It's my pleasure to turn the floor over to your host, Chris Toth of Investor Relations. Sir, the floor is yours.
Thank you, operator. Hello, and good afternoon. Welcome to The Trade Desk fourth quarter and full year 2016 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 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 will be describing will be forward-looking statements that are dependent on certain risks and uncertainties. I encourage you 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 non-GAAP to GAAP measures can be found in our earnings press release.
We believe providing non-GAAP measures combined with our GAAP results provide a more meaningful representation regarding the company's operational performance. I will now turn the call over to founder and CEO, Jeff Green. Jeff?
Thanks, Chris, and good afternoon, and thanks to everyone joining us today. 2016 was a massive year for The Trade Desk. We surpassed $1 billion in total spend, resulting in record revenue of $203 million, an increase of 78% compared with a year ago. In 2016, according to PwC, digital advertising grew by about 14%, while programmatic, according to the IAB, grew 19%. We grew about four times that while producing over 30% adjusted EBITDA margins. 2016 was also the year that our omni-channel offering became more mobile and more video focused than display, and we have significant momentum heading into 2017. I recently spent time with a CEO of an important agency to us.
It was amazing to hear them say point blank, right in front of one of their largest clients, one of the biggest consumer brands in the world, that The Trade Desk is the best demand side platform in the market today and is easier to work with than any partner they've ever had in digital. That was quite an amazing and humbling endorsement, and it really illustrates what we are trying to achieve at this company. I was just as excited when a much smaller and more nascent technology company came to our corporate headquarters in Ventura, California just a few weeks ago. The CEO explained that he is trying to make programmatic easier for the little guy. His mission is to make it so SMB advertisers can buy the rest of the internet the same way they buy Facebook.
He was so excited to explain that his company is growing 60% every month. He's adding hundreds of advertisers every month. He made it clear his business couldn't exist without The Trade Desk platform and especially the expressive APIs that power his entire business. With those two business and hundreds of others, our mission is to change the way all of advertising is bought, an industry which has now reached $640 billion, according to the latest research from IDC. Unlike most technology companies, we are not trying to disrupt an industry. We are trying to enable. In 2016, we powered more DSPs, more data companies, more agencies, more technology companies, and enabled more local aggregators than we ever have before.
While we're known for powering CPG companies and automakers and our especially close partnerships with the big agencies, we are also powering more and more well-rounded and diverse group of enablers than we ever have before. Many of the biggest players in digital advertising and some of the biggest internet destinations claim that their success is the result of them both owning and selling media. We think one of our strongest strategic assets is the objectivity that comes with not owning media. We objectively help buyers look for value and are holistic about their media buying. Today, we are the only purely independent, scaled, omni-channel DSP in the world. Our independence and allegiance to buyers makes it possible to objectively buy media from great companies like CBS, Google, The New York Times, and thousands of others.
I'm going to use the time that we have today the same way that I plan to on most of our future earnings calls, to expand upon our strategy. First, I want to talk about our goals and priorities, in particular, gaining share over maximizing profit in the near term. Second, I'll highlight some of the growth opportunities in our mobile and video channels, as well as our global expansion. Then third, I'll focus on having the best technology platform for media buyers. Finally, I'll talk about how we maintain a strong business model that can build The Trade Desk into one of the most valuable companies in advertising and the ad tech industry. First, let's talk about our goals and priorities. We are in this for the long haul.
We started seven years ago, and we think there are adequate levels of price discovery on about 2% of that $650 billion, a subset of the digital transactions. It's our view that the other 98% will eventually be transacted digitally with higher levels of price discovery than they are today. Our highest priority is to grab land from the remaining 98%. Winning more than our share of the new dollars coming into programmatic is paramount to our strategy. This is why we will regularly highlight that we're growing so much faster than the industry. We do not aim to maximize profitability today at the expense of gaining share, but we also realize we are in a very unique and fortunate position where we can rapidly grow while still remaining very profitable. We are not tempering our growth to keep our EBITDA high. Our growth is not restrained by capital requirements.
We control our growth due to our desire to preserve our culture and maintain our customer retention rate. We're focused on the long term, which is why we are making significant investments in 2017. This year, we expect to keep our EBITDA margin percentages in line with some of the most scaled and most successful SaaS companies in the world while growing faster than most, if not even all of them. Let's discuss our biggest growth opportunities, mobile video and global. Our clients coordinate programmatic dollars across multiple channels and thousands of sites, channels, stations for maximum efficiency. They do so holistically, which is the only way marketers can stay competitive. We think the two most important channels of the future are in mobile and video.
For mobile, there are currently about three billion smartphone users worldwide, and as a result, nearly all of our customers have moved to some amount of mobile advertising. Most were spending nothing with us three years ago in mobile. In Q4, all mobile, including in-app mobile video, represented about a third of our business and grew 135% in 2016, while mobile advertising only grew by an estimated 46%, according to ZenithOptimedia. At 3x the growth of the industry, we are gaining share and our customers can reach billions more people than on desktops alone, and we expect this trend to continue to grow around the world. Another encouraging fact, according to Zenith, is that mobile users spend more than 300% more time on their mobile devices than they do on desktops. In the coming years, this gap is expected to increase.
This is great news for the growth opportunity in mobile. Our platform is allowing ad buyers the opportunity to significantly increase their mobile engagement. Through our omni-channel offering, we can help brands and agencies measure the efficacy of their media decisions, including the cases where TV ads create awareness and the mobile ads facilitate the transaction. Nowhere is this one-two punch of video and mobile more pronounced than on Singles' Day in China. On November 11th, $17.8 billion were transacted in this 24-hour period, 82% of which was on a mobile device. Preceding the day, there is a nationally televised gala that is something like the Super Bowl and the Oscars rolled into one event. Advertisers push in offers, coupons, lock in pricing ahead of the big event to drive awareness and sales. The bulk of the transactions occur the next day on mobile devices.
While display is still an important opportunity and channel for us, mobile is clearly more important to our future. As connectivity and time on mobile increases, we believe mobile ads will also become higher quality and better integrated into the user experience than they are today. We also think mobile video is one of the most untapped opportunities in all of digital. With all of these forces coming together in mobile, we enable unbiased data-driven decisioning to more users and return a better ROI for our customers, not just on one site or on one app, but across the entire internet. Perhaps the only thing I'm more bullish about than mobile is the future of television. Let's talk about TV and video. We think about a few distinct categories in TV and video.
The first, there are web videos, which include premium content like full episode players from the major networks. There are connected TVs, which include the Rokus and Hulus and Apple TVs, and even the Sony BRAVIA of the world. There are the efforts to improve the transactions of linear TV, the third category. However, the future of TV is fully addressable over the internet, and that's where we are placing our biggest bets, in those first two categories. Here's why. The traditional TV model built on bundles, networks, set-top boxes, coaxial cables, and audience panels is under a lot of pressure. Netflix, Amazon, YouTube, and DVRs have changed the game and accelerated a market transformation. As consumers cut the cord and move their TV habits to the internet, every major player in television has been formulating and adapting their own strategies to this new reality.
In 2015, we saw major broadcast and media companies start to test new distribution platforms, and those tests left everyone underwhelmed. In 2016, they decided they needed to monetize their own assets, and we saw large broadcast and media companies either buy sell-side platforms or take action to control their own distribution, as exemplified by the AT&T-Time Warner deal. Now in 2017, we are starting to see this put into practice. This is why the independence of The Trade Desk is so valuable. As broadcast and media companies go to the agencies, which is where nearly all of the TV advertising is bought today, and they ask the question: where should I plug into the programmatic dollars?
The Trade Desk will benefit, as we already work directly with most of the agencies buying TV media, and we are the one self-serve platform that can holistically buy TV programmatically without any competing assets. While consumers love things like Netflix, the ad-funded channels on connected TV are growing exponentially faster than the no-ad offerings. The future of TV is over the internet, and it is ad-funded. As an industry, we have to show fewer ads, and we have to make them more relevant, and that is why we are so well-positioned and optimistic about the future of TV. We are frequently asked about how quickly traditional linear TV is going to transform into internet-driven, fully addressable TV. We think the AT&T acquisition of Time Warner is a bellwether as to how this transition is pacing.
When the deal was announced at the end of last year, the CEO of AT&T, Randall Stephenson, announced that their aim is to provide a bundle to consumers of their favorite channels at significantly less than what is on the market today. Initially, they talked about a $35 price tag. He acknowledged the critical ingredient to this working over the next several years is to utilize 5G technologies and use the internet to deliver all of the content. Everything would be on demand and menus that look more like Netflix than the current cable grids. In order to offer this 100 channels at a lower rate, his plan is to make it up in fully addressable advertising. He stated that his aim is to have this up and running in 2018.
A big $85 billion bet and a fast-approaching date make that an important indication of how quickly the media world is moving to addressable digital advertising, which provides a significantly better experience than what today's coaxial cable providers can provide. Already, leading connected TV companies like Roku and Hulu are embracing data-driven programmatic advertising with significantly higher CPMs, and we believe these are the early adopters. Approximately two-thirds of our customers spent on video in 2016, and our video spend increased by 140% year-over-year, and connected TV grew by 40x between January 2016 and December of 2016. Because of the sheer size of the market, this has the opportunity to move the needle unlike any of our other channels. We are the only scaled independent buyer of media in the video market without the conflicts of interest like the competitors with large TV and video assets.
The activity and enthusiasm, plus the ever-increasing flow of spend we are seeing make me very excited about the future of TV and video. Our objectivity and independence matter in every channel, but it is more strategically valuable to working in TV than in any channel. Next is global. When we talk about global, we are talking about getting Europe to adopt programmatic, to convert their existing advertising to more effective ways of transacting. However, the biggest opportunities in the world are in Asia. The same transitions are happening in Asia as are in Europe, but we predict adoption will be faster since GDP growth is faster there than anywhere else in the world. Today, Asia is nearly one-third of the total global ad spend, and the region is expected to see the fastest programmatic growth in the coming years.
eMarketer estimated that China grew programmatic 70% in 2016 compared to 19% in the U.S. In our Southeast Asia business, according to Magna Global, countries such as Indonesia and Vietnam are expected to grow programmatic ad spend six and eight x where it is today. We've had our offices in Japan, Korea, and Hong Kong, and Singapore for years. Last year, we announced that in the first half of this year, we would open offices in other spots in the world. We're pleased to announce that in Q4, we opened our Jakarta, Indonesia office, and so far in Q1, we have already opened our Paris office in France and our Madrid office in Spain. In 2016, our international markets represented about 10% of total spend but collectively grew about two times faster than the U.S.
Mobile, mobile video, audio, and native are significant drivers of revenue growth outside of the U.S. We are extremely bullish on growth in Asia, and we believe we are the first in the region with products specifically for each market and deep relationships with many of the large global agencies with a presence in Asia. Our independence is proving to be just as valuable in international markets. We are pleased to share with you today that we have recently entered into two large and meaningful inventory partnerships with two international companies, Ambient in Indonesia and Baidu in China. We are excited to be one of the few advertising technology companies in the world to partner with both Baidu and Google. We expect to announce other big partnerships like this one throughout the year.
While the move to programmatic in the United States has been quick, the industry still struggles with channel conflict. We don't see this happening in the international markets because there aren't large established sales teams and channels for advertising like in the U.S. We are seeing customers like American Express go all in for programmatic in places like Asia. I believe international markets will play a much bigger role in our business in 2017 than they ever have before. We are seeing this with rapidly growing content providers like Spotify, which has figured out that a price plus ad-funded revenue stream is the way to go. This represents a tremendous opportunity for us in our Asia growth curve, and we are very excited to work with these early adopters. Now I'd like to speak a little bit about our platform.
We are at the beginning of a massive change for our business and the state of our industry. In December 2014, we crossed the threshold of evaluating up to 1 million ad impressions per second, or QPS, also known as queries per second. Last quarter, we saw a peak of 5.7 million QPS. Programmatic did not grow over 500% in two years. That increase in auctions is caused because of the advent of header bidding. Header bidding is a term used to describe a technical implementation which simply creates competition and enables publishers to run multiple auctions for the same impression. Essentially, publishers figured out that header bidding, or this running multiple auctions for the same impression, was good for them. They can make more money. Advertisers now have to be more careful than ever about which ads they buy, even which auction they buy them in.
The value equation has changed. We evaluate and analyze over 400 billion+ ad opportunities every day in real time, and we help our customers carefully decide which ads are best for them. The barrier to entry for the competition goes up when more ad opportunities come into the digital transaction world because costs are incurred whenever you look at an ad opportunity and will make money only when you win. In 2017, we project QPS could increase upwards of 600 billion ad opportunities every day. This makes the race to profitability even farther away for anyone that's not growing faster than the industry. In Q4, we grew revenue 70% off of a much higher revenue base, while the industry only grew at roughly 19%, and our GAAP profitability grew by 81% year-over-year. The way the industry is shaping out, this provides us with a massive opportunity.
To continue to capture the opportunity, we need to build the best user experiences, we need to improve our platform, and continually develop and release innovative products to keep The Trade Desk far ahead of our competition. In 2016, our engineering team nearly doubled in size. In 2017, we plan to continue this growth and hire more engineers than any previous year while ensuring we operate with maximum efficiency. Internally, we're calling 2017 the year of development. This year, we will lead with product and product marketing, and we plan to roll out a brand-new user experience for our customers with a focus on workflow efficiency during the year. Our goal is to give people time back so they can optimize and better leverage our platform. Finally, I want to switch gears and talk about how we built a strong business model.
As I've stated before, we don't believe that growth has to come at the expense of profitability. We expect to do both and have since 2012, we just reported record profitability for the most recent quarter and year. In 2016, our financial performance was more than we hoped for. Even with our pedal-to-the-metal aggressive investments in mobile, video, and our global growth, as well as the investments to go public, we kept adjusted EBITDA margin at 32% and increased adjusted EBITDA dollars to $65 million, an increase of 66% over the prior year. In addition, free cash flow for the year was also positive, which marks a first for The Trade Desk. This is a testament to the value we provide to our customers and the leverage in our model. We continue to set a high bar.
The industry is continuing to see ad dollars shift to programmatic, our customers are expecting to spend significantly more with us than they did in 2016. For 2017, we expect gross spend on our platform to be over $1.45 billion and revenue to be about $270 million. In the year ahead, we view the aggressive investments we are making in high-growth areas such as mobile, video, and global as critical to grabbing share and deepening our engagement and strategic importance with our customers. This will cause our expenses to grow at a faster rate in 2017, as a result, we see our adjusted EBITDA margin at about 26%. We are not aiming to maximize profit this year, we believe we are doing the best thing for the growth of our business and the ultimate profitability over the long term.
Now I'm going to turn the call over to Rob to discuss our quarter in more detail.
Thanks, Jeff. Good afternoon, everyone. Our business continued its strong trajectory in the fourth quarter, and we ended 2016 with strong momentum. Total fourth-quarter revenue increased 70% year-over-year, led by our mobile in-app channel, which grew by over 400% on a year-over-year basis. Our focus in Q4 was to continue to improve our scale and deliver results for advertisers in our seasonally strongest quarter. We delivered on those goals, and one of the best indicators of this came during the holiday season, where we generated significantly more business from everyone, from large consumer product brands to smaller performance businesses. All of our customer groups use The Trade Desk platform as a core part of their holiday advertising through their agencies. One of the biggest highlights of the quarter was the increasing importance of mobile and video, as Jeff referred to earlier.
One example of this shift is one of our customers, which is a connected TV application customer, that through their agency spent $5 million on our platform in Q4, 96% of which was either mobile or video. We also had an API customer spend over a million and a half dollars in the fourth quarter by aggregating spend from over 400 SMB advertisers, 56% of that spend was on mobile. More and more advertisers of all sizes and across all industries are shifting significant spend into video and mobile. As I described last quarter, from an operational perspective, we really have three core priorities. Number one, that is to be our customers' independent, trusted advisor. Number two, focus on growing our omni-channel presence, and Number three, continuing to grow our international footprint. Starting with the first item, our goal is to be our customer's trusted advisor.
I want to emphasize that The Trade Desk is the only scaled self-service and dedicated buy-side platform in our industry, without any conflicts of interest. In 2016, we focused our hiring efforts in the first half of the year so that each employee was fully trained and ready to advise and contribute to our customers' success during the higher volume second half of the year. Our efforts paid off as we exited the year with 566 active customers, a continued customer retention rate over 95%, and an overall cohort growth of 71%. To help achieve these goals, we ended the year with 467 employees worldwide, with the vast majority of them being customer-facing. Our dedicated sales, account management, and trading teams work side by side with our customers to maximize the value that they derive from our platform as they log in every day.
As a result, our annualized revenue per employee totaled $435,000 for the full year 2016, which continues to stand out amongst other SaaS and ad tech industry peers. Annualized revenue per employee is a key metric that shows how our business model was designed to scale efficiently and that we continue to execute strongly. I want to focus on our commitment to growing our omni-channel presence, which greatly helps media buyers become more holistic in how they buy advertising. In 2015, display was nearly 60% of the total spend on our platform, but as we've developed more robust offerings across mobile, video, native, and audio channels, display is now less than 50% of our total spend.
The strong growth we have seen in areas such as mobile and video have enabled a higher level of coordination and strategic targeting advanced by proprietary data and technology, including both first-party and third-party data across all of these channels. It's making a real difference for our advertisers and agencies. Cross-device targeting is growing rapidly and has seen an almost 200% increase in the second half of the year when we compare that to the first half of 2016. As an example, last year, a large technology company, through their agency, used cross-device targeting as a way to increase retargeting scale and find the same users across multiple browsers and devices. As a result of using cross-device targeting on this campaign, the company saw a 63% increase in subscription conversions, which greatly increased their ad spend ROI.
Our video growth, which includes things like full episode players, connected TVs, set-top boxes, and web video, in total grew by 140% over the prior year. Connected TV, while still in its infancy, grew well over 100% in Q4 versus Q3, as advertisers realized the opportunity to utilize data-driven decisioning to target cord cutters on early programmatic video adopters like Roku, Apple TV, and PlayStation. Given that TV budgets are about half of the global advertising pie, this is only getting started, we continue to view this as one of the most promising areas of growth for our business. In the fourth quarter, we launched several more native inventory suppliers on our platform. Our team spent significant time training media buyers and agencies early in the year on how to incorporate native ad units into their ad campaign strategies.
As a result, we saw over 700% growth in native spend in Q4 as compared to Q3. Digital audio, that also remains one of our fastest-growing channels. It's still the early stages of audio, but adoption is growing more rapidly at this stage than most of the other channels we've introduced in the past. We are currently playing audio ads in more than 100 countries around the world, in large part due to Spotify at this stage, but we are also adding new audio partners such as Triton and AdsWizz in the coming year. One of the more interesting stories of the November election in the U.S. is that while one cannot bring signs and solicit voters near the polls, we were serving ads to Spotify users while they were in line to vote, illustrating the power of programmatic and cross-device targeting.
The third priority we're focused on is widening our geographic footprint to make sure we serve our customers locally in the markets that are important to them. In every quarter of 2016, our international spend growth percentage has outpaced that of the U.S. Exiting the year, our international business amounted to about 12% of our total billings. When you think about places like Japan, which is the third-largest media market in the world, and where we've been on the ground for over two years now, that's a market that has only about 5% programmatic adoption to date. There's just really a massive amount of opportunity for growth in Japan. Then there's Indonesia, where we just opened an office in Jakarta in Q4. Indonesia, which is the fifth-largest country in the world by population, has the fastest-growing middle class and a very concentrated set of publishers.
All of these characteristics make it very ripe for programmatic adoption to happen quickly. Those are just two examples of many international markets of size and scale, that we think have large potential going forward and will continue to drive our international growth in 2017. Overall, we continue to make real progress advancing our strategy and achieving our operational goals. Our revenue and key metrics are growing nicely, and as we look ahead to 2017, we're very confident in the direction of our business. Now I'm going to turn the call over to Paul to discuss our financials.
Thanks, Rob. Good afternoon, everyone. We are all really pleased with our Q4 performance against our key financial metrics, growing revenue 70% year-over-year, growing adjusted EBITDA 53% year-over-year, and GAAP net income 81% year-over-year, all while investing aggressively in areas critical to our future growth and positioning the company to maximize long-term financial returns. We ended the year with over $1 billion in spend on our platform from approximately $550 million a year ago. The increase in spend was driven by video and mobile video, which each grew 111% and 300% respectively. Even display, which is now less than 50% of our total business, grew 57% for the year, three times the estimated rate of industry growth. Revenue for the fourth quarter was $72 million, as previously mentioned, up 70% year-over-year.
This growth reflects both expansion of our share of spend by our existing customers and the addition of new customers. Approximately 87% of our fourth quarter gross spend came from existing customers. We define existing customers as those that have been with us for more than one year. On a yearly basis, revenue for the fiscal year 2016 was $203 million, up 78% year-over-year, with 91% of our year-to-date gross spend coming from existing customers. Our operating expenses increased in parallel with the growth of our business to $48 million in Q4 2016 from $25 million during the same period in 2015. The increase in operating expenses was primarily due to our increased investments in personnel, in stock-based compensation, and in our newer offices. Lastly, we recorded $12.7 million of income tax expense associated with our taxable income.
Our effective tax rate in 2016 was 53%, primarily due to the nondeductibility of warrant expense and the nondeductibility of stock compensation expense, both of which are non-cash. Absent any meaningful tax reform in Washington, we now expect that our tax rate will remain elevated throughout 2017. This is due to two factors. The first is the nondeductibility of stock compensation expense, which is increasing due to the IPO and, of course, our higher share price. The second is due to our international structure, where our international entities may still be operating in a loss position, but we don't get the deductibility of those losses in the U.S., which, of course, has the impact of making our effective tax rate appear high on the face of our income statement.
Should the new administration and Congress pass any corporate tax reform, The Trade Desk would be a huge beneficiary as we generate most of our profitability in the U.S. and have no NOLs, therefore pay corporate income taxes at the full statutory rates. One item I'm often asked about is stock compensation expense. During Q4, stock-based compensation was $4 million, an increase from prior quarters, primarily as a result of the company's new employee stock purchase plan. We're expecting to keep stock compensation expense to mid-single digits as a percentage of revenue and have budgeted about $13 million for 2017. GAAP net income was $10.3 million for the fourth quarter of 2016 or $0.24 per fully diluted share. For the year, GAAP net income was $20.5 million, an increase of 29% compared with 2015.
Our adjusted earnings per share was $0.33 per share for the fourth quarter, compared with $0.28 in the prior year. For the full year, our adjusted earnings per share was $0.89, up 46% compared with the prior year. We use adjusted EBITDA as a core metric for our business, we calculate our adjusted EBITDA by excluding stock compensation expense, and in the case of 2016, also the non-cash warrant expense. Adjusted EBITDA was $28.6 million with a corresponding margin of 39% of revenue during Q4, as compared with adjusted EBITDA of $18.7 million or 44% of revenue during the same time last year. The increase reflects growth of our top line, of course, our operating leverage offset by our increasing investments in product, people, and global expansion. For the full year, our adjusted EBITDA was $65 million or 32% of revenue in 2016.
Adjusted EBITDA increased 67% from $39 million in the prior year, reflecting the leverage in our model, of course, offset by the investments we're making in our people and technology in the second half of the year. Net cash provided by operating activities was $75 million in 2016, compared with net cash used in operating activities of $37 million in 2015. This represents a $112 million positive swing in our working capital situation as we have scaled significantly. During the quarter, we paid down $25 million in debt and closed the year with $133 million in cash. We also have an additional $95 million available on our revolver. Our operating cash flow in Q4 exceeded our internal projections and was the result of initiatives to better align our DSOs and DPOs, as well as the favorable timing of collections, which were much greater in December than we anticipated.
Our DSOs for FY 2016 were 105 days, an increase of 12 days from the same period a year ago. Our DPOs for FY 2016, however, increased by 22 days to 84 days from the prior year, thereby narrowing the gap between DSOs and DPOs. While cash flow from operations can fluctuate meaningfully from quarter to quarter due to seasonality and timing, our net cash position of $108 million and our revolver are more than sufficient to manage the ups and downs. Looking ahead to 2017, we expect full-year revenue to be $270 million on total gross spend of at least $1.45 billion and adjusted EBITDA to be $72 million, or a little bit above 26% of revenue.
For Q1 of 2017, which, of course, is seasonally the slowest in our industry, we are expecting revenue of $43 million and adjusted EBITDA of breakeven ±$2 million, depending on the timing of investments in our personnel, technology, and new offices, such as in Shanghai, Paris, and Madrid. With that, I will hand it back over to Jeff for any final comments and, of course, Q&A.
Thanks, Paul. Before we take your questions, I want to provide a few highlights on the news we issued concurrent with our earnings release today. Today, we filed a registration statement for a follow-on offering. I want to stress that this offering is 100% secondary, with zero primary shares being offered and therefore zero dilution to shareholders. The offering allows us to provide more liquidity for our early investors and employees an opportunity to diversify a small amount of their investment in the company. We also expected to increase our float and therefore reduce the volatility in the stock. In 2016, connected TV ads grew by 40x for us last year. Mobile and video are now the majority of our revenue and growing faster than the rest of our business. Our international growth outpaced the U.S. by more than 2x in 2016.
Finally, one of our strongest strategic assets is the independence and objectivity that comes with not owning media, and it significantly differentiates us from the competition. This is opening doors for us around the world and notably helped us partner with Baidu, which is being implemented and rolled out shortly. We enter 2017 with the wind at our backs and with fewer competitors. We continue to surpass our own expectations and are extremely pleased with our results during the quarter and for all of 2016. We have executed well, and we're hitting on all cylinders as we enter 2017. With that, we look forward to your questions. Operator, shall we begin?
Thank you. The floor is now open for questions. If you do have a question, please press star one on your telephone keypad at this time. Questions will be taken in the order they are received, and if at any time your question has been answered, you can remove yourself from the queue by pressing one. If you're using a speakerphone, we ask that while posing your question, you pick up your handset to provide favorable sound quality. Again, ladies and gentlemen, if you do have a question, please press star one on your telephone keypad at this time, and please hold while we poll for questions.
Let's take the first question, Mike.
Okay. Our first question comes from Shyam Patil. Please state your question.
Hello. Good afternoon, guys. Jeff, could you talk a little bit about how you arrived at the growth spend and revenue guidance for the year? It looks like you're implying growth around or even less than 2x the market after growing much faster the past two years. There's certainly the law of large numbers, but aside from that, can you just talk about the thought process you used to get to the guide? I have one more follow-up.
You bet. First, I'll say that we've guided a little bit conservatively historically simply because so many things have gone right for us. 2016 was an amazing year, in part because we had the Olympics in 2016, we had an election in 2016, then the election sort of pushed some of the retail dollars into the later part of the year, which is what made our Q4 really strong. We also were able to make a bunch of investments for the long term. We know that right now in 2017, we're making investments, some of which won't pay until 2018 and beyond. With all of those things together and trying to win trust with Wall Street, we know that we need to guide conservatively.
I know we threw out a lot of numbers throughout the report that we just gave, but a couple of that I'll just highlight. We grew connected TV by 40x. Of all the numbers we talked about in the last 30 minutes, that's the one that I'm most excited about. Between January 2016 and December 2016, 40x growth in what I think is the most promising channel in media today. We also are able to make investments in Q1 that we previously guided that we would open offices in the first half of the year. The fact that we were able to get Jakarta in at the end of last year, meaning open the office, doing significant revenue there, as well as Madrid and Spain so far this year, we're super excited about the impact that that'll have in the long term.
You put all of that together, and we're extremely bullish on, and frankly bold in predicting that we'll continue to grow at least double the industry.
Great. Just to follow up on the spending. When you look at the implied spending or OpEx for this year, how much of that do you consider to be investment? Can you just talk a little bit more about just what specific areas you consider investment this year?
You bet. First, it's kind of hard to distinguish because growth is business as usual for our business. If you're not capturing share, and particularly in the new channels, you're going to be shrinking, not growing. It's hard to actually think of it as like, what is your base and where is the growth? It's all the same thing to us. Because we've never known a time where we didn't have 95% plus client retention, it's always investment in retaining and growing the clients that we already have. That's sort of first and foremost. By hiring more engineers in 2017 than we ever have before, those are certainly investing for the future. The international, I think it's fair to say that we invest about double what they contribute in all of our international markets, if you're using round numbers.
All of that is investment for the future. Inside of television, we're certainly putting more engineering muscle behind creating a foundation that if you were to measure it on its own, is probably not profitable in 2017. The land grab that that represents is obviously massive. I'm not sure if Rob or Paul, you'd add anything on operating costs. Okay.
Thanks, Shyam. Next question, operator.
Our next question comes from Brian Fitzgerald. Please state your question.
Hi there. Thanks for taking my question. This is John on for Brian. First off, congrats on the quarter. Just a follow-up to the previous question. Once you've made these investments in international and headcounts and on the R&D side also, can you talk to the kind of leverage in the model of how that kind of flows through maybe 2018 and beyond? I have a quick follow-up. Thanks.
You bet. Every market is different. Some markets like Singapore have been pretty linear, where from employee number 6 on, we're profitable. There are other markets like Germany and Japan where we invest for two, three years before we get that return. Part of it depends on how aggressively we're willing to spend in any individual market and how much confidence we have that it will ultimately pay off. In markets like China, which we expect to be a big investment area for us in 2017, we'll go all in, and I expect the shape of that curve to look more like Germany and Japan. Those will take beyond 2017 in order to pay for themselves. The way that we think about it is how certain are we that it's going to return?
How much opportunity is there if we invest now? Because, for instance, in China, we think it's a land grab time where people are hungry for a global, omni-channel, non-conflicted solution. It actually is something of a no-brainer for us to just make a big investment.
Great. Thanks, guys. Just quickly on take rate. It looks like in your guidance that it is kind of stepping down a little bit. I know in the past you have talked about as you expand into digital video and audio, that's kind of expected to happen. Just curious your thoughts on, is that trend kind of playing along the lines with your initial thoughts and any updates there? Thank you.
You bet. I was hoping somebody would ask this question for a couple reasons. Number 1, I just want to reiterate that we internally don't use take rate as a core metric of our business. My hope is that over time we'll develop the trust with Wall Street that we move to what I think are 4 metrics of our business, particularly the growth of EBITDA and revenue, which in our case is net revenue. The fact that we operated at a 32% EBITDA margin for 2016, I think should give people the confidence that I think take rate has been a proxy for in other cases. Meaning for other businesses that have never been profitable, take rate has become the really core metric to their business because it's a proxy for profitability.
Now that said, as it relates to take rate, there's a number of things that can create slight adjustments for take rate. One is volume discounts. Another is there's differences in channel mix and customer mix. I expect in 2017 to be slightly more volume discounts, and that's just a byproduct of us having 95%-plus client retention and cohorts that are super strong. We're going to give discounts to those who spend more, and we've always done that to incentivize them to put more spend on the system. I do think that will have very slight impact on take rate. However, I just want to point to what, again, are the core metrics, which are during that time, EBITDA and revenue will be growing at a very healthy pace. That's the thing that we're optimizing for, not that proxy percentage.
Great. Thanks, guys.
Thanks, John. Operator, next question.
Our next question comes from Mark Mahaney. Please state your question.
Okay. I think that's me. It's Mark Mahaney.
Sorry, Mark.
That's all right. Let me just ask two factual questions, please. I think you said it, but the percentage of your revenue or gross and gross spend in 2016 that was international, then commentary over what's implied in your guidance for this next year, the percentage that comes from international, both with spend and with revenue. Thank you very much.
The first one, I'll just give rough numbers. The spend for 2017 in terms of a percentage coming from international markets is roughly 10% comes from international markets, with again, us meaningfully over-investing. For instance, roughly 20% of our employees are in international markets. Just a commentary on how we're over-investing. Because it's growing at roughly double the pace or more than double the pace of the U.S. market, we expect that to be something more like 15% in 2017 instead of 10%.
Thank you, Mark.
Thank you, Mark. Operator, next question.
Our next question comes from Youssef Squali. Please state your question.
Yes, thank you very much. Two questions. On the margin, longer term, how should we think about your ability to hit that 40%-plus that you guys talked about at the IPO? Has the increased level of investments that you're talking about for 2017 either making that potentially harder to get to, or does it just push it out, but in the meantime, we're getting hopefully higher top-line growth? Can you maybe just parse out the growth in the fourth quarter, that 70% is still very impressive, between maybe pricing and volume? Thanks.
Yeah. I'll ask the people in the room to just make sure I answer all those questions we're just given. Sorry, which was it for 40%?
The EBITDA margin.
Oh, yeah, heading for 40%. No change in terms of us heading towards the 40% EBITDA margin. Just again, to sort of check where we are in our journey. We say all the time, we're 2% done. At the stage we're at, and particularly because of the competition that header bidding is creating, now is definitely the period of time where we want to be land grabbing. Growth in 2017, in some ways, is more of a priority than it's been in recent years for us. Now is the time for us to invest and go get further ahead. That's why we're focusing on engineering. That's why we're opening so many international offices. I mean, last year we opened one international office. This year, I think we'll open five. With that sort of investment, we're going to grab land.
At sort of steady state, we still think we're approaching that 40% EBITDA.
There's just so much to go, the 2%. We're still in heavy growth mode. Over the next several years, we'll be continuing to earn towards 40. By the way, we view the investments we're making now as somewhat of an insurance policy to make us more likely to get to that 40% in the out years.
What was the other component to the question? Sorry.
Q4 growth, pricing versus volume.
Oh, yeah. Was the Q4 growth more the result of us chasing volume?
Right
versus giving price discounts? Am I getting your Help me understand your question again.
No, that's it. That's exactly it. Just trying to understand the dynamics going on in the marketplace competitively, how are volume and pricing driving the top line relative to prior quarters, if there has been any change, maybe there hasn't been.
You bet. I would say in 2016, we started using a mantra, frankly, with our clients and with our own team that we hadn't in years past, and that is: It is not our goal to be the cheapest platform in the world. It's our goal to be the best. While many years ago, we might have tried to underprice others to win business As we try to defend the business that we've already won, the way that we do that is in efficacy and in the value that we add, not trying to be the cheapest price. We spend a lot of time with our customers pointing to the value that we add and not the price that we offer them. Because we think we add so much more value, it's really all a byproduct of volume and growth.
It wasn't us lowering the price so that we could grow faster than everybody else. There's no way we could have put up the profitability numbers that we did last year if we'd taken that approach. I think we were able to thread the needle, which is create optimal growth, which is our first priority, and while remaining as profitable as I think any independent ad tech company ever.
Okay, helpful. Thanks, Jeff.
Thank you. Our next question comes from Kerry Rice. Please state your question.
Thanks a lot. Great quarter, guys. I noticed in the K that you did provide some maybe qualitative data on cohorts. Can you talk a little bit maybe more about that, if you can give us any growth rates of any of your key cohorts? The second question maybe relates to the previous question about volume versus pricing. With the header bidding being a key trend in the marketplace, did header bidding raise CPM prices? Was that beneficial to you? Can you talk a little bit about the impact of header bidding maybe in Q4 and think about it in 2017? Thanks.
Great. Let me start by talking about the cohort, and then if I leave anything out for Rob, if he'll just add to it. I'm really glad that somebody asked about the cohort, in part because I think it's the way a lot of people are thinking about the modeling exercise that our company is. Getting the color around it, I think, helps sort of predict what's going to happen. Just a reminder, the 2012 and 2013 cohort represent sort of smaller customers. When we first got started, we went to small agencies and small shops that we could sort of test with and not make big mistakes and prove ourselves and create case studies so that we could then go to the biggest buyers in the world. Where we really started to see success on that was in 2014.
In 2015, it became, I think, much easier for us to win the biggest buyers and particularly the biggest agencies in the world. We think that those businesses are going to continue to do well, I think there's a case to be made that our 2014 and 2015 cohort will be some of the strongest cohorts for our future going forward. I think it's fair to assume that and model it that way. In 2016, this was helped by the IPO, that's the first time that we started winning on stages and in the spotlight instead of just in conference rooms. I think it is fair to say that all of our wins up until the IPO came in conference rooms.
We have a little bit of wind at our back since then in terms of creating more awareness and winning more customers in sort of the middle size. As a result, those are growing just a little bit faster than our biggest. Still, I don't know that we'll ever have a cohort as competitive as our 2014 and 2015 cohorts. Did I leave anything else out before I talk about-
No. The only other perspective I'd call out is what we talked about during the roadshow is, hey, remember it's still early, right? Still the majority of brands are not spending programmatic in any serious way. We see a lot of growth from brands shifting more dollars from digital into programmatic and frankly from non-digital directly into programmatic. It's still early days, both in terms of the number of clients, if you were to look at all the brands out there, who is spending programmatically, and then secondly, the share of their total advertising spend that's in programmatic. We think the cohorts have a long way to go, the ones we have today. Frankly, the cohort that we signed on in 2016 is the strongest one we've ever signed, and we expect a lot of growth to come from them in the future, too.
Awesome. As it relates to the second half of your question, which is a really complex mathematical question to answer, which is the effect of header bidding because it creates more competition as well as because we are able to see more of the premium inventory as a result of header bidding, it naturally does move prices upward. Do we benefit from that versus any of the cost that the additional volume brings? In other words, because we have to look at more ad opportunities because in some cases the same impression is represented in multiple auctions, in fact that's often the case now, that does increase our cost to examine each individual impression and then figure out the right answer.
Because the price has also gone up, that one because we typically operate on a percentage of spend, we benefit from raising prices, but it hurts us that the volume has gone up. When you put those two things together, they roughly offset each other. It's hard to say what the net effect was because it's almost zero. The most important piece is that the value add that we provide to our customers has gone up meaningfully because the waters have become a little bit more choppy, if you will. The need for you to examine everything and use data to make sure that you're making the best choices has gone up. The odds of being successful without being data-driven have gone down. All of that looks like good news for us.
If you also think that puts pressure on our competitors, most of which are not profitable because of that dynamic, and particularly if you're not growing as fast as we are, which I think there's a strong case to be made that none of them are, then we've separated from the pack while having very little impact upon our profitability and expenses. Overall, especially on a competitive basis, that header bidding dynamic benefits us probably as much or more than every other company in advertising in the world.
Thank you.
Again.
Thanks, Kerry.
Again, ladies and gentlemen, if you do have a question, please press star one on your telephone keypad at this time. Our next question comes from Aaron Kessler. Please state your question.
Great. Thanks, guys, and congrats on the quarter. A couple questions. First, just on the Q1 guide. I believe that implies roughly 40% sequential decline. Should we just think of that, I realize there's obviously seasonality in the business, but is there a little more kind of conservatism built in there? Just maybe an update on thus far kind of what you're seeing through mid-February? I think previously you've talked about, you've touched about 10% of the brands. Just kind of where does that statistic stand now? Finally, I think you mentioned international is about 10% today. What's your thoughts on kind of what that looks like in maybe two to three years? Thank you.
Right. I'll take the first part of the question, then I'll ask Rob to speak on the 10% of brands question. As it relates to the guidance in Q1 or for Q1, it is definitely fair to say that Q1 is the hardest quarter for us to forecast and guide for. It's because there's a seasonal nature to advertising. People tend to start the year reconsidering how they allocate budgets. While we're sort of an install inside of the agencies and technology providers that we power, that doesn't mean that they're not susceptible to that seasonality. Then when you're coming off of the election year and the Olympics and all those things that made 2016 unique, forecasting for the year is hard.
Particularly when you had such an amazing year like we did, we want to make certain that we provide guidance that we know that we can hit. While we're super excited to report that we are growing at at least double the industry, we also recognize that we have a lot of trust that we have to gain on behalf of sort of the ad tech community, we know that has to be reflected in the guidance that we provide.
Great. I'll jump in. In terms of brand adoption of programmatic, as touched on briefly earlier, just to reiterate and be more clear on that one. Yeah, most of the big agencies that have rosters of advertisers in the hundreds, there's still sort of double digit, 10%, 20%, 30% of the brands inside many of the agencies that have adopted programmatic in a serious way. There's another incremental percentage that's experimenting with very small parts of their advertising budget. When we look at the penetration within the agencies and then their brands that they aggregate and service, we just see a long runway to go. It's way less than the majority of brands that are spending in programmatic in any serious way.
We think in our cohorts that Jeff talked about in 2014, 2015 and 2016, as we signed most of the large global agencies, there's just a ton of runway to go in terms of brands adopting programmatic and them working on The Trade Desk platform. I think the last question you asked is around international growth, Jeff mentioned it perhaps early in the call, roughly about 10% by the end of last year. We're over-investing, particularly in the key international markets like China and Indonesia. We've got really strong positions throughout Southeast Asia with our presence in Singapore. We've been in Korea and Japan for three years. We've been in Europe for four years. We expect all of those to continue to grow much faster than the U.S., and therefore, take relative share as a % of our total business.
I think we talked about heading towards a 15-ish % by the end of this year.
Got it. Great. Just one quick follow-up. Election sounded like it was a decent traction with the election in Q4. Can you give us a rough sense maybe how much gross spend was in the election that you benefited from?
Yeah. Directly, it was in the low single digits. It wasn't like some massive amount of our spend. As we think about the future, it's not like you should massively downgrade what would happen in mid-2017 because there aren't any U.S. elections. There's nothing like that at play. That said, when you combine the advent of header bidding with the increased competition that comes from having the election, and most notably, as a consumer, I'm sure everybody on this call, if you were looking closely, noticed that there was way less retail advertising prior to the election in November. There was a little bit of a sense of making up for lost time, particularly inside of TV. That had an interesting dynamic on prices that had a bigger impact on our business than just that low single-digit percentage that was added through elections.
Got it. Great. Thank you.
Okay.
Thanks, Aaron. Operator?
Thank you. There appear to be no further questions at this time. This will conclude today's conference. We thank you for your participation. You may disconnect your lines at this time, and have a great day.