Last month, in a fairly dramatic spectacle, two anti-trust lawsuits were filed against Facebook on the same day: one from the Federal Trade Commission, and one from the Attorneys General of dozens of US states. The suits are multi-faceted, and this post isn’t about the details of the arguments they make, but both are centrally focused on the notion that consumer harm has been inflicted through regressive user privacy: Facebook initially competed with rivals on the basis of superior privacy controls, and the power of those controls deteriorated over time as Facebook eliminated its competition. This argument is best articulated in the canonical paper on the subject by Dina Srinivasan, The antitrust case against Facebook.
Both the US States and the FTC are suing Facebook under Section 2 of the Sherman Act, which relates to the abuse of monopoly power, but the States are also suing Facebook under Section 7 of the Clayton Act, which relates to M&A activity for the purposes of reducing competition. The FTC suit seeks to unwind Facebook’s acquisitions of WhatsApp and Instagram and force their divestiture, breaking apart the company into three separate entities. Facebook has called the government’s claims that its acquisitions reduced competition “revisionist history” since the FTC declined to take actions against those acquisitions. But as legal scholars Tim Wu and Scott Hemphill point out, the FTC doesn’t “approve” mergers but rather either takes action or doesn’t — and not taking action does not prevent future enforcement.
These lawsuits are monumental, and their outcomes will likely dictate how the consumer technology landscape is shaped for decades to come. Putting aside the core content of the suits relating to user privacy — which is an important consideration and deserves a dedicated treatment — it’s interesting to think through the consequences of a Facebook breakup on advertisers. How would the partitioning of Facebook into three separate companies change the advertising ecosystem?
A first pass of this exercise requires acknowledging how advertising changed when these companies were combined. As I wrote in 2017 in Was it a mistake for Instagram to sell to Facebook?, Instagram gained a great deal when it was subsumed into Facebook:
The second thing to unpack is an understanding of what business Instagram, Facebook, Snapchat, et al are in. These companies serve ads: that’s how they make money. And Facebook is better than almost any company on the planet at turning behavioral and demographic user data into insight that can be advertised against in a visual feed. Feed-based advertising is Facebook’s home turf: competing against it is very difficult, but partnering with it can unlock untold value from feed real estate. Would Instagram be worth as much as it is today without Facebook’s data infrastructure, ad serving tools, and product support? Almost certainly not. But would it be worth more than $1BN? After all, Snap is.
The union of Facebook and Instagram unlocked vast value: Instagram’s ad impressions combined with Facebook’s advertising infrastructure and core user data set created powerful economic opportunities for advertisers, from game developers to fashion labels to DTC brands. How much is Instagram ad inventory worth if Facebook is no longer operating its auction “plumbing”?
It turns out that Facebook’s advertising plumbing — its targeting mechanisms, engagement identity system, and conversion probability models, as I outline in this QuantMar thread — has a tremendous impact on the price that advertisers are willing to pay for the inventory that it traffics. Facebook estimated that 50% of the CPM delivered to publishers on its Facebook Audience Network (FAN) is derived from its ability to personalize ads. That number is equally applicable to Instagram.
Facebook’s pixel is ubiquitous on the web and allows for ad attribution and conversion tracking, although Apple via ITP has made inroads into curtailing Facebook’s ability to track users across sites over long periods of time. It is not just historical conversion data that Instagram would lose access to were it orphaned from Facebook: it would also have to replicate these measurement touchpoints with advertisers in order to attribute ad performance.
Although Facebook — partly as a result of ITP and in anticipation Google’s upcoming blockage of third-party cookies in Chrome — has attempted to move as much of the conversion process into its apps as possible in order to preserve attribution. Instagram Shopping, Facebook Shop, Instant Articles, Instant Games — all of these experiences take place within the app and without having to rely on click tracking or conversion measurement (first-party cookies are dropped into instant article pages, for instance). That data is all centralized by Facebook and indexed with the Facebook ID, and Instagram would lose access to that in a divestiture, meaning advertisers would lose access to it.
Facebook is able to absorb these interactions into the app experience because of its size: retailers and publishers are willing to have their relationships with consumers intermediated (reluctantly) because of the increased efficiency of in-app content interaction. But would they be so generous with an independent Instagram?
If it is spun out of Facebook, Instagram will need to build all of this plumbing from scratch. This would be an enormous undertaking. Snapchat didn’t launch lookalike targeting, which is table stakes for digital display advertising, until 2016 — five years after its launch. If Instagram is either divested as an independent entity or acquired or merged with another social platform, it might take years to replace the capabilities it currently provides to advertisers.
This would be massively disruptive to the advertising ecosystem. Instagram is one of the most important advertising channels on mobile. Eliminating Instagram’s targeting, measurement, and identity infrastructure would remove billions of dollars from the digital economy: Instagram did $20BN in revenue in 2019, more than YouTube. Advertisers would struggle to re-route the budget they currently allocate to Instagram to other channels in the case that Instagram was cut off from Facebook’s infrastructure, and they would sustain a real economic loss as a result.