The Trade Desk S-1

As many of you know, The Trade Desk (TTD) filed its S1 with the SEC this week. This was alluded to in an earlier Lift Letter (AppNexus and The Trade Desk S-1s) that delves a little deeper into what an S1 is, going public, and the like. You can read the TTD S1 here: When you read the S1, you'll note that there are a number of blanks around the price per share and amount raised. This is because the offering is not yet priced - the banks are going around drumming up interest. The round will eventually be price when it goes public and the number of shares * the offering price is the amount they raise. Below we discuss some note worthy points from the S1. Another interesting note is that they're offering 2 classes of shares - class A is the issuance, and class B is held by employees, current investors and executives (until they're sold, then they become class A). Class B shares get 10 votes per share while class A has 1, meaning the class A shares - because of the current makeup in ownership - cannot make any decisions.


TTD only makes its money as a percent of media spend through its platform - there is no media arbitrage, only long term master service agreements with fixed media fees. In 2015, there was $550M spent on their platform, from which its revenue was $113.8 - meaning their fee is about 20% (higher than I thought it was). In 2014, its revenue was $44.5M. The company was profitable by $0.005M in 2014, and by $15.9M in 2015. On June 30, 2016 they had 387 employees, which is half the size of MediaMath, and a quarter (or less) the size of AppNexus. But TTD is likely worth a similar amount, if not more, meaning it is very effective at generating revenue on a per-employee basis (something they make quite clear in the S1), and that it is easier for them to be profitable. Also, both TripleLift and TTD are quickly growing companies whose revenue is impacted by media spend. We projected our Q4 to be 40% of annual media spend, and they saw 42% in 2014 and 37% in 2015.

The key trends they highlighted as the basis for their business are the shift to digital, consumption fragmentation across many websites / apps, shift to programmatic, automated ad buying, and more data in ad buying. This largely agrees with our views, though it is - in my view - is more retrospective than prospective. Perhaps more prospective on their part, native advertising was mentioned 6 times in the S1. They also view programmatic TV as key to their future. TTD expects their margin to decrease as they move into TV.

TTD highlights a few key risks. They don't have exclusive relationships with agencies and it is very easy for agencies to move budgets from one DSP to another. Omnicom and WPP are each more than 10% of their revenue. They are concerned that SSPs may cut them off from high quality inventory at any time. And - a specially called out item for them - training for their platform is so important that they mentioned that if they fail to do it at a high level it could be an existential risk for their platform.

The ad tech industry has not done terribly well in the public markets over the past couple years, so TTD does go to some lengths to differentiate themselves. The discuss their view of the industry, and their position as a non-arbitrage, pure-technology, pure buy-side business with transparent optimization - implicitly contrasting against other public companies. They speak frequently about their customer retention (95%) and their rate of growth (>100% annually).

AppNexus and The Trade Desk S-1s

Recently, there were reports that AppNexus and The Trade Desk were on the verge of filing their S1s ( What is the significance of this news? What does it mean for us and the industry?

We've previously discussed what it means to IPO (What Does It Mean to IPO). This included conversations around public v private company regulations, as well as the process. That discussion was largely phrased from TripleLift's perspective. Filing an S1 is the initial securities registration with the SEC (the securities regulation agency in the US). When you file your S1, you submit for comment from the SEC information about your business, disclosures, risk factors - all the things a reasonable public investor might want to know, all of which is then publicly available. Here's a copy of Google's:, Rocketfuel:, Tremor Video: They're pretty interesting and revealing.
download (6).png


The efficient market hypothesis states that public securities are priced to perfectly incorporate all publicly available information, and effectively that pricing is devoid of emotion. This is, entertainingly, both a presumption in the federal courts ( and largely incorrect - even the weakest form ( This derives from two different perspectives: 1) the US courts want to ensure that all material information can reach the public markets so that investors could use that information to make a decision, which is a noble goal, and 2) the markets are gonna do what the markets are gonna do, which often means not being "efficient," but instead of having moods.

In the current environment, IPOs for tech companies have not been received well. There was no "efficient" answer for this, it was just a cyclical thing that the investors were spooked by a few ideas, including a lot of unicorn bubble talk, fears about China's declining economic growth, oil prices, etc. Because every company IPOs precisely once (unless blah blah blah), no one would want to do it in a market that wasn't receptive. Recently, however, a brave soul (Twilio Inc.) decided to go public. Twilio is a telephony-based tech company, and its IPO was a smash success. People look at the first day's results, its first several months, and how it performs relative to projected earnings for the first few quarters. For Twilio - so far so good (initially priced at $15, currently at $34).

Ad tech is also in a bit of a slump. A lot of the companies that have IPO'd did not do well in the public markets, including RocketFuel, Tremor, Yume, Millenal (now AOL/Verizon) and several more. A couple others did well enough, including Criteo and Rubicon (to a much lesser degree) - but there was more bad than good for ad tech. These mediocre IPOs were often after raising a ton of VC money, in many cases less than the market value. This trickled down into the VC world, meaning VCs thought the IPO market wouldn't necessarily welcome ad tech, which meant it would be hard for them to turn a profit, which meant they would be less likely to invest.

All this gets to the point that AppNexus and The Trade Desk just filed their S1s. These are two good companies. The Trade Desk is probably the largest independent (meaning not Google) buy-side platform, and AppNexus may have the most money flowing through its platform of any independent (also meaning not Google) ad tech company. The fact that both of these companies are on the verge of filing their S1s means that they are seriously considering going public. It is quite expensive to prepare, revise, and repeat to the level that the SEC needs, and almost always means the company will - when they're ready and have the prospectus and buyers lined up - actually go public. So one can imagine that two strong ad tech companies will IPO at some point in the next several months (timeframe TBD). 

This will have a positive cascading effect through the industry. It will show that the public markets are ready for tech IPOs again. It will show (hopefully) that ad tech companies can be legitimate, successful, public companies. It will show to other ad tech companies that the market is increasingly receptive, which may lead to more solid ad tech companies filing IPOs. And it will show to VCs that ad tech should be considered ripe for investment.

What Does It Mean to IPO

Many startups aspire to IPO. This means they want to hold an initial public offering. If TripleLift is particularly successful, we may do the same. What does this mean? What's the point? What are the pros and cons? 

When a company IPOs, the process is similar - in certain ways - to any round of fundraising for venture-backed companies. The company will issue new shares of stock and sell them to buyers (investors) at a specified price. Like the other rounds (and as previously discussed in the Preferred Shares Lift Letter), this is dilutive to existing shareholders - meaning there are now more shares outstanding than there were previously - and there are certain rights associated with the buyers of the shares.



If TripleLift were to IPO, all of our preferred shares would convert to common. In our case, assuming the IPO is above a rather low threshold, preferred converts 1-to-1, meaning each preferred share gets one common share. Companies may have very complicated conversion mechanics, but generally the take home is that preferred shares turn into some number of common shares. These common shares are the same as those that are traded on the exchanges - so they too can now be sold. 

There are a number of reasons to IPO, which include: 1) the ability for owners of company stock to freely sell shares, 2) the ability to access the public markets as a whole, which includes selling more equity at a later date, and to obtain debt financing at a scale often not accessible to private companies through debt offerings and the like, 3) vanity, and 4) the elevated reporting and governance standards of public companies (debatably not a "pro").

To be publicly traded means that, literally, the public can trade the stock. It "goes public" on generally the NASDAQ or the NYSE after a whole song and dance where the company woos buyers for the initial offering through a "roadshow" with investment bankers. The company must file a ton of documentation with the SEC about its records, history, risks, prospects, etc - all in the name of public investor protection. After the SEC gives its signoff, the company has its buyers, and the exchange is all set - the stock goes public on a particular day, meaning the ticker corresponds to a certain type of share in a certain company, and buy and sell orders can now run through the exchange's order management system between any parties, at any time that the exchange is open (except right after the IPO, when employees generally have contractual lock-up provisions for between 90-180 days). 

This is markedly different from today's state of affairs for TripleLift, where there's actually no price or value to TripleLift - whereas in a public exchange, it would be traded every second so there would be a regular "market cap" to the company. In order to be publicly trade-able, you have to satisfy all the rules, requirements, and processes of IPO-ing, which we have not done. This also means shares in TripleLift, as a result, are not publicly trade-able. The SEC has requirements on how many shareholders a company can have before it must satisfy the requirements of being publicly traded, as well as the nature of transactions that are allowed - which often means that a buyer of a non-publicly-traded company has holding requirements of several years, if the transaction is permitted at all. All this means that there isn't a ton of liquidity available for private companies nor is there regular pricing information (there are certain, limited exceptions for companies that expect to go public in the not-too-distant future - Second Market being a good example).

That said, it's not all gravy to go public. The SEC has a huge number of intensive, ongoing requirements for any public company, including regular disclosure, quick public reporting requirements, Sarbanes-Oxley, etc. It also means you're now subject to the whims of public market investors who act very differently from private market investors. Private market investors have the restrictions of generally needing to stay in the company for a while, so they have to care about it, to at least some degree. Public market investors have a different set of incentives. Many, like retirement and mutual funds hold shares for a long time. But high-frequency trading and activist investors can change the strategy, board of directors, and senior management of a company with the intent of making a quick buck - perhaps at the expense of long-term strategy. 
But at some level, large successful companies where the shareholders want liquidity for the equity owners will generally need to go public at some point in order to handle the number of likely buyers. There are noteworthy counter-examples, like Koch Industries, Bloomberg, Advance Publications, Cargill, Saudi Aramco and others. These companies are generally family- or state-owned, and produce enough capital for the owners on an annual business that they have no interest in selling shares in the company. 

Rocketfuel IPO


On September 20, 2013, Rocket Fuel IPO'd. It had a volatile couple months, and then cratered precipitously. The stock has yet to recover (see below).

Back in the halcyon early days of RTB, Rocket Fuel was indisputably a leader. It only raised around $75M before going public, and had demonstrated a consistent and significant top line revenue growth. In fact, Rocket Fuel may have had significantly higher revenue than its demand-side competitors, including Turn, MediaMath, etc.

download (3).jpeg

What happened?

No one knows for sure, but I believe it's a very unfortunate combination of situations. First, when a company goes public, there a few very important numbers and events. The ability of the company to successfully IPO at or above the range it had announced, the closing price of it's first day of trading - and each day for a few weeks (also the market), and the company's first earnings report. Companies that miss their first earnings report may be punished by reduced analyst coverage, which translates to reduced aggregate demand. On Jan 22 2014, their first earning's report came out. Gross revenue showed significant YoY growth, but the company sustained large and increases losses for the year. In May 2014, their second earnings report as a public company, they continued to showing a large - growing - top line, and ever-increasing overall losses.

Rocket Fuel developed a model that - while the media was transacted programmatically - they never developed programmatic relationships with their customers. This meant - by and large - campaigns were won on an IO-by-IO basis. This isn't a problem itself - BuzzFeed is in the same boat and has a solid valuation. But while Rocket Fuel had very high margins on their IOs, in part because of their technology, their business required very substantial marketing, sales, and R&D costs that made the company unable to generate a profit.

As mentioned, their first two earnings were disappointments, so analysts dropped coverage and interest as a whole waned. The complexity of ad tech only added to things, as did certain possibly-questionable practices around bots and fraud (, and their being shut off from Facebook's exchange. Most public market participants never really understood their business. Instead, they began to be see the company as a media company - because they had to rely only on IOs. As a result, the multiple of the business changed from being a technology company to that of a media company (which is significantly lower).

Rocket Fuel is still a large and important company in the space, doing over $111M in Q3 of 2015 in top-line revenue. This may actually represent a situation where the markets are incorrectly valuing a company (this isn't investment advice!!!). But as Keynes said "the market can stay irrational longer than you can stay solvent" - and people would actually have to come around and want to buy the company. We'll see what happens. But the first step was their replacing their founder CEO (George John) with someone that may be more focused on controlling costs and driving the company to an actual profit (Randy Wootton).

Here's their most recent quarterly report - it's interesting: