In a sort-of-recent podcast1 from In Bed With the Right, Adrian, one of the hosts, talks about Meta’s big, successful acquisitions but then points out some of Meta’s less successful acquisitions. He argues that Meta’s strategy has been somewhat haphazard and inconsistent. But, the three examples he gives seem almost to argue the opposite, that Meta’s acqiusitions are often quite strategic, giving the company either a critical knowledge advantage, key technical talent, a technology / user base moat, or combinations of these.

Let’s run down the three companies:

  • Oculus - I don’t think Oculus has made a profit yet, but Oculus has utterly changed the direction of the company. Zuck even changed the name of the company to highlight Oculus’s efforts to make the Metaverse into something real.2

  • Onavo - The data from Onavo was apparently instrumental to the WhatsApp acquisition - that probably justifies the acquisition alone. But, even when I got to Facebook, the Onavo data was still reliable and was the singular data source for very subtle questions3. These analyses were how some really nice new projects got approved. (And Facebook Dating, but that is a different story.) Also, it was not that expensive; the estimate is $[120M, USD, 2013], or about $150M in 2025 dollars. It might be justifiable simply for that Guy Rosen came with it, and he’s been fundamental in Meta’s trust and safety strategy.

  • Beluga - This was a small acquistion; according to Claude, it was a team of 2-3 people, and a SRP of $[15M, USD, 2011], about $20M in 2025 dollars. Yet, it was the foundation for Facebook Messenger for many years, and was only replaced as part of the push for end to end encryption (or maybe it was cross-platform messaging4?).

I had a some passing exposure to purchase decisions, and it was not particularly haphazard. Further, I think most companies would be very happy with these purchases.

  1. I can’t remember how I found In Bed With the Right, but it’s a consistently well-made and thoughtful and insightful. More than once, I’ve had a question about something, and discover it’s the next episode in my queue. (Most recently, it was about Roy Cohn’s influence on Trump.) Give it a try. ↩

  2. In 2015, a lot of the executives at Facebook were telling us (UXR) to read Ready Player One to understand the potential of VR. There was a samizdat PDF of the book that was circulating, so I asked for a copy. One of the thigs that struck me about the book was that it was set in a remarkably diminished US. The glide path down was never really explained, but it must have been quite extreme. Every so often, I wonder if Trump is trying to re-create the those conditions. He’s doing a great job so far. ↩

  3. For instance, one of my candidate projects for Data Camp was to look at Grindr usage at churches. The challenge here turned out to be two-fold, neither of which was Onavo-related: first, there was a data quality issue where different regions coded “church” in contrdictory ways. “Church” here came from a Facebok table that would join to the Onavo data. Second, the query was not fast. There was no straight-forward way to sample the data, which meant I was doing many full-table scans of big-even-by-facebook-standards tables. At some point, I’ll write up my argument about how Facebook’s fundamental data model make some questions very hard to answer. ↩

  4. I don’t have good enough notes to write a full post on this, but cross-platform messaging was barely used, and where it was used, it was used by scammers. ↩

Updated: