChatGPT expanding to other models when Sam got fired and for the weekend and things like that. So two years we're here, and you already have the customer ready to go. How hard was that, right? What's the barrier there for others doing it?
If you do not have the data, you simply cannot do it. I think that is the biggest barrier. I would say the advantage we have is the amount of information we have within our walls. Also, we have made it very easily accessible within our walls. It is one thing to have data; it is another thing to have data in accessible formats. I think those two are the foundational things that will stop anyone from taking advantage of this at scale. You will see a lot of things pop up, and companies come up with brand new ideas and look very interesting. They have a plateau because they simply cannot continue to spend into using third-party LLMs with no enrichment of their own.
This is an important fact for us too. We're doing this within our existing model, right? We are creating the efficiencies and the operating leverage. We're using that leverage to reinvest back in the business. It's not changing our overall goals of, "Hey, we're targeting 33% by the end of the year for next year for the full year." We're doing that under an environment that we're investing in AI and these agents and getting them to market. I'll give credit to Gourav and the team. Their key is they're laser-focused on the jobs to be done by our customers. That brings in our ability to get more efficient, and that brings in our ability to meet their needs. That keeps the LTV part of the equation going.
Yeah. No, it's amazing. Maybe give an example or two of your existing customer base. Is it completely greenfield to sort of implement agents with existing customers, the existing small and kind of micro businesses that you guys serve so well? Maybe just give an example or two of how you could go in and enhance somebody's business that is functioning fine today, but they don't realize the efficiency they could get out of it, just so we can sort of conceptualize.
I think for most existing businesses, if you ask a small business owner, they'll say their number one problem they want to solve is to get more customers. In order to get more customers, you have to have better marketing. In order to have better marketing, you have to have a clean and compelling brand. A brand has to be present more than likely in social media or other forms where the customers are present. Think about content like images, obviously, is very basic, but video content. How do you create video? How do you personalize a message for every person that buys something from you? It's now possible to do that. Now, obviously, the cost has to be managed because video content creation is still quite expensive. We're not going to offer it until we can optimize it efficiently.
Fundamentally, as a seller on the internet, you could theoretically send a personalized message in your voice with your face, thanking your customer for buying from you. That really changes the narrative of what it means to have a personalized experience from a seller, a small business seller. I believe we can do that. It is well within reach. Technology is already there. The price point just needs to be managed. Those are some of the things we are working on. I am sure that is the pattern that will start. Over time, I believe there will be more such elements that will pop up.
Just kind of a market question. There's obviously a ton of vibe coding applications out there, and one new one comes to market almost every day, I feel like. Clearly, I think from the data we've seen in some of those early players, they've done really well. I think the common shortcoming has been is they can ideate really well, right? Throw an idea up on a whiteboard, digitally bring it to life, so to speak. To sustainably run it, right, you need some of those back-end tools. You need marketing. You need payments, inventory, or half a dozen other sort of plugins and so forth. The software investor view is that eventually they will get there because it's AI and it's easy to build and all that stuff. You're a product guy. You do this for a living.
Why is maybe that not the right way to think about it, or is it?
If you want to issue an SSL certificate, you have to be an issuing authority. You have to have a certificate authority. Not everyone has it. We do. Now, can someone go and resell somebody else's SSL certificate? Sure, you can. Then your margins are affected. For us, it's like we can generate the certificates ourselves and deliver it to customers, and we can do it at massive scale. We support some really large players who are resellers of our certificates. Same things hold true for domains. Same holds true for actually hosting the app. The foundational things like getting a customer's app with a domain, hosting, security is just built into our narrative, and it just scales. Other things like we've talked about ANS, for instance.
We are going to market with a concept and trying to get people to adopt a technology that is open, that makes agentic technology broadly available to everyone. That is very different from what's happening with the walled gardens who want to keep things within their purview. When new players come in, they actually want to work with somebody like GoDaddy because we open up the world to them. I do not necessarily think of these vibe coding apps or people building these apps as being adversaries. They literally will have to use the backbone that we have already built. To us, they are more customers that we can work with. I think it is an opportunity for us to actually increase our base and actually have our core products offered not just to our direct customers, but to the customers of our customers.
Yeah. And then the idea of custom agents. I mean, you guys are building, what was it, 22 at this point, more to come in a couple of weeks, it sounds like. The idea of custom agents within this ANS framework is super interesting. What could that even turn into?
I mean, historically speaking, if you think CompuServe wanted a walled internet and that never happened. Here we are with millions and millions of websites out there primarily because DNS made it easy to discover, a browser made it easy to access, and SSL made it easy to trust. Those three things still hold true for agents. We need to make discovery easy, and we have built ANS as a standard on top of DNS. It is not a net new standard. It is an existing standard that we extended. It is an open standard, so anyone can adopt it. We still need certificates to create trust. All these factors that we are building with ANS will just, I believe, lead to the same kind of proliferation of agents that we saw for websites.
Interesting.
That's important for our customer base when you think about the micro business doesn't want to be in a walled garden where they lose control of who and where they can sell. They want to be able to sell across multiple platforms because that's important to their ability to grow a micro business in and of itself.
Totally. Let's talk pricing and bundling, your favorite topic.
Yes, thank you.
How, put this into context for us? We know what pricing and bundling looks like today in the traditional product set. We're talking about a lot of new things. How does that change?
Yeah. Think about it as these are additive to each other and the incrementality around it. Pricing and bundling still is an initiative that is working on GoDaddy.com that allows us to value-based price to customers based on the value that they're getting from multiple different bundles. It's an ongoing strategy for us. We saw the results it has given us in 2024, now in 2025. That continues on. This is an incremental layer that has paywalls within it that we're experimenting with. That can ultimately become part of pricing and bundling. It also can become its own separate tier. It can be a premium service. It allows us to look at the value they're getting out of it and ultimately add to it later on based on what we're seeing.
It is being done on Airo.ai because that allows the experimentation to take place there where we can see where the value is being delivered and not get muddled up within the core platform of GoDaddy.com. More to come on how we ultimately do that. It is all part of the same schematic of pricing and bundling, seamless experience, and Airo being the driver of it all at the end of the day.
I think 2024 was kind of the first real big year for pricing and bundling. 2025, kind of a, we'll call it a tougher comp a little bit.
Thank you.
We know you're always, particularly Q2, we know. As you look towards 2026, right, you always talk about this. You're always testing and iterating, right, and seeing what works, what people like, what people will attach. How central was all of this to some of that commentary? Because you've been giving that commentary for a while, right? That's not a new comment out of you. How central is all of this in the context of that prior sort of pricing and bundling commentary as we think about next year and lapping 2025?
Yeah. Obviously, we'll talk about 2026 a little further, but it is a, I would say, it works within the existing model that we've built that we are still looking to what our customers need, what are their jobs to be done, how do we bundle that into products, how do we put that at a price point that works within our customer base, how do they continue to get value, and then how do we add to that on a continuation basis. The great thing is this isn't a new initiative for us. We have several years of experience with it now. We know how to test, to experiment, to look at the results, and to launch. I would say the biggest change we have right now is when we first launched it, it was very product-specific.
We are moving it to be very customer-specific as to what that customer is getting for value. This will continue the journey for us, right? This is where we get to that LTV, and this is why we feel comfortable that our model continues to work. I have to say it continues to compound free cash flow.
Yes. Got it in there. Thank you. Yes. We love that. And then just on that point on playing to the customer, you guys have talked about each of your customers uses the products in such different ways, right, that there is going to be a, I do not want to say a bespoke pricing, but there is going to be sort of custom buckets of pricing. I do not think you are very far along in that, or maybe I am wrong about that. Where are we in the journey of sort of optimizing price for what value your customers are really getting out of the products?
It's just a matter of signal collection. Let's talk about the agentic experience. One of the things that has been a priority before we ship anything is telemetry. We have to get signal on every single user's, customer's usage of the product. We do evaluation on the answers we gave them to see the quality of the answers that are being generated. Evaluation and usage behavior both tell us based on reinforcement, "Oh, this is working well for the customer. This is where they got value. This is where they did not get value." That is inherently built into the product before we even ship. What that does is when we go to value-based offers on the pricing side, all those signals are readily available to the algorithm.
Day one, it starts to learn based on the reinforcement by the customer and their usage pattern, whether they're getting value or not. Unlike other products that were legacy products that had to be retrofitted, these products have been built ground up with the signal collection.
Got it. Okay. Long way to go. Noted. Just as we think about your favorite topic, customer count and everything, let's go there in a second. Small business formation, right? It has kind of been stagnant the last few years. It exploded during COVID. There were probably a little bit more of sort of, we'll call it defensive or reactive businesses formed. You have had to sort of churn that out of the system. Here we are, 2025, headed into 2026. This technology seems like it has the potential to sort of almost change new business formation structurally. Are you guys seeing any indications of that? Do you have any thoughts on that? Just curious.
Yeah. We'll take it at the highest level, and then we can talk about specifically what we're seeing today. As we've gone through these, we've been around for 30 years. We've seen a lot of trends, and we've lived in a lot of ups and downs, and we've been able to grow and grow with entrepreneurs through that entire time period, right? The cycle around people coming to market with ideas, ideating, and wanting to launch things has always been a consistent, steady pattern that maybe spikes here or there in certain years based on some economic factors. Right now, what we're seeing is that stable environment, right? We're seeing customers coming in with ideas, high-intent customers wanting to get those ideas online, whether they're selling a product or trying to share content, could be any of the above. That hasn't fundamentally changed throughout this year.
We saw variations in COVID, no doubt about it. Since we've come out of that cycle, we've seen a stability that the only thing impacting that has been the actions we've taken internally going after the high-intent customers, which we talk about. We've seen the right behaviors. They're coming in. They're attaching faster. Their conversion's better. We're getting to that $500 mark a lot faster than we used to. We have perfect retention in there. All the metrics we like to talk about because the strategy is working, because ideation is now getting to launch a lot faster. What we're talking about, Airo.ai, just closes that gap even further to being able to get online. Think about it where a small business maybe had to hire four or five people to do certain tasks. Think about the writing a business plan concept in and of itself.
You used to have to pay someone to go do that for you. Now it's done right there for you with all the market research, and you can decide whether you want to launch this in your geographic area or you want to move somewhere because it's a better market. All those things now just create a better environment for entrepreneurs to be able to get through all the hurdles that they used to have to get their idea online and to sell them. We think that is exactly where we need. That is why we champion the open internet too. They need to be able to get it to wherever they want to go and be able to sell it everywhere.
Got it. One last one. We can't have a 30-minute discussion without talking customer count. Can you update us just on how you, and we've talked a lot about clearly there's a lot of potential with these new products, but just setting that aside, customer count, you've had some things where you've done to the business, hosting migrations, tech migrations, little divestitures that affected that metric. Investors, obviously, are kind of waiting for that to stabilize, if not turn positive. How should we think about that?
We've talked about this a lot, no doubt about it. We are through the divestitures, through the things we did that impacted that number. We're in a steady state stabilization now we're going through. A reminder, our strategy isn't to grow customers just for the sake of growing customers. Our strategy is around higher intent customers, customers that are going to come in, attach product, want to get that idea out online. The $500 has become a big mark for us that represents when a customer has to be attaching products most of the time and has to be getting some value to renew. That's the customer we're going after, and that's the customer cohort we continue to see driving our ARPU going forward and continues to be a larger and larger part of our bookings number as we go forward.
That is our strategy, and that's what we continue to focus on.
Got it. Understood. That's great. I think we're out of time, but Gourav, Mark, thanks for being here. Anybody who's in the Phoenix area December 2nd, come out to the Investor Dinner.
Just call Christie. Beautiful headquarters. She'll get you a seat.
Sounds good. Thank you, guys.
All right. Thank you.