Greetings, and welcome to The Trade Desk second quarter 2019 earnings conference call. At this time, all participants are in a listen-only mode. A question and answer session will follow the formal presentation. If anyone should require operator assistance during the conference, please press star zero on your telephone keypad. I would now like to turn the conference over to your host, Mr. Chris Toth. Thank you, Mr. Toth. You may begin.
Thank you, operator. Hello, welcome to The Trade Desk second quarter 2019 earnings conference call. Our call today is taking place from our Ventura headquarters. On the line is our Founder and CEO, Jeff Green, and Chief Financial Officer, Paul Ross. A copy of our earnings 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'll be describing will be forward-looking statements which are dependent upon certain risks and uncertainties. I encourage you to refer to the risk factors referenced in our press release and included in our most recent SEC filings. In addition to the 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 that providing non-GAAP measures, combined with our GAAP results, provides a more meaningful representation of the company's operational performance. I will now turn the call over to founder and CEO, Jeff Green. Jeff?
Thanks, Chris, and thank you all for joining us. I'm excited to announce that our revenue growth accelerated to 42% in Q2 compared with 41% last quarter. Our accelerating growth is a testament to the increasing trust that major global advertisers are placing in us as they shift more of their advertising dollars to programmatic. To understand the significance of our growth, we need to provide context for what is happening to advertising from a macro perspective. In 2019, according to IBC, global advertising spend will be about $725 billion, up over 4% from 2018. At current growth rates, global advertising will be a trillion-dollar industry in about seven years, one of the only handful of industries with a TAM over that mark. Programmatic is still a relatively small part of total global advertising. It is estimated at around $34 billion in 2019.
It is growing five times faster than total advertising at around 20% year-over-year, according to Magna Global. We maintain our prediction that before long, most advertising will be digital, and nearly all of it will be transacted programmatically. This puts us in the fastest-growing segment of an expanding industry where we expect to continue to aggressively take share. We saw this in our results last year. We are continuing to outperform the industry again this year, which is reflected in our biggest Q2 to date. More significant than comparing our company's performance now to our performance in the past, our growth rate is 10 times the pace of overall advertising growth, and in fact, our growth is nearly four times the growth rate of digital. There are many reasons for our industry-leading growth rate.
Some of our growth is coming because we're executing well and capturing opportunities, but other growth drivers are secular. The digitization of advertising, particularly TV, is massive. The shift to data-driven decisioning versus guessing or intuition is game-changing. These changes in the landscape significantly benefit us. We have created our strategy, our technology, and our business model to take advantage of these shifts. As a result, more advertisers are standardizing on our platform. Every brand and every agency knows that they need to be engaged in programmatic advertising, which is another way of saying that every brand has come to understand that advertising using data-driven decisions is much more powerful and effective than simply following intuition alone. Our results are proof that this is occurring. For example, in the second quarter, we signed 55 new MSAs representing some of the largest brands on the planet.
This is the highest number of new MSAs we have signed in one quarter since the end of 2016. Through the first half of the year, 60% of the Fortune 500 companies are now growing advertising through The Trade Desk. More brands, nearly 50,000 of them, mostly through their agencies, are moving more of their ad budgets into data-driven advertising. These numbers are very telling because once advertisers embrace this model, they start to shift more of their budget into it. Once one person on an agency account or a brand starts to realize the value of data-driven precision, then others start to adopt it across more channels. We see evidence of this everywhere. For example, in Q2, EMEA had continued solid growth and share gains. Our offices in both Madrid and Paris achieved record spend in Q2.
We added more major advertisers and more accounts grew their spend with us. During the quarter, we added a large regional hotel chain and a smart home energy company, to name a few. We are seeing similar trends in APAC, where our offices in Hong Kong, Jakarta, and Sydney all achieved record spend. We are also seeing exciting growth in many of Asia's fastest-growing emerging markets. eMarketer defines Southeast Asia as Singapore, Indonesia, Thailand, Malaysia, Vietnam, and the Philippines. They predict digital mobile growth there to be about 35% in 2019. In several countries, we are growing exponentially faster than that. For example, Vietnam, we grew 125% year-over-year. In these smaller but fast-growing markets, data-driven advertising is the perfect fit. The cost of sales and distribution is way lower, and market-moving shifts require adoption from fewer people.
I continue to predict that these emerging markets will be nearly 100% data-driven before the U.S. To continue demand in these emerging markets, you have to grow supply. That's why I'm excited about the premium inventory partnerships we have developed in Asia. For example, in Thailand, we access the country's top TV digital platforms, Channel 7 and Channel 3. We also access TrueID TV, the second-largest broadcaster in Thailand. In Vietnam, we partner with the largest free TV station, VTV. Globally, our commitment to invest and innovate in the channels that matter most to advertisers is a major driver of growth for The Trade Desk. In Q2, 47% of spend on our platform was in mobile. That's a new record. Mobile video spend growth was about 50%. Mobile in-app spend growth was about 63%.
Data spend, again, was up 80%, and cross-device spend was up over 250% for the third quarter in a row. Audio, which I believe is the best value in programmatic today, grew almost 3x in Q2 for the second quarter in a row. Connected TV spend growth was also amazing, growing nearly 3x from a year ago. I said before, we will likely never see a channel larger and more full of opportunity than we have right now in CTV. Much of what we've done over the last decade has simply been a dress rehearsal for the digital shift happening in TV right now. This is important because as we drive towards a trillion-dollar total advertising market by 2027, about half of that market will be in some form of video, and most of that will be in premium TVs.
We are at the very beginning of the digitization of TV advertising. More and more consumers watch TV content through connected devices. As they do that, advertisers are getting a much better understanding of who is watching their ads and how they are responding. This enables advertisers to apply real data to their large TV ad campaigns for the first time, making those campaigns more precise and more effective. This kind of targeting is simply not possible with linear TV. This is especially important for major consumer brands who are among the most aggressive in embracing data-driven advertising. For them, brand differentiation is critical. They are facing unprecedented margin pressures in their core business. Getting smarter, more precise, and more scientific about how they invest their ad dollars is a very effective way of maintaining and enhancing brand value. Nowhere is this more apparent than in TV.
Nearly every discussion I've had over the last few months with advertisers or content providers starts with the potential of connected TV. It was the number one point of discussion during the recent upfront season and at the recent Cannes Lions conference. Every top advertiser wants to know how they can best access CTV inventory at scale and how they can apply programmatic to it. On the publisher side, all of the major premium TV content providers in the world want to know how they can make more of their premium content available for programmatic demand. This is driving our premium inventory supply growth that we see every quarter. In Q2, our CTV inventory growth was up over 3x from a year ago, but our momentum doesn't stop there. We recently announced that Amazon Publisher Services is working with The Trade Desk.
This is a breakthrough deal in connected TV for several reasons. As always, let's start big picture and then zoom in on what this means directly for our business and our advertisers. This announcement is a victory for the open internet. Many decisions by big tech players have created walls around their content and reduced transparency for advertisers. In this deal, Amazon has taken a different path. By making this announcement, they are staking four claims. One, they are joining the open internet. Two, aside from advertising on amazon.com, they are centering their efforts on the sell side. Three, they are going to reduce the cost of sell side fees significantly while being completely transparent. Four, they are partnering with the demand side instead of the go-it-alone strategies that the walled gardens often deploy. Let me tell you why I believe these points are significant.
Advertisers on our platform now have access to 100% of Amazon's third-party content providers. Think Discovery, NBC, and ESPN apps on your Amazon Fire Stick. This is quality inventory on premium content. With APS, we have a joint focus on improving the consumer experience by doing better with ad frequency. Amazon is providing an anonymized ID, very similar to the IDFA anonymized ID that Apple shares in its mobile ecosystem, and we can use this to measure reach and frequency across the entire internet. It also allows brands to unlock the power of their own data and insights, their most valuable data, to drive ad relevance across the TV ecosystem. In so doing, Amazon Publisher Services is supporting the open internet, in contrast to big tech walled gardens. It's a strategic move which I believe will put significant pressure on other CTV aggregators.
Finally, Amazon is being transparent about their fee structure and exactly how much the publisher will receive, something that's very important to us and something I'll come back to shortly. We are equally excited to be working with other major TV content providers. For some of them, it's the realization that more than half of their viewership is now coming from connected devices. For others, it's the realization that live TV events, such as sports and political debates, are driving massive new viewership on connected devices. They need to overhaul their models to allow for ad price optimization during unpredictable spikes in demand. This requires a new business model, and these providers are coming to us to partner on that strategy work. We are driving similar relationships internationally.
In Europe, for example, we are working directly with many of the largest media companies, such as RTL Group, ProSieben, Mediaset, and others. Across Asia, we have access to many of the top OTT video streaming services such as TVB, HOOQ, iflix, Viu, and Line TV, just to name a few. Across all of these relationships, our focus is on monetizing premium ad-funded content. This contrasts with walled garden platforms which prioritize user-generated content. Premium content is more valuable to advertisers because it's much higher quality, it's inherently more brand safe, viewers cannot skip through these ads typically, and there are better content-to-ad ratios. The shift toward premium content will also force big walled garden platforms to reevaluate their take rates and margins. All in, they are operating at something like 50% today.
Emerging platforms, including Amazon and others, who are just starting to make their content available, are operating at a fraction of that, and often with significantly more premium content. In this environment, more advertisers, such as The Hershey Company, are shifting more of their campaigns to The Trade Desk. In fact, Hershey has consolidated much of its programmatic advertising on our platform. They have gone from several DSPs down to just one, which gives them simpler line of sight into the effectiveness of their programmatic campaigns. They share our view that data analytics, transparency, and user experience are critical to driving greater effectiveness. As they emphasize TV across all their campaigns, they saw in The Trade Desk a partner who can help them drive greater relevance and success. They've built a small internal team at Hershey that manages their agencies and their relationship with us.
That team focuses on areas such as audience targeting, social media marketing, and data science. They are building in-house programmatic capability because like an increasing number of major advertisers, they want to get a much better handle on supply transparency. Which brings me to another point that I want to make. Since our last report, we have started to see a lot more regulatory focus on big tech, and much of it has to do with how they manage their ad platforms and the consumer data that they manage through it. Many of you have asked what this means for our industry and for our company.
To answer that, I want to take a step back and try to give you some context as to why we do what we do, because I don't think we've talked enough about this or taken enough credit for the pressure we are putting on the advertising ecosystem to be more transparent about data, pricing, and value. As I mentioned before, the total addressable market for advertising is on a growth march towards $1 trillion, driven by growing demand from advertisers to apply data to their ad campaigns and drive greater precision and relevance. This growing ad market is also driving the internet as we know it today. This includes not only content delivered over computers and phones, it's also internet-fueled TV that is right now the most exciting growth area within advertising. To sustain that great content, advertising must keep up and keep innovating.
In the near term, as CTV advertising grows rapidly, that means fewer ads with greater relevancy. The data that will fuel more relevant advertising won't come from TV viewing data alone. It will come from where you click on your many devices, what other content you interact with, and an understanding of what else you are interested in. Let me give you a sense of how we approach data and how we put it to work, and a couple of the initiatives that we have pioneered in our industry to improve transparency, confidence, and the power of data analytics. We have a very consumer-friendly position on data. We do not transact in directly identifiable data. We don't want the data that comes from owning a search engine, and we don't want the personal information and social graph that comes from operating the social network. We don't need it.
This is a huge distinction for us when compared to the walled gardens, who are under huge scrutiny on this issue. We don't know your name, your address, or the details of your health. We know about the products you are interested in and the content you tend to like. We can put that data to work to help our advertisers build an understanding of the kind of advertising that should be relevant to you. All of this data is anonymized. Our acquisition of Adbrain has been hugely important in our ability to advance this kind of work. Adbrain's AI helps us integrate insights across environments in a way that protects consumer data by using only anonymized information. This is a differentiator for our advertisers who are hyper-focused on protecting the integrity of their consumer relationships, which sometimes go back decades. Let's look a little closer.
Because of our leadership position on the demand side of advertising, advertisers also expect us to leverage that position to apply pressure to the supply side of ad tech. Over the last two or three years, we have done exactly that, often behind the scenes. In doing so, we've used our strong position as the leader in the open internet to make the internet better. As we do this, our focus is on improving the ad ecosystem for all players so that everyone benefits. We firmly believe that the more transparent the ecosystem is, the more confidence advertisers will have in it, and the more demand it will drive. That's why we developed and then gave away our Unified ID solution. As the largest independent demand-side platform, we have a massive cookie footprint. Leveraging this scale through Unified ID radically improves match rates across the ecosystem.
This is true even when the transactions aren't on our platform. Both demand and supply-side companies can use the Unified ID to drive a much more transparent view of users across the ecosystem. This is all done anonymously with no directly identifiable information. We have seen match rates go from what was considered decent by industry standards, around 50% to 60%, to close to 100% for those who use Unified ID. On the supply side, we have seen CPMs more than double in some cases. Every day, more and more companies are deploying it. It's great for advertisers who have a much cleaner view of their market. It's great for publishers who can drive greater value. It's great for other DSPs, as it takes one of their major challenges and radically simplifies it. Of course, it's great for us.
While it may seem counterintuitive to create this technology and then give it away, we believe that it helps create a growing competitive market, and we believe we will take an outsized share of that market. Another major initiative on the supply side has been our partnership with White Ops to scrub all ad inventory for fraud. Three years ago, if you were at any digital advertising industry conference or event, the number one issue was fraud. Advertisers were worried that they would inadvertently buy fraudulent ad spots. The problem at that time was that those companies who claimed to help advertisers manage fraud were really just taxing it. Their vested interest was not in preventing fraud itself, but taxing it, levying a fee for layering their technology onto every impression. Working with White Ops, we upended that model.
We wanted to scrub all inventory before it ever came to the buy side. They reduced their rates, we significantly increased their volume by insisting all of our suppliers work with them. In doing this, we put pressure on the entire ecosystem, big and small, to scrub for fraud. Our leverage was such that we could take this position, make it affordable, and make it pervasive. Today, this is not a major issue in our industry anymore. These are a couple of the examples of how we are working to improve the digital ad ecosystem for all parties. There are many others.
As I mentioned before, I don't think we've taken enough credit for the work we've done, but I think in the context of all the scrutiny of our industry right now, it's important to understand how the industry is changing, what's driving that, and how The Trade Desk is creating value. There are times when some supply-side players get uncomfortable with these efforts to improve transparency and efficiency. That's because many of them are trying to protect their business models that drive high margins while providing comparatively little value. This is the natural cycle of a maturing industry. Transparency will force price and value discovery, and every market participant will ultimately have to account for the value they provide and the prices they charge. In this climate, I don't think it's coincidental that our business continues to thrive and more advertisers standardize on our platform.
They want to align themselves with a partner who ensures they can act with transparency, objectivity, and integrity. As I've stated many times before, we believe our business model is exceptional. We have the luxury to not have to choose between growth and profitability. We are doing both, and we have pretty much done so since the beginning. In Q2, our financial performance, and particularly our $58 million of adjusted EBITDA, was significantly better than we had estimated. I want to give a little context. We've increased our operating spend by more than $41 million year-over-year in Q2. This means our investment was up 50% over the previous year. Even at that rate, we couldn't invest fast enough. Of course, part of that is our investment discipline. We invest quickly but carefully, always focused on the needs of our customers worldwide.
We are also careful to ensure that we maintain our culture, which we believe is key to our success. In Q2, some of the hiring and investments we did not complete are already happening in Q3. As a result, we expect adjusted EBITDA to be $45 million in Q3. If we expect additional investments will lead to an outsized return, we will invest as fast as we can. That's how we operate. We consider this a great position of strength. We can produce EBITDA margins that meet or beat most other SaaS companies, including those much bigger than us. Hopefully you found this context useful as you think about the nature of the market that we operate in and why we believe we will continue to grow share in the fastest-growing part of it, data-driven advertising.
Our focus is on advertisers and agencies, our commitment to the open internet and everything that means in terms of transparency. Our pragmatic approach to rapidly investing our profits in future innovation means that we will accelerate our leadership in this market. With the biggest shift in media and advertising we've seen in a generation now underway. We are in a great position to grow share moving forward. I'm going to turn the time over to Paul to discuss our financials.
Thanks, Jeff. Good afternoon, everyone. As you've seen in our press release, Q2 was another outstanding quarter. Revenue increased 42% year-over-year and accelerated versus our Q1 growth of 41%. Adjusted EBITDA increased to $58 million and GAAP net income increased 45% to $28 million. We achieved this while we continued to invest aggressively in areas critical to our future growth, such as on our platform and adding engineering and sales talent. Revenue for the second quarter was $160 million, which was above our prior expectations and reflected increased spend by our existing customers, plus the addition of new customers and advertisers. For the quarter, approximately 89% of our second quarter gross spend came from existing customers who have been on our platform for over a year. Q2 marks the 22nd quarter in a row where customer retention was over 95%.
With the growth of our business, our operating expenses grew to $128 million. This increase year-over-year was due to sales in marketing and engineering as we continue to scale for future growth. The year-over-year increase also reflected higher G&A expenses, which takes into account stock-based compensation, and we expect G&A to moderate as a % of revenue in the back half of the year. Income tax was a $5.6 million expense, mainly due to the tax benefits associated with employee stock-based awards, the timing of which can be variable. GAAP net income was $27.8 million for Q2, or $0.58 per fully diluted share. Our adjusted net income was $45.6 million or $0.95 per fully diluted share, compared with adjusted net income of $27.2 million or $0.60 per share in the comparable period.
Adjusted EBITDA was $58 million, with a corresponding margin of 36.2% of revenue during Q2. The increase in adjusted EBITDA dollars reflects the strong growth of our top line, offset by our increasing investments in product, people, global expansion, and corporate expenses. Net cash provided by operating activities was $11 million in Q2, and our trailing 12 months of operating cash flow and free cash flow were $95 million and $60 million, respectively. We continue to have zero debt on our balance sheet and our total cash equivalents and short-term investments exiting the quarter was $231 million. Our DSOs at the end of Q2 were 100 days, an increase of one day from the same period a year ago. DPOs for Q2 were 81 days, a decrease of one day from the same period a year ago.
For Q3 of 2019, we are expecting revenue of $163 million and adjusted EBITDA of $45 million. For the full year 2019, we now expect revenue to be at least $653 million, revised upward from $645 million last quarter. Adjusted EBITDA is now expected to be $201 million, also revised upward from last quarter, or about 31% of revenue. I will now turn the call back over to Jeff for final comments and of course, Q&A. Jeff?
Thanks, Paul. Q2 was another very encouraging quarter for The Trade Desk as we continue to see our strategy and execution pay off as more advertisers commit their budgets to us. We exceeded our expectations for the quarter and are raising them for the year. The fundamentals of our business are solid and we continue to scale our business across markets and channels. As the worldwide advertising market moves towards $1 trillion in a few years, we are well positioned to win a lot of the programmatic portion of that market, the fastest growing segment. We invested early in key markets and channels, and while we are pleased with our initial gains, we see far more upside yet to come. That concludes our prepared remarks. Operator, let's open it up for questions.
At this time, we will be conducting the question and answer session. If you would like to ask a question, please press star one on your telephone keypad. A confirmation tone will indicate your line is in a question queue. You may press star two if you would like to remove your question from the queue. For participants using speaker equipment, it may be necessary to pick up your handset before pressing the star keys. One moment please while we poll for questions. Our first question comes from the line of Shyam Patil of Susquehanna International Group. Please proceed with your question.
Hey, guys. Congrats on a great quarter. Jeff, I wanted to ask about the Amazon partnership. As you mentioned in your prepared remarks, it seems like it's a big step for the industry. Can you just talk a little bit more about what you think it means for the industry? I know it's super early, but how do you think about the opportunity for The Trade Desk? Thank you.
Awesome. Thanks, Shyam. First, let me talk about Amazon just generally before I talk about the deal. In talking about Amazon and what they're up to, a lot of people have asked, what are they doing and what does third-party apps mean and what is Amazon Publisher Services? My understanding of Amazon Publisher Services is the primary thing that they're monetizing is Amazon Fire, which is the sticks and the hockey pucks. Inside of that, of course, you have Amazon's own app where you watch Amazon Prime Video, and then you have Netflix, and you'll have YouTube. You have a lot of third-party apps where there's ads.
They're going to those third party apps, whether that's Sony Crackle or CBS or anybody else, and saying, "We can help you monetize that." Doing it in a very different way, which is why I'm so excited about the deal, and let me talk about what this means for us. Let me just summarize the deal. First, Amazon is joining the open internet in the sense that they're using their ID to make it possible for us to measure what's happening in that ecosystem that I just described, as well as on the rest of the internet. That makes it so measurement is meaningfully better than what we would get inside of any walled garden. Secondly, they're being more aggressive in economics, from what I can tell, than what anybody else has been on the sell side for connected TV.
What this does for content owners is, if you rewind two years and you're looking at this through the lens of a content owner, an owner who's 90% of their revenue comes through linear television, and the money that you have to divide with somebody like YouTube versus the money you have to divide with an MVPD partner is roughly the same, then you're in no rush to move to digital. In fact, digital just represents risk for you. If you make targeting better because of the use of that ID, and so you get the efficiencies that digital can provide, and then you also get the chance to keep more of a dollar. You get a greater percentage of the deal. It becomes economically irrational to not race towards digital.
What I think is especially significant about this deal is that it just became more incentivizing for content owners to move into digital. It also gives us a little bit of insight as to what Amazon might be thinking in terms of overall strategy, which is to be way more focused on the sell side when it comes to everything that's not on amazon.com. I think it's a game changer for TV. I think it's one of, if not the most significant deal that we've done in television to date. We hope it's the first of many, both with them and with others, and really excited about the opportunity that it represents.
Thanks, Shyam. Next question, Tim.
Our next question comes from the line of Youssef Squali of SunTrust Robinson Humphrey. Please proceed with your question.
Hi, this is Sagar for Youssef. Wanted to ask a couple of more questions about the Amazon partnership. First question is, do you have access to Amazon's first party shopping data for targeting? Second is, how do you compete with Amazon's own DSP?
Great. First one's easy. No. Amazon would never want to do that because of the risk that that would represent in terms of data security. They'll continue, I suspect, to do that on their own. To answer your first question is no, we don't. Secondly, in terms of how we compete with Amazon's DSP, I'd like to take a step back and just talk strategically for a second. If you look at this through the lens of some of the biggest advertisers in the world that are CPG companies, Amazon as a distributor has more power and to them can be a little bit more scary than any distributor that they've ever had. If you, for instance, compare them to Walmart a decade or two ago, while Walmart may have been the most powerful distributor they had then, Amazon has more power than they did today.
That's for a number of reasons, in part because Amazon has become just such an amazing retail engine that has done lots of consolidating, made it possible for other merchants to come onto their platform. Also because Amazon has gotten into so many other businesses, including selling CPG products themselves. The conflict they have with those companies can create some pause.
When you add the fact that AWS often stores the data for all of these companies, I think it's a really hard pitch for Amazon to go to a CPG company or really most of the biggest advertisers in the world and say, "We know you give us a lot of money and you trust us a lot to do all the distribution on our site, but we would also like to ask you to give us all of your marketing budget to do all of the spend off of amazon.com." They have, in that sense, a bigger objectivity problem than any company in the world, where in effect they're saying, "Trust me with your money, trust me with your data, and trust me with your entire marketing spend." I think what I anticipate is that the core of Amazon's buy side efforts will be to monetize amazon.com.
While they do have a DSP that's helping do some other things today, I don't expect that to be the core of what they're doing. This does shed more light on the significance of the sell side approach that they're taking in Amazon Publisher Services, where I anticipate that we'll be partners with them for a very long time. When you look at sort of the strategic hand that Amazon has dealt, I think it makes a ton of sense that while we do compete with Amazon DSP, the much more significant headline is that we're partnering with Amazon Publisher Services and really excited about what that means for the future.
Thanks, Sagar. Next question, Tim.
Our next question comes from the line of Michael Levine of Pivotal Research. Please proceed with your question.
Congrats again on the quarter, Jeff. Terrific acceleration. Wonderful to hear the detail about Amazon, but one of the other things that was interesting to see come out this week was about the partnership with Disney and also in the context of their basically bundling Disney+, ESPN+, and Hulu ad-supported at probably a much lower price point than I think a lot investors had expected. Love to hear your thoughts about that and how impactful you think it could be to the business.
Fantastic. As you know, Disney has been so aggressive in the last 18 months and it is actually so exciting to watch. A few years ago, I was really touting the forward thinking of AT&T, and I feel just so excited by what Disney is doing as a partner and as a consumer as well. I think the new bundle they have coming out with Disney+ and ESPN+ and Hulu+ for ads is really great, and I think is going to be successful. In about that same time, about 18 months ago, they approached us and we just started working together. We learned something in those discussions, which is that more than 50% of the views that they're getting on their content is coming from connected devices, and they wanted to figure out a better way to monetize that.
As somebody who's watched them for a long time as an investor in their business, I've watched them try to figure out what do you do with ESPN in a world where cord cutting is happening, and especially with just the business model that they had for ESPN, say a decade ago compared with today. It's actually why I am so excited about this partnership because one, we anticipate getting access to a significant amount of inventory as they become more digital. Also, I think programmatic is better suited for live events than any other way of monetizing. If you think about it, when a game goes into extra innings, it's not really conducive to planning months in advance as to how many ads you're going to see. Most advertisers aren't going to spend a bunch of time thinking about, well, what if it goes into 13 innings?
Instead, you make your plan, often what those companies have had to do in the past is give away those ads, the extra innings for free. Instead what they can do is, in real-time, check what demand is out there and also make certain that those ads are relevant and not overly repetitive like they often are in those situations. You make more money, make a better user experience. It's really critical when more and more of the views are coming online, and the only way to support that content with the optimal user experience of today is to welcome programmatic demand so that you get higher CPMs on each of those ad views. It's the reason why I think us and Disney are strategically sort of stuck with each other. We have a fantastic partnership.
It's not because we don't love working with each other. We absolutely love working with them. It is strategically obvious that product, live content, and the amazing offering that they have is going to continue to grow together. Our great partnership with Hulu, I think, is just indicative of what's to come. A lot of bullishness on our partnership with Disney.
Great.
Thanks so much, Michael. Tim, next question.
Our next question comes from the line of Brian Schwartz of Oppenheimer. Please proceed with your question.
Thank you, Jeff. Congratulations, too. It's just great to see the company executing so well here. I was hoping maybe we could switch subjects over to what's happening at Google. Can you please provide an update on any impact or opportunities you're seeing or expecting to see from the Chrome browser changes? Separately, you talked a lot about the Amazon Transparency Initiative. Can you also provide an update on the Transparency Initiative that you have been working on with Google for some time? Thanks.
You bet. Thanks for the compliments, Brian. Yeah, happy to give comments on Chrome. Just a reminder that Chrome has roughly 60% of the browsing market. I talked, I think it was last quarter or the quarter before, at length about their initiatives, including SameSite. I'm one that wants to really sing the praises of Google. I know sometimes because we compete with some divisions of Google, because they're so big, a lot of people are quick to pick on them at times, at least the parts that we compete with, we've talked about. What they're doing in Chrome, I think, is exactly what the industry needs, and I think they should be praised for what they're doing. They're making the Internet better, and they're making privacy better. I think it's a positive change for us and it's good for the industry.
I think the fact that it is positive for us and good for the industry is indicative of how we've aligned well with just, if you do the right thing, then you're going to be set up well for the long term. It's really great to see Google, under the pressure that they are, sort of reaching the same conclusion. I'm really optimistic about it. They haven't been specific about when they'll roll it out. I think they just want to make sure that they dot every I and cross every T, which is exactly what I would do if I were them, too. I suspect we'll be well into 2020 before we see the actual changes implemented. There's been little more discussion than just sort of coming soon to a theater near you. Excited about that.
Can you remind me, Brian, of the second part of your question?
Yeah. It was just on the transparency initiatives I think you've been working on for some time here with Google. I know you talked a lot about where Amazon's heading on transparency, and was just wondering if you can provide an update on what you're doing with Google on the transparency initiatives. Thanks.
You bet. I know there's been a lot in our trade publications talking about just our desire to put a lot of pressure on the sell side to just be more transparent. Essentially, the way that I look at it, most advertisers, when it comes to digital, have two options, two types of options. One is walled gardens, which are very good at making it easy to spend your money. The other is the open internet, where you have more data, more transparency, you have more optionality, you have more power, but it's more complicated. Sometimes what has happened to the open internet is because of that complexity, some players have used that as an opportunity to hide, to hide margin, to charge more than they add in value.
I actually think that has been the thing that has held back the open internet while the walled gardens have grown so aggressively and so impressively. I view our role as leading the open internet. What that means is that we have to go to all of our suppliers and say, "We expect you to add more value than you extract." As we get bigger, we're going to demand that we have more transparency on what we buy. We are in this really strong position that because there is so much media out there, and because the world of media and the internet is producing so much content so fast, we are in the power position, if you will, in that we are able to be picky. We get to choose what we're going to buy and what we're not.
We just demand a lot of our suppliers. Those that are transparent and open and honest and give us a fair environment to participate, we spend way more with them, and we're growing very aggressively. Those that aren't, we don't grow as aggressively. I think that approach is putting pressure on a number of players on the sell side, and that is a great thing for our industry. It is great for the internet. It's great for consumers because that will make a better internet. We're on a mission to do it. We'll keep doing that. It's something that we spend a lot of time talking about with Google. Overall, Google is interested in the exact same thing.
It is good for our industry to continue to work for a better internet, and I'm really proud of all the efforts that we've done to date on that front.
Thank you.
Thank you.
Next question, Tim?
Our next question comes from the line of Nick Jones of Citi. Please proceed with your question.
Hi. Thanks for taking my question. I just wanted to touch base on the Unified ID. How would you try to benchmark adoption of this, and what kind of pushback are you getting from various players who don't want to adopt the ID?
Yeah. I don't know that we've been public yet about just the adoption numbers, but I'll tell you just in terms of commentary, the number of companies that have adopted it is overwhelming. I would say that it's very difficult to be a top 10 supplier for us in any channel if you're not adopting it in some way. I do think it's been massively successful. I think at this stage it's nearing inevitability in terms of its ubiquity. We've seen adoption all over the world, so it's not isolated to North America. It's happening all over the world. It is a better way to operate the internet. It is better for users. One thing that I don't think I've said enough is that we basically have used the standard with Unified ID that operates in mobile with IDFA.
This means that our privacy standard or the way that we're creating this ID and the degree of privacy that this provides to consumers is just like IDFA, which is, of course, what Apple uses to operate its ecosystem and what the entire mobile community uses. This ID, while the way to sort of go to market or get its adoption has been unique, in effect, especially to the consumer, it looks just like IDFA, which is part of the reason why we're so proud of what we have been able to accomplish in getting adoption to Unified ID. I expect it to be one of the few standards for IDs and therefore targeting on the internet as we march towards the future.
I'm sure at some point we'll give updates in terms of where adoptions happen, but it's happened on the buy side, on the sell side, on the data side, all over the internet, all over the world.
Thanks so much, Nick. Next question, Tim?
Our next question comes from the line of Mark Zgutowicz of Rosenblatt. Please proceed with your question.
Thank you. Jeff, thanks for all the detail on the Amazon partnership. It's helpful. I was just wondering if you could separate the anonymized ID, which you're getting from Amazon, from the shopper data, which you're obviously not. I'm just trying to understand how the advertiser sort of closes the loop, if you will, with attribution. Without having the shopper data, do they go back to Amazon to close that loop? How do you close the entire attribution loop within your platform?
Yeah. This is actually something that's, I think, important to understand about walled gardens is that when you operate a walled garden, you can't really provide attribution for the rest of the internet. You simply say, "This is how many people I touched in my walled garden, and this is how many conversions came from those people that we touched." You can't really make many arguments about causation or correlation because you're just measuring in isolation. That is sort of what you sign up for when you create a walled garden. The reason you have to create a walled garden is so that you can use powerful data that isn't owned by the advertiser.
In order to use the data that another company like an Amazon or like a Google or like a Facebook have, you have to operate a walled garden and to make certain that data never leaves that garden. Otherwise, you run the risk of sharing that. That includes closing the loop on attribution. We're not able to say how did that shopper ad on Amazon actually convert when they went and bought on the rest of the internet. We can't weigh in on the effect they had on the open internet, nor can we weigh in on the open internet's effect on sort of bringing them into the walled garden and participating in sort of the supply chain once you get onto Amazon, because Amazon's not going to provide that outside of their own shopper experience in order to utilize that data.
Walled gardens, by their very nature, make attribution hard for everybody else, and that is why it is so important for more and more advertisers to be spending on the open internet. It is why, as connected TV continues to grow, and it will be extremely fragmented because TV has always been extremely fragmented, that the bias will be to the open internet. As an advertiser sits there, says they want to spend in connected TV, the first place to start is on the open internet. Then you'll give the leftovers to walled gardens because, one, that's where all the premium content is. Two, I expect that's where all the volume will be too of users because they'll spend more time there than they will on user-generated content, even though that's also a huge pool.
Just by their very nature, you're not able to connect the dots between walled gardens because that is what makes them walled.
Thanks so much, Mark. Tim, next question.
Our next question comes from the line of Mark Mahaney of RBC Capital Markets. Please proceed with your question.
Hi. Thank you for taking my question. This is Benjamin Wheeler for Mark. Two, if I could, just firstly on China, if we could talk about that. I understand that it's not going to be material for a while, but you are partnered with three of the largest Chinese media companies. I'm just wondering about the short-term limits to growth. Does that have more to do with, one, a lack of demand from the global advertising partners you have? Two, these media partners limiting your access to their inventory, or three, you kind of self-capping the spend that you're bringing onto them in order to maintain positive relations with these partners?
Just secondly, if we go back to Amazon TV, is there any way you could kind of quantify in relative terms just the incremental amount of CTV inventory that this Amazon TV deal brings to you, like relative to the CTV inventory you had prior? Understanding that you do have access to 100% of those impressions. Thank you.
Sure. As it relates to China, as you point out, it is very early stage for us. I wouldn't characterize anything that we've experienced today as limiting or anything like that. We started a partnership with Baidu, Alibaba, and Tencent. We're growing that. We are seeing growth every single month, and we're going slow and being very deliberate. We, of course, are going to continue to roll out in a bigger and bigger way, and there's more to come before the end of the year on just some of our plans there. Things in China are going well, and we're playing the long game. There's nothing that's happened in China that I view as a setback.
It's us just staying the course and making certain that we tap into the media market that's about half the size of the United States, but growing at double the pace, and is the second-largest media market in the world. Perhaps the biggest opportunity geographically that we'll ever see again. Huge opportunity for us that we want to make certain that we don't mess up by being hasty or by burning bridges, and I'm really excited about what we've done so far. Second, to your question about incremental inventory. I can't comment on how much it will add to our total inventory, but I can reiterate that we will see 100% of every impression on Amazon. We will also compete with all other demand, and that includes anything that the Amazon sales team sells. Sorry, messed that sentence up. Let me say that again.
We will compete with all other demand, and that includes competing with anything else that the Amazon sales team sells. That means they haven't carved out the 20% or 30% of inventory that they then sort of give it a carriage deal that is separate, and then make a much higher margin on that than they would on the rest. They're putting it all together, and that's all competing. We're just as eligible to win any impression as anybody else is, including Amazon. Given that that represents 100% of every impression on those third-party apps, that is without a doubt significant.
Cool. Thanks, Ben. Tim, next question.
Our next question comes from the line of Robert Coolbrith of Wells Fargo. Please proceed with your question.
Great. Thanks for taking our questions. Jeff, wanted to go back to your comments earlier about programmatic so far as a dress rehearsal for CTV and where you're going. Programmatic was initially sort of a non-premium opportunity, which ended up, in some cases, impacting premium pricing integrity. Given that, just wondering if you could talk a little bit about the PMP model and how some of the premium publishers and networks you're partnering with now are getting comfortable that we won't have sort of a repeat of that history with the creation of sales channel conflict or impacting pricing integrity. Then just one quick one on Amazon, just wondering if that deal could potentially extend to their ad-supported IMDb TV product. Thanks.
You bet. Let me answer the last one first. It doesn't today, but I see no reason why it couldn't in the future. There's no plans for that that we have talked about today. On the pricing, I love this question, first of all. You're absolutely right, that programmatic was born out of sort of a bottoms-up growth or evolution in the sense that when we first started monetizing display 10 to 15 years ago, we were taking the remnant, the leftovers, and running an auction for it.
That was really scary for companies like MSN and Yahoo, because they were saying, "Uh-oh, I make 90% of my revenue on 10% of the impressions, and I don't want necessarily everybody to know that I have more supply than I do demand." programmatic shined a light of transparency on things, and it did make prices go down temporarily. What's happened over the last few years is that that's helped create better or fewer ads per page. It's helped make the internet overall better, just economics getting more effective. especially what's happened in connected TV has been exactly the opposite, which is exactly what you want to do. programmatic, in a way, should have never started with display, because when you have more supply than you do demand, an auction is not the best way to monetize that.
When you have more demand than you have supply, which is exactly what's happening in connected TV, because we're taking the number of ads in a commercial break and reducing them. Because cord cutting has accelerated adoption, we're getting record demand for connected TV. That's made it possible when you join that with the increased efficacy that comes from being data-driven instead of just sort of the spray and pray model that most linear is. Most CPMs have gone up pretty meaningfully. It's not uncommon to see a $10 CPM in linear TV and see the same content monetized at a $30, $40, or $50 CPM in digital.
Programmatic is exactly what that needs, and I don't think that that's going to create any pricing erosion. What that has done as well is created a desire for both the buy side and the sell side to get together and create sort of a faster distribution channel. How can we reduce the number of hops, the number of taxes that happen in the middle by creating private marketplaces or PMPs or private deals, so that you can reduce the tax that the people in the middle take? That is something that we're working really hard on with suppliers. We want to make certain that as much money as possible, is kept by the ends, if you will, advertisers and publishers, so that they continue to do really well.
We think that's one of the things that we're extremely good at, which is adding more value than we extract in the middle and making certain that everybody along the way is earning their keep. The evolution you're describing, it's a really important moment for those macroeconomic forces to be changing media, because what audio and Connected TV have is a secular tailwind that display and native never had.
Thanks a lot, Rob. Tim, we have time for one more question, please.
Our final question comes from the line of Vasily Karasyov of Cannonball Research. Please proceed with your question.
Thank you. Good afternoon. Jeff, I wanted to ask you about a topic that we don't talk about much on these calls. That's audio spend. I think in your prepared remarks, you said that you think it's the best value in advertising right now. Looking at the growth rates, it's growing a little faster than connected TV. I believe that's from a bigger base. My question to you, can you talk a little bit more about that? Why do you still think that connected TV is a bigger opportunity for you here? Maybe talk a little more about how audio is growing, how the adoption by advertisers is going. How far do you think this growth will extend? Anything that you think is worth sharing with us would be interesting because it is growing fast, and we don't talk about it as much.
I'm so glad you've asked this question, in part because we get to rectify exactly what you point out, which is that we have not given audio, and that includes Spotify and Pandora, the word count that they deserve. We did mention in our prepared remarks that audio grew by 3x for the third quarter in a row, which given how long we've been in audio, is unbelievable. I am blown away by that number. I'm looking at the trajectory. I'm just so excited by what's happening. Part of the reason why I call it such a great value is when you compare it to some of the other forms of advertising, including other forms of programmatic, I don't think that there's a form of advertising that captures your attention much better than audio.
You think about the way most of us consume Spotify, for instance, is you're in your car or you're exercising, you're walking around, you're way less likely to skip an ad than if you're watching YouTube or something where you're sitting in front of a computer. You're also more likely to give it more of your attention in some cases. When you look at the fact that the rates of those audio ads are exponentially or a fraction of the price of more premium television content, I think it represents a big opportunity. When you look at what's happening in podcasting and some of the other ways that those companies are starting to monetize, it's no wonder that we're growing at the pace that we are, and I think it continues to represent a huge opportunity.
One thing that I also think is just worth pointing out, because I think we can learn from it as we look at other markets, and that includes other geographical markets. Because some of the markets we're just about to go into are smaller markets than where we've entered before. They're not top 10 media markets. They have something in common with audio, which is that audio is operating on such tight margins that they can't really afford to do what radio used to, which is pound the pavement looking for dollars and selling to SMBs. You have to sell in a more automated way, and you have to reduce your cost of sales in order to be competitive.
As Spotify and Pandora are competing around the world to get as much listening time as possible, they have to operate really efficient engines, and there's nothing that helps the efficiency of the distribution of ads more than programmatic advertising. It's another one of those places where we're sort of destined to work together, and I'm just so excited by what we're doing together.
Thank you, Vasily.
There are no further questions over the audio portion of the conference. This does conclude today's conference. Thank you for your participation. You may disconnect your lines at this time. Have a wonderful rest of your day.
Thank you.