A Call to Israeli Engineers! Adtech Is Not For You.
Years ago, in Part 1 of the Humus Manifesto, I wrote about the talent shortage in Israel which partially derived from talent trapped at a lot of small companies that would not scale. Today, I think that Israel has a much, much larger talent pool to tackle some really big problems, by leveraging big data, building marketplaces and creating serious technological innovation. However, much of this talent is now trapped in a lot of adtech companies that will neither change the world nor deliver much of a return to this engineering talent. I think these engineers, the country and the innovation economy would be better served if we recycled these great engineers into other endeavours.
In the following paragraphs, I plan to argue that if you are a talented engineer, it is a bad financial and career decision for you to go an adtech company or stay at one (the Yahoo acquisition Brightroll notwithstanding). Additionally, I would argue that the current trajectories of VC-backed adtech companies are unsustainable and will not reward you. Lastly, if you are a great engineer, you should be targeting more ambitious missions than extracting better advertising margins.
The salient points of the argument are:
- Adtech is on the back-end of the innovation curve. Its technology breakthrough glory days are are in the past.
- As online advertising in various formats becomes a larger part of marketing budgets, advertisers (and publishers) are attacking pricing and cost more aggressively.
- The industry has structurally poor margins over time.
- You need to raise a lot of money to build an adtech company, so engineers are buried under a lot of liquidation preference. However…
- You are not rewarded with premium public market multiples due to the structure of the industry, its margins, its dependencies and competitive dynamics.
- You should really be spending your time and talents solving bigger problems and challenges.
Israel has an abundance of adtech companies. I think a lot of adtech founders and engineers see easy money so they deploy their awesome engineering talents there. The reason for this perception is that there is a ton of money sloshing around in advertising and a quick path to revenues and even minimal profits. Some of these adtech companies are venture backed and others are bootstrapped. In my opinion, the VC-backed ones will struggle to deliver their engineers much of a return. In fact, adtech is a value trap and is the farthest thing from easy money at scale. Let me try to explain why:
Is Adtech Still Truly Disruptive Innovation?
Today, adtech is on the back end of the innovation curve. I think it would be correct to say that it is on the downward side of the slope (see figure below). It is not surprising anyone but rather optimising existing models and inventory.
That does not mean you cannot make money as a founder of an adtech company. What it does mean is that adtech and ad networks are no longer disruptive and, therefore, will have difficulty attaining sustainable hyper growth, defensible moats and outsized returns needed to return a significant multiple on capital.
Today, most adtech companies are exploiting features that are missing on the core platforms of Google, Facebook, and many of the already public companies. They are optimising and brokering between technology platforms (mobile and web), exchanges and advertisers. However, information is nearing perfection in this market, making it difficult to build a moat around businesses and maintain margins. This flies in the face of Peter Thiel’s oft-stated mantra that the best businesses and startups you can build are ones that can become natural monopolies (at least in a niche).
Adtech is heavily dependent on the vagaries of distributors and is “sandwiched”
Let me simply describe the structure of the adtech/advertising industry to illustrate this point more effectively. Essentially, there are three players in the advertising industry — the advertiser, the publisher and the broker or middlemen — who decide, how, when and where to put ads. The advertiser — let’s call them Coca-Cola — has a budget for ads and their task is to find the most effective way to place those ads to bring as many customers as possible for as little money as possible. On the other side of this equation are the publishers — let’s call them the New York Times or Facebook’s feed — and they have the eyeballs or customers that Coca-Cola wants to access. They want to capture as much rent as possible from the advertisers, so they can optimise their return on “eyeball inventory”. Essentially, since the publisher controls the customer, they are the power broker and determine the price to access the customer. That price could be discovered through a rate card or some other discovery mechanism, like bidding.
If you are an adtech provider, network of advertisers or a data provider, you are playing to earn the spread between what the publisher can capture and what the advertiser is willing to pay for optimization. For example, if an advertiser, using his own tools, believes he must pay $1 per customer and he uses a network or adtech provider to optimise his delivery so it only costs him $0.80 to reach that customer, the adtech provider will share in the $0.20 delta that he helped the advertiser save through whatever method he was able to access the publishers’ customer for less money. On the flip side, publishers have limited sales forces and cannot reach all of the advertisers who may want to reach their customers so they often turn to the brokers or adtech companies and ask them to use their networks to fill the page inventory of the publisher.
A credible publisher is essentially in the content and communication business and hence the real estate or inventory he has for advertisements is necessarily limited. Within reason, the publisher is trying to optimise his earnings and will balance the types of ads between banners, search, content links etc. across his site. The adtech providers fill this void by providing the various networks (content, banner, search) that enable the publisher to fill his site or mobile app. Here too, the adtech provider is playing for the brokerage spread between the publishers (he controls the customer) need for earnings and the advertisers need for optimized customer acquisition cost (he controls the budget).
The adtech provider or broker actually has no control in these cases as the publisher controls the customer and the advertiser controls the budget. What this means is that an advertiser will always look to trade away for a lower cost of customer acquisition and the publisher will look to optimise his rent and holds no loyalty to the broker network. This means that from an ad and content quality perspective the higher paying ads win which leads to a race to the bottom (bikini content links and viagra ads still pay better than scholarly articles or considered purchases). Further as opacity in the ad industry fades and information becomes more perfect, the publisher will, inevitably, demand more of the margin than the adtech provider is collecting.
This is why gross margins are flat or declining for adtech companies and why public market performance of adtech companies is so poor (more on that below). What makes Google so powerful is that they both control the customer (across mobile and web) and the network. So Google.com gives them customer control, price discovery and liquidity in their ad network. They also then directly market to the advertiser with a bidding model which gives them an understanding of the full loop of price elasticity. They then deploy this network to other sites but you cannot disintermediate them because they are the market and a publisher. The same goes for Facebook today and their nascent Atlas network ambitions that they will likely succeed with, leaving a wake of adtech value destruction in their wake.
Today, this same lack of control is also vexing content publishers who rely on search engines or social networks for bulk of their traffic. This excellent piece, on Monday Note entitled: “Facebook and Google Now Control Media Distribution” lays the issue out beautifully.
“The first concerns the intrinsic valuation of a media so dependent on a single distribution provider. After all, Google has a proven record of altering its search algorithm without warning. (In due fairness, most modifications are aimed at content farms and others who try to game Google’s search mechanism.) As for Facebook, Mark Zuckerberg is unpredictable, he’s also known to do what he wants with his company, thanks to an absolute control on its Board of Directors (read this Quartz story)…None of the above is especially encouraging. Which company in the world wouldn’t be seen as fragile when depending so much on a small set of uncontrollable distributors?”
Adam Singolda, CEO of Taboola, who is trying to build a full solution to get around some of these challenges, understands and also makes this point extraordinarily well:
“…the underlying risk is not that Facebook may send publishers less traffic than what they send today due to algorithmic changes. It’s that one-day Facebook will take these publishers out with the flip of a switch…
For those in the industry, Facebook is a source both of great abundance and great fear.”
Well, adtech and ad networks are equally fragile as they are completely dependent on publishers (many of whom themselves, as Adam points out, are dependent on Google and Facebook.)
Are Up Front Payments the Answer?
The way that adtech companies deal with the lack of control over the end customer and the possibility that the publisher will swap them out is that they cut long term deals with publishers. Essentially, they pay up front or commit to paying up front for inventory that the publisher may have. Publishers have gotten very good at this game and they bid the adtech guys against each other. To use an example from Israel, AOL effectively bid Taboola against Outbrain for the right to provide content links to AOL’s properties. For years, publishers would bid Conduit* against IAC for search distribution. This is true across the adtech industry, which means that, in order to grow your business, you must lock down these long term contracts and you need the cash to pay for them. Therefore, adetch companies must raise ever larger amounts of capital to grow. Those capital raises are the rub if you are an engineer at an adtech company because it piles lots of preference on the cap table. If your adtech company has raised $100MM that means that investors will need to get back their $100MM + 8% per year before you, the brilliant engineer, see a nickel.
An adtech founder, who sold his company for tens of millions of dollars, shared the following insight with me,
“The challenge of getting to cash flow positive is not only about squeezed margins. It’s about the fact that ad tech companies are backend collection businesses, which have to pay for the inventory they use on time or risk being kicked out of the big platforms, but then have to go collect from their customers, advertisers and agencies which are notorious in their payments delays. That causes the effect that as you grow, especially if you grow fast, there’s a growing working capital gap you have to deal with. This is the total opposite of a business like groupon, where you get paid sometime months before you have to pay. Small adtech startups have zero leverage against the publishers and against the agencies/advertisers.”
At this point, you might be saying to yourself, what is the big deal? Uber raised over $300MM, Wix** raised almost $100MM. Why is my adtech company different? Here is why: Uber** owns its customer and has his credit card on file. They have brand loyalty and a deep customer connections. They distribute directly to the customer and are not dependent on anyone else. Wix owns its customer and has a direct subscription relationship with it. They market directly to the customer and control the cost of goods!
They have customer loyalty as well. People love Uber. They Love Wix. They Love AirbNb. They love Salesforce.com and Tableau. Nobody knows who the heck the adtech guys or the Ad Networks are with the exception of Google and Facebook.
Adtech companies are sandwiched between the publisher that owns the customer and the advertiser that controls the budget and both of them are always optimising at the expense of the broker. Therefore, as adtech businesses raise capital*, it is actually not that complicated to grow revenue since you lock in publisher contracts to provide ads from your network.
However, then you need to transfer that inventory to advertisers for a slice of the spread. When you then sell that inventory via exchanges or directly, you will actually sell it at ever declining prices. This is why average CPM is way down as inventory on publisher sites increases and the networks become every more efficient. Therefore, it becomes very hard to grow gross margins sustainably.
Said differently by the adtech founder I quoted above,
“when you buy up front but have to sell in a volatile marketplace, you neither have guarantees on margins and you also have to increasingly build algo trading/option pricing expertise, regardless of what was your original secret sauce so that you can sell in the market. In effect, you start an adtech company with a nugget of a new idea/new data but you end up spiraling to a state where 80–90% of what you build is the same old rapid serving and bidding mechanisms.” This of course puts you in competition with everyone, further eroding margins and control.”
Gross Margins are the most important indicator of your moat
This is why gross margins are so important. If your company is growing revenue at a very rapid pace, you may think that you will still get paid for your toil. However, revenues are not the most important proxy or indicator of ultimate value. Gross margins are, perhaps, the best proxy for the defensibility of your business model and hence your value. As we all saw in 1999–2000, you can buy revenues with venture capital money. However, you cannot buy sustainable long term gross margins with venture capital driven revenue growth.
My partner Bill Gurley lays this out very well in his post on why all revenue is not created equal. And, I would add, ad tech and ad networks suffer from a bunch of the issues Bill describes.
Bill asks does your revenue or company have the following characteristics?
“Do you have a sustainable competitive advantage?” or as he quotes Michael Porter, ““How easy is it for someone else to provide the same product or service that you provide?”
Given the number of direct competitors in each category of Adtech, it is clear that there is not a long term competitive advantage here.
2. “Are there high switching costs? “
In ad tech, the answer is most definitely No. Ad Networks have made it so easy to embed their code and generate contextual ads or content that they have not only lowered their own barriers to win customers but lowered the barriers to enable competitors to easily compete with them.
3. “What kind of growth do you have?”
Writes Bill,
“So growth is good, correct? … While growth is quite important, and even though we are in a market where growth is in particularly high demand, growth all by itself can be misleading. Here is the problem. Growth that can never translate into long-term (emphasis added M.E.) positive cash flow will have a negative impact on a DCF model, not a positive one. This is known as “profitless prosperity.”
“In the late 1990s, when Wall Street began to pay for “revenue” and not “profits” many entrepreneurs figured out a way to give them the revenues they wanted. It turns out that if all you want to do is grow revenues, with disregard for the other variables, it is quite simple to “manufacture” awe-inspiring revenue growth. …There is another situation where growth can be misleading. If a company stumbles on to a hot new market, but lacks “barriers to entry” or does not have a sustainable competitive advantage, there will eventually be trouble. In fact, the very success of the first company in the field will act as a siren inviting others into the market, which, in the absence of a competitive advantage, will lead to margin erosion. Many electronics products follow this trend as some hot new product is quickly commoditized.”
Adtech companies follow this same trend as they too are quickly commoditized.
Are adtech margins defensible?
Adtech revenue is the revenue with the least defensible margins over the long term because you are sandwiched.
Let’s look at the lifecycle of an adtech company:
Often in adtech businesses, you discover a piece of screen real estate, an ad format or a data approach that gives you an edge. For a few years, you grow revenue quickly, exploiting your algorithm and market opacity. As we described above, you commit ever larger budgets to secure that inventory. Then, either competition or transparency begins to emerge and the publishers get the upper hand. Then, by bidding that format or real estate out to multiple bidders/providers and optimising the RPM (revenue per page) they get to do price discovery on exactly what his customer is worth. That is when margins begin to flatline or decline in adtech businesses.
Additionally, an astute Wall Street analyst covering adtech pointed out to me recently, that as each category of online becomes a larger portion of both the publisher’s and advertiser’s budget, “they tend to sharpen their pencils and claw back margin from the adtech providers.”
Compare the gross margins at Amazon to those in the adtech business. The chart below shows Amazon’s Gross Margins over the last 4 years. Why are they increasing? Because Amazon controls their customer directly and interacts with him directly. Amazon has built customer loyalty.
Ad Networks have neither B2B nor B2C customer loyalty. So lets compare Amazon’s margins with the flat gross margins in Adtech and Criteo’s down margins (which I think augurs a trend for the rest).
The green indicates the quarter of IPO. What you see is that margins increase up until the IPO. That margin increase shows momentum and may fool some investors. Thereafter, the margins flatline to decline as the opacity recedes and the publishers and advertisers understand how much more of the pie they can eat from the ad networks. In fact, when an adtech company or adnetwork publishes its S1, the first thing it does is expose its margin structures to publishers, advertisers and competitors who then know exactly what pricing to target. This is another reason why margins of public adtech companies flatline and decline.
When I first arrived at Benchmark 10 years ago, I set about trying to figure out what would make a company get an outsized growth multiple from Wall Street and I came to two conclusions: First, it needed to be unique in its business at scale. That is a proxy for building a Warren Buffet style moat and being the only way public market investors can play a key trend. The second was that it needed to have either steady or increasing gross margins. Those are the two factors that very smart Wall Street investors value. Adtech has neither.
An IPO is not an exit for engineers:
Using Bill’s model for the value of companies with high revenue growth but low quality revenue, the following chart will not surprise you. You can see here the 5 adtech companies that went public in the last 18 months.
Here is what you see. All but one of the companies is trading measurably below its IPO price (first day close and IPO price). On average, the group is down 49% from their IPO day prices. This is important because the IPO is not an exit for employees and engineers. Sometimes it is for founders but almost never for employees. This means that when you finally get released from your lock up, you are looking at a lower stock price. The stocks are trading down because Wall Street figured out that there are not moats around the business and that the next great algorithm or ad format would be bid against these companies by the publishers who control the customer.
Let’s look at how much capital each of these public companies raised and their current valuations. Remember, they need to raise capital to grow revenues and that capital is what sits on top of engineers. please note that this does not include the money raised on the IPO which is further dilutive! That reduces the multiple by almost 20% on each! With exception of Criteo, the multiples are not inspiring returns.
Additionally, the revenue to valuation of these companies shows the low value Wall Street accords to the revenue of these companies:
In my experience, some adtech companies even have complex preferred securities at the top of their cap table so even if you think the nominal valuation of the company when it is private is very high, the day it goes public, that preferred will collapse on the cap table and further bury the engineers who hope to get a return.
Turning To Israeli Adtech:
Now, lets look at the preponderance of Israeli adtech companies. My crowdsourced chart (please help finish filling it out) shows almost 60 companies. Mapped in Israel shows closer to 90 companies in adtech. The first interesting observation is that even in this list there are companies that compete directly with each other. As I mentioned above. In and of itself, that makes for a difficult financial play.
Outbrain: Taboola
Perion: Ironsource
Kontera: Infolinks
To name a few. Others are competing for the same real estate or budget with different formats.
Most of the Israeli adtech exits to date are in the tens of millions. The exceptions to that rule are: Conduit, Kontera and Amobee. Amobee raised $XX in capital. Kotera raised $XX in Capital and Conduit raised $8.4MM. With the exception of Conduit, none have been venture scale returns.
This chart shows the multiple on capital raised in each of those M&A Scenarios:
Company
$$ raised
Exit Value
Multiple =(Exit/$$ Raised)
Amobee
$72MM
$235,000,000
3.2X
Conduit
$8.4MMM
Dividends + PERI (M&A $640MM in PERI stock at acquisition time)
NA
Kontera
$36.1MM
$150,000,000
4.15X
The above ones are the best ones. By far. Most of the exits are in the tens of millions of dollars which does not yield a great result for engineers. Let’s assume that the average adtech company that sells for tens of millions raises only $5MM which I believe to be a pretty low number. Let’s also assume for this example that the company sold for $40MM. Also, a pretty aggressive assumption. Let’s assume that the average engineer has 0.1% of the company’s equity. When the company is sold, there are $35MM of proceeds (actually less due to 8% preference) so the equity held by the engineer is worth less than $35,000. That is still 3 months of salary or so but it is not a bonanza by any stretch of the imagination and this is before a company gets to scale and engineers are diluted by the massive capital needed to accomplish what I wrote above.
I reiterate that I believe that founders can make money in adtech by exploiting these business holes for a while. In addition, even as you raise capital, founders will likely be taken care of in an exit either by an acquiring company or the board of a start up. The ones left holding the proverbial “bag” in an adtech company are the engineers and the rank and file employees and they might end up looking back at years of blood sweat and tears with little to show for big revenue growth and a lot of effort.
You can do better!
There are a lot of problems in the world that the talented engineers of Israel (especially our core 8200 grads) can solve. Many of these problems and challenges will also reward you financially. The poll at the right from our Aleph Karma app is irrefutable. Great engineers want to believe in a product and mission to join a start up/ As we watch many of you go into adtech, we cannot help but wonder, who will solve the great challenges of the future and why would you waste the best earning-potential years of your life in a slog for profitless prosperity? Do you really believe in the product and mission that adtech offers?
Thanks to Oppenheimer & Co for help on some of the data.
Thanks to all those who reviewed the post before I published it: JW, RG, ES, ES, OR, JH and others.
* It is not an accident that Conduit, where I was an investor, raised less than $10MM. Ronen Shilo and team built a business that did not require capital to grow which is what made it investable as a VC investment. It is also one of the reasons that a well-thought-out Ronen went for dividends instead of going public at the time.
** Wix and Uber are both Benchmark companies and I served on the Board of Wix
