Good day, ladies and gentlemen, welcome to The Trade Desk second quarter fiscal year 2017 earnings call. All lines have been placed on a listen-only mode, the floor will be open for questions and comments following the presentation. If you should require assistance throughout the conference, please press star zero on your telephone keypad to reach a live operator. At this time, it is my pleasure to turn the floor over to your host, Chris Toth, Head of Investor Relations. Chris?
Thank you, operator. Hello, good afternoon. Welcome to The Trade Desk second quarter 2017 earnings conference call. On the call today are Founder and CEO, Jeff Green, Chief Financial Officer, Paul Ross, and Chief Operating Officer, Rob Perdue. A copy of our press release can be found on our website at thetradedesk.com in the investor relations section. Before we begin, I would like to remind you that except for historical information, the matters that we will be describing will be forward-looking statements that are dependent upon certain risks and uncertainties. I encourage you to refer to the risk factors included in our press release and in our most recent SEC filing. In addition to reporting our GAAP financial results, we present supplemental non-GAAP financial data. A reconciliation of the GAAP to non-GAAP measures can be found in our earnings press release.
We believe providing non-GAAP measures combined with our GAAP results provides a more meaningful representation regarding the company's operational performance. Lastly, I would like to highlight our participation in the following investor relations events. On Wednesday, September sixth, we will be at the Citi Technology Conference in New York, on Wednesday, October fourth, we'll be holding our first-ever investor day in New York City. Professional investors and financial analysts interested in attending the event should contact myself in advance, as registration is required. I will now turn the call over to Founder and CEO, Jeff Green. Jeff?
Thanks, Chris. Good afternoon, thanks to everyone for joining us today. The first half of the year, in particular Q2, have been great for The Trade Desk. We posted record revenue in the June quarter of $72.8 million and adjusted EBITDA margin of 35%, which exceeded our expectations. I'll talk more about that in a bit. What I get excited about most was that during the quarter, we had some of the highest spend ever in multiple channels, in mobile video, native, and audio. I'm super excited about each of these channels, more excited that they are all growing together. This is a sign that we are becoming more diversified as we add unique supply partners. This helps brands reach consumers wherever they go. We help them buy holistically. This well-rounded growth is as important as the growth itself.
We had one of our biggest quarters ever in display, but I'm even more excited that it was the lowest percentage of total spend ever, now below 40%. Audio is one of our most promising channels, and it's growing at an astounding pace, almost 400% since December. Audio is still nascent and has so much upside as more inventory comes online and more users leave traditional radio. If you believe like I do that audio and video are more effective advertising mediums, the fact that we are moving in that direction means that we are moving towards more effective advertising, which in turn makes our business more sustainable as we provide more value. It also is evidence that we have done as good of a job as anybody in programmatic at being on the channel.
Our customer base remains extremely loyal, and retention continues north of 95%. Similar to the past 3 quarters, we have accurately predicted the performance of this core base. Plus, we continue to see incremental upside from existing agencies with new advertisers from across many different industries, such as banking, communications, food and beverage, automotive, and consumer electronics. Additionally, data usage continues to play a larger role, growing 75% from a year ago. Through Q2, the number of customers using third-party data is nearly 100%. Year-over-year, our international operations grew about 3 times as fast as the U.S. Mobile continues to lead our mature channels in terms of growth, with mobile video leading the way, growing at 171% year-over-year.
Given mobile represents a scaled channel at just over a third of our revenue, growth this rapid is something we had not predicted. While we continued to grow quickly, I do wanna mention that the advertising industry is not without its challenges. While we consider ourselves a SaaS company that operates in advertising, many ad tech and some agencies with different business models than us have struggled. There have been a lot of ad tech M&A deals at prices that feel like capitulation. I also recently met with a CEO of a large agency and the first thing that he said to me was that business is tough. In our view, both of these trends are actually healthy. The ad tech deals are cleaning up the space. Companies that can't afford to be standalone are joining the ranks with those who can.
Changing business models to something more sustainable is a good thing. The changes in ad tech are making the future brighter, not worse. The same is true of ad agencies. They are having tough conversations with their clients about their business model and how they charge. Getting agencies more aligned with their clients is a good thing, though, in the short term, it's painful. Because we enable ad agencies instead of disrupting them, our relationships are getting deeper, and they are stronger than ever. We have worked really hard to win the trust of agencies and make ourselves indispensable. We're getting closer to that goal, which is good for The Trade Desk and good for the programmatic industry. Let's not lose sight of the fact that most people predict the evolutions of this space will lead to a happy ending, myself included.
This massive $650 billion advertising industry is both growing and transforming. As it does, we expect there to be fewer ads, they will be more relevant, advertisers will waste less money, and publishers will make more money. This is a win-win for everyone, including and especially the consumer. However, as the industry transforms, there will be more companies joining forces, more deals, and companies that are features but not businesses will sell for cheap. Old business models, void of transparency but rich in fees will go away, and this is not a tragedy. Some negative headlines fail to acknowledge that this time is one of the best ever in media, and the efficacy of programmatic advertising is unprecedented.
Many of these companies are aggressively trying to seek an exit. It is because the industry is weeding out those that have not improved their business models or are not adding more value to the industry than they extract in fees. They will continue to struggle and even go away as the industry matures. As this happens, The Trade Desk is winning more and more share and building more credibility with agencies and advertisers due to our transparency and objectivity and business model. As CMOs and agencies are digesting all the changes in our space, they are coming to us. Our objectivity is more of an asset today than it was last quarter, or frankly, ever before. We expect to continue to outpace the growth of digital, in large part because we buy holistically and objectively, and we don't own any media.
Transparency is an ideal that gets talked about a lot in advertising, but it too often means something different in practice. To be clear, we think we've set a new standard in advertising transparency. If an advertiser or an agency wants to know the details of every single ad impression they buy, we can and will provide the details, including price and all the metadata on every impression they've ever bought. Now I'll provide an update on the largest growth opportunities for our business, global expansion, mobile, and TV. For the better part of June and July, I spent time meeting with our team, agencies, potential customers, and data and inventory partners in both Europe and Asia, and I came away more bullish than I have ever been on our global prospects. In Q2, outside of the U.S., our year-over-year growth rate was over 130%.
We had record quarters in literally all of our offices outside the U.S., which includes London, Singapore, Hamburg, Tokyo, and Hong Kong and more. Some countries, like Germany and Japan, which had been slow and steady in their programmatic adoption, recently started to hit inflection points. Year-over-year, our Hamburg office grew nearly 150%, and Japan grew by even more at roughly 300%. We have been investing in these countries for over three years. These inflection points serve to remind us that out of the growing $650 billion advertising pie, about two-thirds of that comes from outside the United States. We have only scratched the surface. At this stage, being a public company is an advantage for us.
Outside the U.S., publishers and advertisers see that we are stable, that we're profitable, that we're growing quickly, have superior technology, and that we have a lower take rate than nearly all of the local DSPs, which puts us in the pole position to win over the long term. While we've never aimed to be the cheapest, our offering and rates often make our value proposition overwhelmingly competitive. We recently opened our Shanghai office, putting us at 20 offices worldwide and firmly rooting us in most of the major media markets around the world. Our focus has shifted from opening offices and getting on the ground to now growing the offices we currently have. We can grow in existing markets like Japan, Korea, China, Hong Kong, Germany, and France, where the media markets are already so large.
In Indonesia, where the total population is similar to that size of the United States, the opportunity is massive compared to opening any new office in a smaller country or smaller media market. We expect to talk about our global opportunity in more detail at our upcoming investor day. Our newest office is in Shanghai. We are building out the team and have begun working with some of the largest agencies in the world there. By our estimation, China is about seven years behind the U.S. in programmatic adoption. The overall QPS available there is a fraction of what is available in the United States. Price discovery requires trust, meaning publishers need to share info about the inventory that they're selling so that buyers can make informed decisions. The buyers need to trust that the data that they are utilizing is valid, it's accurate.
Creating price discovery and liquidity are two of the most fundamental and foundational principles of any marketplace design. China is still in the early process of developing a healthy programmatic market. We need to remember that China has massive publishers in the industry. There are tons of potential growth opportunities ahead. I don't expect China to pay dividends for a number of years, just like many of our other offices outside of the U.S. that have just started adopting programmatic more rapidly. In this case, the next few years will be slower growth than other offices have been as a percentage, but we expect the market to move quickly after the groundwork has been established. Some companies have a China strategy of go big or go home. We don't take that approach.
We approach the China market the same way we have entered all other major media markets around the world. We invest ahead, we build trust, we demonstrate our value add. Then we see those investments start to pay dividends over time. We reinvest as those markets adopt programmatic. It has worked in the U.S. and in every other worldwide location. We expect it to work in China too. We're off to a great start. We are already partnered with Baidu. We continue to have great dialogue with other major publishers in China. We expect that we will have many important partners there in the future, the same way that Google or Spotify or Oath and others have partnered with us here in the United States. Many companies have tried to go to China and failed to win the hearts and minds of Chinese consumers.
We are different in that we are not trying to win them to our brand. Most will never know our company. We go to publishers in China and bring demand from global advertisers because we represent the largest multinational brands in the world on our platform. Global brands that Chinese consumers already love can go to China with a coordinated and global message. This puts us in a great position with our customers to win spend, as we can power advertising budgets on a global basis with an omni-channel presence, completely independent and without bias to any specific media asset. Multinationals we already power want to spend in China. We expect that major publishers in China will find it hard to say no to big $ checks from the most premium brands in the world.
On to two of our most important channels, mobile and video, where we continue to make great progress. In Q2, all mobile, including in-app and mobile video, grew significantly and represented over a third of our business. We continue to see exceptionally strong growth in mobile video and mobile in-app, which grew 171% and 87% respectively compared with Q2 a year ago. Mobile continues to grow rapidly, we are well-positioned to win additional spend. As consumers spend more time on mobile devices and as advertisers become more sophisticated with video, we believe mobile ads will also become higher quality and be better integrated into the user experience than they are today. Mobile video is still one of the most untapped opportunities in all of digital, with its potential to connect with audiences since consumers are spending more time there.
With all of these forces coming together in mobile, we enable unbiased, data-driven decisioning and return a better ROI for our customers, not just on one site or one app, but across the entire internet and all of media. Perhaps the only other area that can drive more spend than mobile over the long term is connected TV, which leads to two questions that I get regularly. Why isn't connected TV growing faster? When will connected TV hit a tipping point? Before I answer the questions, I want to state that I firmly believe that no one is better positioned to take advantage of the move to connected TV than The Trade Desk. I believe most people are looking at it the wrong way. There's not really a question that things are moving fast. Connected TV is growing. It is growing fast.
No ad subscriptions are actually slowing down, consumers can't take many more ad-free subscriptions. They can't afford it. In our view, this is why ad-funded inventory is growing so much faster than the growth of connected TV. The amount of connected TV inventory we access has gone up by 10X since a year ago, the spend on connected TV on our platform is up 167% Some have failed to consider that subscription saturation will make the adoption of ad-funded models more hockey stick than linear. More importantly, traditional TV is a ticking time bomb. Traditional cable TV is losing millions of subscribers every year to streaming, and on top of that, millennials aged 19 to 29 spend more time on OTT than they do on linear, according to Morning Consult, a brand intelligence firm.
At The Trade Desk, an informal poll showed more than half of those in their 20s have never had a cable bill, ever. The average cable bill in the U.S. is over $100 a month, which is almost 50% more than 10 years ago. The number of ad commercial breaks has gone up over the same period, making the appeal of ad-free environments like Netflix more successful. Cord shaving, cord cutting have accelerated, while the younger half of millennials and nearly all of Gen Z are cord nevers. Since both economics and a better experience are causing consumers at every income level to long for internet-powered TV, the more interesting question is: when does traditional TV end as we know it? Traditional TV is now showing more ads to fewer people at a higher price than ever before, that is simply not sustainable.
As we've said before, we think the AT&T and Time Warner deal represent one of the biggest moves in TV ever. This signifies some of the biggest names in TV trying to win in the transition, that's dependent on ad-funded models. The question, when do others follow, is a much more interesting question than when connected TV takes off, because I would argue it's happening right now. Advertisers are realizing at an unprecedented rate that traditional advertising isn't producing the same efficacy that it once did. We frequently hear from large advertisers and CMOs that they are horrified that millennials don't know or care about their brands. They believe that they have not been able to reach them effectively. These are from progressive marketers who are early adopters in digital and programmatic.
They have put more dollars into traditional TV, all while OTT is adding people every day, especially from young millennials and mature Gen Z-ers who are in a critical stage of life where they develop brand loyalty. Not only do CMOs need programmatic TV desperately, TV content producers need programmatic. It is the only way for ad-funded content creators to make the change from broadcast to connected TV. While it takes big media companies time to turn the ship, they are turning now as they realize that understanding and adopting programmatic advertising is existentially critical to their businesses. The sheer size of the TV market, approaching $300 billion and growing of the worldwide advertising market, can move the needle unlike any other channel, which is why we are investing so heavily now in more inventory integrations, more partnerships, and more product development.
Finally, I'll take a minute to talk about our business model. As I've stated many times before, we believe our business model is exceptional. We don't believe that in our case, we must choose between growth and profitability. We can do both and have for three years now. In Q2, our financial performance, and particularly our adjusted EBITDA, was significantly better than we had expected. It came in at $25 million for an adjusted EBITDA margin of 35%. This is meaningfully higher than we'd expected, I want to give a little context. First, because of the operating leverage inflection point we've reached in the last year, we've increased our operating spend by more than $20 million year-over-year. That's the increase. This means our investment was up 65% over the previous year.
For our 3 major growth initiatives, mobile, TV, and global expansion, we spent more on product development, but we couldn't invest fast enough. We refuse to spend frivolously or foolishly. The precedent and the moral hazard is way worse than whatever benefit we would get from that. Since we are playing the long game, we are much more concerned about the cultural impact on making those types of decisions. Some of the hiring and investments we did not complete in Q2 are already happening in Q3, and we are adding to our technology investments and into the expansion of our global offices. We expect EBITDA to be 27% in Q3 as a few of our Q2 investments slip into Q3. We consider this a great position of strength.
After all, we are investing as aggressively as we can, and we're still producing EBITDA margins that compete with all of the other SaaS companies, including those much bigger and more mature than we are. We are setting the bar for what software companies have to produce to compete in the advertising business. The advertising industry is continuing to see ad dollars shift to programmatic, and we continue to win new business, and our existing customers are spending more on our platform. For 2017, we are increasing our expectations for the year. We now expect to be at least to $303 million in revenue, and we are increasing our adjusted EBITDA estimate as well. We expect adjusted EBITDA percentage to be at 29% for 2017, which revises our adjusted EBITDA guidance to $88 million for the year.
Agencies and brands are looking for a data-driven, easy-to-use platform solution that delivers a better ROI for their ad dollars. The Trade Desk is the answer, and this is why we are the largest objective, independent market share leader in the programmatic space, and why we believe we will continue to gain share in the future. I'm going to turn the call over to Rob to discuss our quarter in a little bit more detail.
Thanks, Jeff, and good afternoon, everyone. Our business continues to deliver outstanding results, and we delivered an all-time record in the June quarter with $72.8 million in revenue, which is really phenomenal given how well we did back in the December 2016 quarter. Every single office outside the U.S. set all-time records in Q2, led by Seoul, which grew 372% year-over-year, our Tokyo office growing 279%, and our Singapore office, which grew 220% year-over-year. From a channel perspective, our growth was driven in part by our mobile video channel, which grew 171% on a year-over-year basis, and our connected TV product, which grew by 167%. Our cross-device products, which clients can purchase to track user IDs across multiple devices, including smartphones, tablets, smart TVs, and personal computers, that also grew by 163% on a year-over-year basis.
Our Q2 was all about continuing to build momentum and trust with our customers, training more people on our platform and onboarding new customers to get ready for the seasonally stronger second half of the year. Our execution has been really strong. A good example of that came from our Hamburg office in Germany, which in Q2 grew by 146% year-over-year. The German market has been growing steadily and we have been making inroads there, but hadn't quite hit the inflection point we knew was coming. After years of hard work, investment, and patience, we broke through in Q2 and started to see stronger momentum from some of the largest agencies in Germany.
Starting in Q2, they began moving their spend for several of their larger clients over to our platform, and we are now helping them in the pitch process to win other business together as we move forward. As I've described before, from an operational perspective, we have three core priorities that we focus on. First, which is to remain the objective and independent advisor to our customers. Two, to continue to grow our omni-channel presence. Three, continuing to grow our international footprint. There are many ways we win the trust of our customers. I regularly highlight how hard we work to win new business, to train our clients, and to become their independent, trusted partner, and this happens every single quarter.
Today, I'd like to highlight a couple of things I don't usually talk about, but provide good insight into how we have maintained a 95% customer retention rate for 14 quarters in a row. It's a great example of how our engineering and business teams work together and the grit that our team demonstrates to build trust with our customers. In this example in Q2, we lost some spend with a large consumer technology company due to some temporary technical challenges. Our engineering team stepped up, as they often do. They jumped right on the problem and solved it quickly. We were able to go back to the client with a solution, and we said to them, "Hey, this is what we do when things go wrong.
The world isn't always perfect, and when it isn't, we do our best to fix it and we fix it fast." The client's response back to us was very positive, and just a few weeks ago, we were informed that we won back over $1 million of their spend for Q3. This is a great example of what happens when our business and engineering teams work together. The client told us specifically that they now know they can always count on us when there's an issue, and that is not true with their other partners. We earned a lot of points with them, and it's just one of the many great things that our team does to win trust one advertiser at a time. Another example I want to call out is from our user experience team.
We listen to media buyers who sometimes spend the entire day in our platform. One of the features they were asking for was a way to streamline campaign setup workflows in our system. Our UX team revised our workflows for campaign setup and launch and reduced the number of clicks by 30%. That may not seem like a lot, but when you're a media buyer and you're setting up hundreds of different campaigns on a platform for a client, a 30% reduction in clicks per campaign is a lot of time saved. We delivered on this initiative in Q2, and the feedback we've received from agency trading desks has been very positive. They feel like we've given them some meaningful time back in their busy day.
It's our ability to produce continuous improvements in our platform like this that goes a long way in helping to win the trust of our media buyer clients on our platform. Next, I want to focus on growing our omni-channel presence. The strong growth we have seen in mobile and video have enabled our clients to have a higher level of coordination and consolidation of their marketing spend across the whole marketing funnel, from brand awareness to consideration and then to purchase. Now, with more and more inventory coming online, we are working to add connected TV into that mix too. As an example, our team recently was working with an agency trading team to incorporate more connected TV buying into the client's overall programmatic strategy.
There was a lot of product work involved and a lot of trust to build on this newer channel, but our team crushed it and secured a $1 million opportunity in connected TV for the second half of 2017. In addition, large broadcasters are also starting to note this and are looking to us to monetize their connected TV and OTT inventory as they are realizing the opportunities in front of them now. Over the past quarter, two of these large firms have expressed their interest in establishing inventory deals, both for their own inventory library and with key shared large global agency clients to begin offering ad inventory that they have on their standalone connected TV apps. This app inventory is accessed through apps on places like Apple TV or Hulu or Roku, for example. We are building relationships and trust with the broadcasters and delivering ad demand.
As Jeff has said for years, all of our efforts in programmatic to date are only a dress rehearsal for the day when the traditional TV model breaks down, and we will be positioned better than anyone to capture that opportunity. Also in the second quarter, our native channel again saw robust growth as media buyers continued to use more native ad formats on our platform. The time we spent over the last few quarters training media buyers at agencies on how to incorporate native ad units into their ad campaign strategies is now paying off. Through Q2 of this year, our native channel could be the fastest-growing channel we've ever had at The Trade Desk, and we're still in the very early stages of growth for our native channel. One last area I will touch on is audio.
In terms of pure percentage growth, audio was our fastest-growing channel in Q2, massively exceeding the growth of other channels, albeit from a much smaller base. We are already in the early stages of audio going programmatic, but we're already playing audio ads in more than 90 countries on a daily basis, and it is a very effective medium for advertisers. Completion rates for audio ads are regularly in excess of 90%, and advertiser key performance metrics are regularly exceeded. The digital audio industry needs programmatic advertising as it benefits content providers like Spotify with a lower cost of sales, while also delivering more relevant and effective ads, which leads ultimately to a better consumer experience.
The final priority we are focused on is extending our geographic footprint to make sure we serve our customers locally in the markets that are important to them, our strategy there is paying off. We had massive year-over-year growth rates across the board, and during Q2 had all-time record spend in every single office located outside the U.S. International spend growth percentage outpaced that of the U.S. by a factor of about 3x. The adoption of programmatic and the market growth we are seeing across Asia is actually accelerating, and we believe the same progress will happen over time in our newest location in Shanghai. Our focus has shifted from opening offices and getting on the ground to now growing the offices we currently have.
In addition to product development and inventory and data integrations, we are hiring more people across our client-facing business teams to support our growth. We are adding business intelligence, partnerships, marketing, and business support functions in the regions so that we can increase our response time and serve our customers better locally every single day. We feel really great about what we've accomplished in the second quarter and the momentum we have entering the second half of the year. Our revenue and key metrics are growing nicely. We are consistently gaining incremental spend from existing customers, and we're winning new customers. The advertising industry is still in the early stages of programmatic transformation, and we are very confident in the direction of our business. I'm going to turn the call over to Paul to discuss our financials.
Thanks, Rob, and good afternoon, everyone. As you've seen in the numbers, the first half of 2017 is off to a record start, and we were particularly pleased with our Q2 financial performance and execution against all of our key metrics. Revenue increased 54% year-over-year, adjusted EBITDA increased 60% year-over-year, and GAAP net income was an all-time record of $18.8 million, a 148% increase from a year ago, all while investing aggressively back into our business in areas critical to our future growth. Revenue for the second quarter was $72.8 million, which was above our expectations and reflects both the expansion of our share of spend by our existing customers, plus the addition of new customers and advertisers.
For the quarter, approximately 87% of our second quarter gross spend came from existing customers, whom we define as existing customers that have been with us for over a year. Our operating expenses scaled efficiently with the growth of our business to $53 million in Q2 of 2017 from $32 million during the same period in 2016. The increase in operating expenses was primarily due to our increased investments in personnel, mostly in technology and development, an increase in platform operations expenses, which reflects hosting costs to support the increased use of our software platform, and in general and administrative expenses, which now reflect the costs of being a public company. Total other expense net was $1.3 million, and income tax was a benefit of $450,000 in the quarter.
Similar to last quarter, our Q2 income tax reflects a discrete tax benefit of $8.6 million, primarily related to incentive stock options. GAAP net income was $18.8 million for the second quarter of 2017, or $0.43 per fully diluted share. Our non-GAAP adjusted net income was $23 million for the second quarter, or $0.52 per fully diluted share, compared with non-GAAP adjusted net income of $8.2 million or $0.22 per share in the comparable period a year ago. Adjusted EBITDA was $25.3 million, with a corresponding margin of 34.7% of revenue during Q2 2017, as compared with adjusted EBITDA at $15.7 million or 33.4% of revenue during the same time last year.
The increase in adjusted EBITDA dollars reflects growth of our top line and the leverage of our business model, all while we aggressively invested in product, people, and global expansion, in addition to incurring public company expenses compared with a year ago. Net cash provided by operating activities was $10.7 million for Q2 2017, as expected, given the seasonality in our business. Over the past trailing 12 months, we generated $40.4 million and $25.8 million of operating cash flow and free cash flow, respectively. Our DSOs at the end of Q2 were 95 days, an increase of seven days or 8% from the same period a year ago. DPOs for Q2 increased nine days to 73 days, or 14.1% from the same period a year ago. Our net cash position is now at $89 million. Moving on to our guidance.
For Q3, we are expecting revenue of $76 million, adjusted EBITDA of $21 million or 27.6% of revenue. Updating full year 2017 expectations, we now expect revenue to be approximately $303 million and adjusted EBITDA to be $88 million or 29% of revenue. Total other expense net is expected to be between $2 million and $4 million, and our income tax rate for the full year is expected to be in the low 30s, reflecting the discrete benefits we received during the first half of the year. With that, I will hand it back over to Jeff for any final comments and of course, Q&A. Jeff?
Thanks, Paul. In summary, the programmatic revolution continues to gain momentum. The Trade Desk remains the clear independent industry leader as we move forward. For all that we have accomplished in the past six months, I personally believe that we are just beginning to scratch the surface of what is to come. Our objectivity that comes from being buy-side only and not owning any media is mattering more and more as time goes on. CMOs and agencies are in need of partners that align their interests. The digital world is so much bigger than two or three websites. Programmatic is only 2% of the entire $650 billion advertising industry, which is growing. There is a long way to go.
More importantly, our inventory partners, particularly in broadcast TV and the agencies and their clients, are only beginning to realize what success they can achieve from programmatic advertising on our platform. The first half of 2017 has been an important time for The Trade Desk. I would like to give thanks to our teams and our partners and our investors for helping to make this a success. I look forward to reporting on our progress and to our upcoming investor day on October 4th, when we plan to provide a more detailed look into The Trade Desk. With that, we look forward to your questions. Operator, let's begin.
Thank you. The floor is now open for questions. If you do have a question, you may press star one on your telephone keypad at this time. If your question has been answered, you can remove yourself from the queue by pressing one. Our first question comes from Shyam Patil from Susquehanna. Go ahead.
Hey, guys. It's Shyam. Congrats on a fantastic quarter. First question, Jeff, just following up on your connected TV commentary, can you talk about how you see that progressing in terms of platform adoption and materiality to The Trade Desk? Recently, we saw the CEO of GroupM North America leave to join AT&T. Just curious to get your thoughts on that as well, and how you see that impacting the overall opportunity.
Awesome. Thanks, Shyam. First, let me just say that I think the AT&T Time Warner deal is something of a North Star deal that everybody in TV and advertising should be watching. I'll come back to that. Let me just talk about the connected TV opportunity first. I actually think the most important metric that I can share to underscore the bullishness that I have for the space is the 10x increase in inventory over the last year. What that represents, I think, is in part that consumers are tapped out on subscriptions where there are no ads, and they are more and more open to the subscriptions that come after HBO and Netflix to include ads. That's what's creating so much additional inventory. The Time Warner AT&T deal, as I said before, I think is just so critical.
I think it's one of the biggest deals in recent media history. The fact that one of the most powerful people in advertising, Brian Lesser, is heading to AT&T, I think is a commentary on his bullishness on the opportunity because I truly do think that if they execute well, they'll accelerate the move to connected TV. To just elaborate just a little bit on that deal specifically, essentially what their vision is to take 5G technology using mobile to make it easier to install cable. Instead of using cable, they're actually going to use the internet, of course. What they're going to try to do is gain more customers by selling it for less than what cable companies do today by making everything on demand and then making it so there's fewer commercials. They fully recognize that Netflix has a great thing.
While most cable companies have said, "Hey, we're going to stick with the status quo, we've got a long time," you have somebody in traditional television heading for the door as fast as they possibly can so that they can take advantage of the migration. That's forcing everybody else to worry about it. When Disney announces in the last couple of days that they're going to try to control distribution themselves, that's the exact same reason that Time Warner does that deal with AT&T. I think that what we're seeing with that North Star deal is everybody heading in that direction, and with 5G expected to be here in 2019 or 2020, I think it's happening faster than people think. I think we sometimes underestimate how few ad-free subscriptions people can afford.
That increase in inventory is perhaps the most important metric that we shared today. The last thing I just want to say on connected TV is just that because we do not own any media, we operate at a massive advantage to those companies that have bias. By not owning a TV station or YouTube or anything like that, we have the opportunity to represent advertisers in an objective way, and that makes it easy for us to partner with all of those players. We think that we're one of the few companies that has a good chance at partnering with Netflix and Comcast and Disney and AT&T, where others will have buyers.
Great. Thank you. That's very helpful. I had one follow-up. Jeff, you also talked about M&A in your prepared remarks. We've seen quite a few deals, particularly on the demand side with Fuel, Tremor and AppNexus. Can you just talk about how you see this shakeout impacting The Trade Desk? Is this going to be something that's primarily a tailwind here in the U.S., or is it something that you expect to feel globally? Thank you.
Yeah. I expect it to continue globally. I know I talked about it in the report, I just want to reiterate, this sort of industry cleanup is a very good thing for the industry, and that this whole story is going to end with a happy ending, which is consumers are better off, advertisers are selling more product for less money, and publishers get paid more and more of the dollar goes to them. All of that is forcing the industry to move in the very direction we've been saying everybody has to go for the last 10 years, which is you have to choose buy side or sell side. You can't have that level of conflict of interest. You have to be global, and you have to be omnichannel.
The players that are being acquired most quickly and are most aggressively trying to sell, I think, are those players that are either geo-specific or channel-specific or have conflict of interest, and they need to adjust their business model, and they'd rather do that with more capital behind them in the hands of another company. That's good for the industry.
Great. Thank you.
Our next question comes from Mark Kelley from Citigroup. Mark, go ahead.
Great. Thanks a lot for taking my question. The first one is just on the changes to Safari that Apple's making the rollout towards the end of this quarter. I'm sure you've had the beta version for a little bit now. Curious to hear your thoughts on what you think the impact could be to, A, the entire ecosystem and then just for your business as well. Second, the international side, that growth sounds like it's going as planned. Just can you expand a little bit more on the Germany and Japan offices hitting inflection points this quarter? What's driving that now, given that you've been investing in those regions for some time? Thanks.
Okay. Awesome. Thanks, Mark. On the Apple stuff, first, I just want to provide a little bit of perspective. I love this report that Forbes wrote where they quoted that market share when they said that Safari has a 3.5% market share in the browser world. I think people have a tendency to think of Apple as such a dominant company, that they think they must dominate in everything that they do. They don't when it comes to the browser. In fact, Google's is north of 60%, while theirs is below 4%. At those levels, I don't think there's any move that Apple can make that has a significant impact on our business. I don't believe any changes in their intelligent tracking will have an impact on our business.
What I hope that they continue to understand is that the internet is powered by a quid pro quo, which is that publishers have to make money so that they can afford to create content. If Apple takes away from them the ability for them to monetize that optimally, then they're going to stop providing content, or they're going to steer users towards other browsers that are more friendly. Perhaps Apple makes an aggressive move so that they can sort of mess up the plans of Google. Given Google's dominance in the browser world, if they do that, I think it will backfire, and Google will benefit more from that than anybody.
There's a number of things in the prototypes that we've seen that make it so that it will actually have very little technological impact on us, but it is still very early, and really a lot hinges on how intelligent the intelligent tracking is, and then also how well they respect that quid pro quo. There's no scenario where I see it having a material impact on our business. On the second piece of it, Rob, I don't know if you want to comment on Japan and Germany specifically.
Yeah, absolutely. Happy to. Thanks for the question. Things are going great, as we've said on our prepared remarks across the world. Jeff has talked about, I think earlier today, we've talked about in the past that different markets have different characteristics. Germany and Japan have some similarities in that it takes time to build trust. We've been on the ground in both of those markets for more than three years now. In the case of Germany, four years. We've done well. In every year since we opened the offices, those markets have grown faster than the U.S., but of course, off of a very small base.
As they've scaled over the last couple of years, we've done what we always do, which is to go in, build relationships, get a chance to prove results, train people on our platform, they bring us more. When I say more, that means both more spend into programmatic, but also incremental advertisers from our existing agencies. That trend has happened in both markets this year, just at a higher order of magnitude than it has in the past. Specifically in Germany, there were a couple of agencies that we've been working with on a very small basis, a test basis for the last 12 months or so, that really leaned in in Q2 and starting to do business with us in a serious way.
They've, to summarize, invested more in their own in-house talent to be able to do programmatic buying in a better way, therefore we became even more of a natural partner. In Japan, a very similar story, slightly different dynamics around the agencies construct in Japan, it really comes down to that we've been there for three years, we've built trust and won more business than we ever have with folks that we've worked with for three years.
That's great. Thanks, Rob. Thanks, Jeff.
Sure thing.
Yeah. Thanks, Mark.
Our next question comes from Kerry Rice from Needham. Go ahead, Kerry.
Thanks. Congrats on a great quarter. It is nice to see some positive news after this trading day. Two questions. First, we have heard some reports of some lower pricing from SSPs and maybe the ability to de-dupe some of the impressions across exchanges. Can you talk a little bit maybe about the benefits that would be, or if you have seen those benefits at this point? The second question is on demand trends. We saw a couple of reports earlier, or maybe late Q2 or early Q3, about P&G pulling back on some programmatic spending, Unilever pulling back on some programmatic spending. Maybe you can put that in context, because I think P&G is probably growing with you guys, and I am not sure about Unilever, but I think it is a pretty big client through the agencies through you guys.
If there was more context on what that means, that would be helpful. Thanks.
You bet. It is a really interesting time. Let me just preface by saying there is a lot of nuance and esoteric details that we could go into on the pricing, especially the effects of the market pressures on SSPs. We will talk more about that on Investor Day. I just want to preface with that. On the de-duplication, that is always good for us. We have a relatively simple P&L where something like 60+% is in employees, 30% in machines where we operate our platform, 10% in the rest of our expenses. That 30%, when we can reduce the number of ads that we need to look at, or especially reduce the duplication of the ads that we look at, we are better off.
As SSPs and exchanges are looking to reduce their costs and the market gets more efficient as players get sort of weeded out, that results in lower costs for us and is more streamlined. As well as more pressures get on those SSPs or affect those SSPs, it makes it so that we can choose more deliberately where we buy impressions, especially when impressions are represented in multiple places. As we do that, it cleans up the industry, which is, I think, exactly what P&G and Unilever are looking for, to just segue into that second question. Yeah, there have been some public statements made about their reluctance to put more dollars into programmatic. I can tell you, as the platform that powers, I think, the vast majority of large CPG companies, there is a general anxiety in both directions.
They're afraid in programmatic that they don't understand it, that there's lots of nuance, that there's a bunch of speed bumps that we've hit in the last couple of years. When I say we, I mean the industry, which are things like viewability and brand safety and all those things that are simply that. They're speed bumps. They slow us down and make it so that we have to do some course correcting, but it does not change the fact that programmatic is the most efficient advertising, I'd argue, ever. They're also afraid that when they pull that back, that they are going to lose a generation, that people in the millennial generation as well as Gen Zs are not watching commercials. They didn't grow up on Saturday morning cartoons to learn about all the brands that all of us learned about.
How do I win their hearts and minds without video, without motion picture, and how do I not get too dependent on Google or Amazon as I sell my products? They have that anxiety, too. As they're wrestling with that, they're trying to right-size programmatic. What they're essentially doing when they get on stages and do that is they're begging us to get better. That, to me, is the most bullish thing in all that. If it were really bad and they were just moving budget over, they would do it quietly and go back to what was working. They're afraid that it's not, and they're just asking us as an industry to get better. We think the role that we play as the biggest independent is that we can lead the industry to something better.
Thank you.
Our next question comes from Youssef Squali from Cantor Fitzgerald. Go ahead, Kip.
Hi. Thanks for taking my question. Just a couple from me. Jeff, you had a very interesting and I think must-read piece on Recode yesterday. As a follow-up on that and your comments today, just curious what your thoughts are on premium digital video CPMs and where they can go relative to broadcast and cable CPMs. While digital video ads will likely be shorter than the 30-second slots on linear TV, in a people-based marketing world, they should be a lot more targeted. Appreciate any color you can add there on ad pricing as this rapid shift away from linear to digital materializes. Second, do you think the shift can expand or accelerate the overall video ad market despite cannibalization of linear? This would really be enabled from better targeting, I would think. Any color there would also be appreciated. Thanks.
You bet. On pricing, there's no question that the CPMs in digital are going to be higher because they're exponentially more effective. If you think about how many commercials are run in your house that you don't watch because there are just so many. Commercial breaks has become a time to leave because you've got time, or else a time to hit skip on your DVR. It's been really hard for the traditional measurement companies to figure out how to measure that. The thing that is so great about digital is we can measure everything. We measure how many seconds you watched it. We do that today in the digital ads that we run. In some of them, they're skippable, how many seconds did they watch, and how does that affect performance? We can measure everything.
As things are closer to 100% viewable and closer to 100% measurable, as well as targetable, instead of the sort of spray and pray methodology that has been used on TV for the last 75 years or 100 years, we can instead be very deliberate about what ad gets in front of which customer on which device. That is exponentially more effective. If you look at the gap today in CPM, they're actually not as exponentially different as I think the efficacy or performance is. Perhaps both have to give, because I think that perhaps traditional TV is overcompensated because we haven't fully measured the impact of all the people leaving. That's going to move both up. Can that happen without cannibalization or during the transition so that there isn't cannibalization in TV, to your second question? I absolutely think that that's possible.
But it just depends on some of the moves the big players make. Those like AT&T, are they ready by 2019 and 2020? How smooth is that transition? What's the execution look like is what's required to answer your question.
All right, great. Thank you, congrats.
Thank you so much.
Our next question comes from Brian Fitzgerald from Jefferies. Go ahead, Brian.
Hi, yes, this is John on for Brian. Thanks for the question. Just wondering if you could highlight any differences that you're seeing in cohort growth. I know earlier in the year you added unprecedented net new clients. Is there any difference in the growth of spend there across channels versus earlier cohorts or anything interesting to call out? Thank you.
I'll point this over to Rob Perdue or Paul Ross, if you want to comment at all on the cohorts.
Sure.
I'll have you comment on that.
I'll give a little color and Paul Ross or Jeff Green, you want to jump in? Yes, we had, as we said in the last earnings call, even more wins than we expected in Q1, which was great. We talked about that being a quarter where we batted 1,000, and that won't happen every quarter. That said, we ended Q2 really right in the zone of where we expected to in terms of new customer adds, in terms of growing spend from our existing clients, which I think Paul Ross called out some of the statistics there. It was a great quarter. It wasn't that Q1 knock it out of the park quarter, but it was among the best quarters we've ever had.
The one thing I would call out in terms of color is newer cohorts are just like all the other ones in the sense that they tend to start with us in a channel with an advertiser, right? They start in display, or they start in mobile video, we prove success in that channel, and they expand across channels. Our older cohorts tend to work omni-channel. They tend to work with us in three and four and five and even six different channels as they've consolidated more spend programmatic. The newer cohorts do just what all the other cohorts have done, which is to start in one or two. We prove value. They start to learn how attribution and getting full transparency works with reporting, and they tend to spread out both in terms of channel and also in spend.
Nothing different really, other than we have more to offer than we ever have in terms of channels. We continue to help them bring more of that spend into programmatic.
That's illustrated by the statistic that 87% of our gross spend came from existing customers who've been with us for more than a year.
Great.
Our next question comes from Dylan Haber from RBC Capital. Go ahead, Dylan.
Hey, guys. Solid quarter. Congrats. In terms of connected TV, what OTT platforms are you connected to now, and what platforms would you like to expand to? Just a follow-up. For the full-year outlook, unlike revenue, you raised EBITDA guidance by slightly less than the Q2 beat. Can you provide more color on where you are investing those incremental dollars? Thanks.
On the OTT stuff, in terms of partners, I'll just say that every major ad-funded player we're either in discussions with or partnering with, not all of them can be talked about, and frankly, some of the ones I'm most excited about, I can't talk about. As a result, I'm not going to go through the list, but every major player in ad-funded OTT we're either talking to. You're absolutely right that the guidance is lower given where we just landed, meaning that we're going to lower our EBITDA percentage in Q3, which I think is what you're asking about. The thing I just want to highlight is this is land grab time. This is the time where we should be spending as much as possible.
I almost felt a need to apologize for the 35% EBITDA margin instead of the 27% that we're implying going forward, because I want to take as many of those dollars and reinvest as possible. We want to invest in international. We want to continue to invest in TV. We're going to continue to invest in mobile. A bunch of those investments, especially in international, just crept into Q3. I'll just remind you that two-thirds of the global advertising pie is outside of the United States. We hope for it to represent 15% of our spend today. That means a decade from now, I hope this to represent over 50% instead of 15%. In order to do that, we just have to build up our organization.
There's more that we'll talk about on Investor Day as well, just because there's so much to talk about in the individual opportunities in some of the countries that we've recently gone into that I'm just so bullish. Because of the long game, I refuse to make bad investments or to be flippant about the investments. I'd rather be deliberate. The moral hazard of hiring the wrong people or investing too far ahead or investing in an office that we know we can't spend time in order to actually make it a success in the same way that we have the other 20 offices around the world, that's the reason why our EBITDA was 35%. Just remember, our EBITDA margins are as close to every scale SaaS company in the world, and yet we're still investing as aggressively as we possibly can.
I'm so happy with the status quo.
Great. Thank you.
Great. Thank you.
Thank you. This does conclude today's conference. We thank you for your participation. You may disconnect your lines at this time, and have a wonderful day.
Thank you.