For more than a year, Facebook has faced a rolling public relations debacle. Part of this is the American public’s shifting attitudes toward Big Tech and platforms in general. But the driving problem has been the way the platform was tied up with and perhaps implicated in Russia’s attempt to influence the 2016 presidential election. Users’ trust in the platform has been shaken, politicians are threatening scrutiny and possible regulation, and there’s even a campaign to get people to delete their Facebook accounts. All of this is widely known and we hear more about it every day. But most users, most people in tech and also Wall Street (which is the source of Facebook’s gargantuan valuation) don’t yet get the full picture. We know about Facebook’s reputational crisis. But people aren’t fully internalizing that the current crisis poses a potentially dire threat to Facebook’s core business model, its core advertising business.
Facebook is fundamentally an advertising business. Almost all of the company’s revenue comes from advertising that it targets with unparalleled efficiency to its billions of users. In a media world in which advertising rates face almost universal downward pressure, Facebook’s rates have consistently risen. Monopoly power may drive some of that growth. But the key driver is efficiency. If old-fashioned advertising shows my advertisement to 100 people for every actual buyer and other digital platforms show it to 30 people and Facebook shows it to 5 people, Facebook’s ads are just worth a lot more.
As long as the rates bear some relationship to that efficiency (those numbers above are just for illustration), I’ll be happy to pay it. Because it’s objectively worth more. Indeed, as the prices have gone up, Facebook has actually gotten more efficient. As one digital ad agency executive recently told me, even if Facebook jacked up the prices a lot more, his firm would likely keep using them just as much because on this cost to efficiency basis it’s still cheap. This is the basis of Facebook’s astronomical market capitalization which today rates at over $450 billion, even after some recent reverses.
So the money comes from the advertising. And the advertising comes from the data and the artificial intelligence that crunches it and models it into predictive efficiency. But what if there’s a breakdown in the data?
Starting in a early March, a number of marketers running substantial sums on the Facebook ad engine, who’ve spoken to TPM, started noticing a new level of platform instability and reductions in targeting efficiency. To understand what this means, think about it like how an efficient debt or equity market operates. If there is relatively accurate information, no big external shocks and enough buyers and sellers, pricing should have relative stability and operate within certain bands. Accounting for some reasonable amount of bumpiness that’s what Facebook’s ad engine has looked like for a few years. But starting in March, if you’re down in the trenches working with the granular numbers, something started to look weird: price oscillations, reduced targeting efficiency and even glitches.
We’ve talked to a number of advertisers who’ve reported this. We’ve also talk to others who haven’t. But the ones who have tend to be the ones more tightly tied to the numbers and in marketing operations with tighter ROI (return on investment). Where we’ve seen the most of this is with so-called DTC (direct to consumer) marketers. Facebook is an amazingly large ecosystem. And it’s all a black box. So there’s no way for us to talk to a representative sample of advertisers. But something is going on in at least substantial sectors of Facebook’s ad engine. What is it? Marketers who’ve asked mainly get told it’s their creativity. In other words, the ad you’re running isn’t working. Come up with another ad. Here at TPM, we operate in a different part of the programmatic ad universe. You hear comparable things like that a lot. And it’s hard to ignore. But we’ve talked to people with different people with (by definition) different ads in totally different industries. So that’s not it. Something is happening.
So what’s up?
One thing is already being discussed widely in the trade press. In response to the rolling public relations debacle Facebook has already dramatically reduced or has announced that it will reduce advertisers’ ability to use third party ad data on the Facebook platform. That is a big deal. What’s that mean?
Let’s clarify some terms.
As you know, through your activity on Facebook, Facebook collects lots of data about you that it then uses to target ads. That’s “Facebook data” (or it’s yours, but you know what I mean). Facebook also allows advertisers to upload “1st party data”. What’s that? That’s if my book publishing company has a list of 50,000 emails, I can upload those emails to Facebook and run ads to those people. Then there’s “3rd party data”. That’s if the advertiser or Facebook itself goes to another personal data broker, buys access to that data and pours it into the Facebook ecosystem for more efficient targeting.
If you’re not versed in the world of data and digital advertising, there’s a ton here to keep up with. But here’s the key. How reliant is Facebook’s advertising cash cow on third party data? Not just the third party data that Facebook allows advertisers to put into its ecosystem for better targeting (which is now being phased out) but 3rd party data Facebook uses itself to improve its ad targeting? As one data industry executive put it to me, sure Facebook can crunch its own data to find out all sorts of things about you. But in a lot of cases it may be easier, cheaper and in some cases simply more effective to buy that data from other sources. We don’t really know – and no one outside Facebook really knows – how good Facebook’s AI really is at modeling user data entirely on its own without other sorts of data mixed in. It’s a black box. It matters a lot in terms of Facebook’s core revenue stream.
Here’s another question: when you consider Facebook’s own data, how reliant is Facebook on ways it collects and processes Facebook data which it may not be able to do any longer either because of new regulations that come into effect later this year for the EU or because of new regulations Congress may put into effect as it puts new scrutiny on Facebook’s behavior?
My hunch is that the answer to most or all of these questions is “a lot more than most people realize.” We already know that Facebook is making a lot of changes to how it uses data, especially third party data and how it allows advertisers to use data. Some of this is already public. Indeed, it’s getting discussed a lot in the trade press – particular how Facebook will implement with and cope with the new regs from the European Union. So why all the choppiness in Facebook’s advertising and targeting metrics? I suspect that Facebook is trying to rejigger its algorithm on the fly more than people realize in order to see if they can get it to work as effectively for ads without a lot of data sources or data uses they really aren’t supposed to be doing or which they suspect they’ll lose access to in coming regulation. That is the most logical explanation of the instability in their reporting.
If you talk to ad industry people, they treat it as a given that Facebook is already having to “rebuild their platform basically from the ground up” as one top agency executive told me, in response to “fake news”, propaganda campaigns, privacy scrutiny, etc. – all the stuff we’ve read about over recent months. But it’s Facebook. They’ll work it out, is what these people figure. And they’re probably right. Facebook is huge, has massive resources and access to the world’s largest audience for anything ever. They have oceans of data and a massive leg up on everyone. Down at the more granular level though, even in the industry press, it is treated as a given that the already publicly announced new restrictions on third party data will likely lead to at least some migration of advertisers to new platforms. Ginny Marvin, a top trade press reporter working at the granular ad tech and marketing level rather than up in tech big think land, tweeted this on March 30th: “FB removing 3P [3rd party] data is a big change for advertisers. But at FB’s scale, you’re not going to see advts sharply pivot elsewhere en masse. This will look more like a slow moving ship of budgets diverting to other media if they don’t get performance they want from FB.”
For now, as Marvin notes, Facebook’s advertiser lock-in, market power and simple price value make it highly unlikely that there’s going to be any dramatic near-term move from Facebook even in the worse case scenario. But Facebook isn’t just making money hand over fist. It’s market valuation rests on the assumption that it will keep making that amount of money hand over fist and indeed keep increasing the amount of money it makes hand over fist. Any breakdown or significant slowdown in that growth and consistency is a big problem. Years ago, everyone counted Facebook out as a true profit platform, until it exceeded everyone’s expectations. Now, even with all the bad press, most figure that it’s profitable forever. Both conventional wisdoms were wrong. For now, keep in mind that Facebook isn’t just dealing with a reputational crisis. It’s having to clean up the reputational mess by rejiggering parts of its core revenue stream it’s not clear it really knows how to do. That creates a lot of unpredictability. More than most people seem to realize.