Facebook says it deleted 1.3B fake accounts in six months; fake accounts are being created, probably for commercial purposes, faster than Facebook can delete
Sheryl Sandberg's testimony to Congress revealed that fraudulent pages are being created as fast as the social network can delete them. Tweets: @goldinagency Tweets: Goldin Agency / @goldinagency : Another incisive piece by @sivavaid, this time for @nytopinion, on Facebook's uphill battle to combat fake accounts: “we judge exterminators not by the number of roaches they kill, but by the number that survive.” http://www.nytimes.com/...
Context & Ripple Effects
Sheryl Sandberg's testimony to Congress put a number on what had been an abstract problem: Facebook's own detection improvements had already pushed its fake-account estimate from roughly 1% to 2-3% of the network, and the company was now deleting 1.3B accounts in six months without denting the supply. The reason is structural rather than operational — as a BuzzFeed investigation found months earlier, fake accounts are cheap, always in stock, and sold openly on marketplaces, so deletion is fighting a replenishment loop, not a fixed stock.
First-order effects
- Congress now has Sandberg on record admitting creation outpaces deletion, making Facebook's quarterly integrity numbers the de facto measure of whether the platform can be trusted at all.
- Advertisers and election officials face the immediate implication that any audience or engagement metric on the platform includes an unknown share of commercially motivated fakes that no cleanup has caught.
Second-order effects
- The gap between takedown counts and the active-fake estimates Facebook reports in filings becomes a credibility liability in itself — later analysis showed those two figures diverged significantly, inviting regulators to ask which number reflects reality.
- Commercial fake-account sellers face no demand shock from deletions, since each purge simply reopens inventory slots their marketplaces refill, keeping pricing low and supply elastic.
Third-order effects
- If the pattern holds, enforcement becomes an arms race measured in billions — Facebook's own later disclosures show takedowns climbing past 2.19B per quarter and then 6.6B a year once machine-learning detection scaled — while the underlying question shifts from 'can fakes be removed' to 'can authentic activity be distinguished at all'.
- A platform whose integrity depends on perpetual automated deletion invites external oversight: dedicated monitoring operations like the Dublin election war room foreshadow regulators treating platform hygiene as a compliance function rather than a product choice.
The trend: Platform integrity is becoming a treadmill economics problem, where detection scale determines reported takedowns but commercial demand for synthetic accounts keeps supply ahead of deletion.