/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

Analysis: Facebook's estimates for active fake accounts it reported in filings each quarter was significantly smaller than the number it said it had taken down

Jack Nicas / New York Times : Tweets: @jacknicas , @jacknicas , @jacknicas , @jacknicas , @jacknicas , and @jacknicas Tweets: Jack Nicas / @jacknicas : New: In the small print at the bottom of its earnings slides, Facebook just disclosed its count of fake accounts on its site jumped 27% to 116 million, while duplicate accounts jumped 12.4% to 255 million. Our story on its dubious fake-account figures: https://www.nytimes.com/... Jack Nicas / @jacknicas : The figures also imply Facebook takes down roughly 90% of fake accounts on its site. Facebook said such success was possible because most of its removals are of bots that are easy to spot. It takes down most within minutes of registration. Manually created fakes are tougher. 6/ Jack Nicas / @jacknicas : Then everything got murkier when Facebook disclosed last year that it took down 2.8 billion fake accounts in a recent 12-month period, or 7.7 million a day, revealing the scope of the problem was far larger than the estimates in its SEC filings had suggested. 4/ Jack Nicas / @jacknicas : Facebook had previously said about 3% to 4% of its accounts were fake. The new figures showed the accounts taken down each quarter were equivalent to 25% to 35% of its active users. (The new figures aren't counted in the active-user tallies because they've been removed.) 5/ pic.twitter.com/LASpuL3ehy Jack Nicas / @jacknicas : And we haven't even yet mentioned duplicate accounts, which are more numerous than fakes, according to Facebook's figures. Its estimates of duplicates have also seesawed: 8/ pic.twitter.com/NImqntWW6q Jack Nicas / @jacknicas : First, take a look at how Facebook's estimates for fake accounts, made in securities filings, have oscillated in recent years. Facebook cut its estimate significantly in 2016, then more than quadrupled it a year later. 2/ pic.twitter.com/nf4pGXPzk3

New York Times Jack Nicas

Context & Ripple Effects

Facebook has been publishing ever-larger takedown figures for years: it reported deleting 1.3 billion fake accounts in just six months in 2018 while admitting creation was outpacing deletion, and later said it banned 2.19 billion fake accounts in Q1 2019 alone. What was missing was the other side of the ledger — how many fake accounts it believed were actually live at any moment.

This analysis closes that gap by reading the small print of the earnings slides: alongside the SEC-filed estimates, Facebook disclosed 116 million active fake accounts (up 27%) and 255 million duplicate accounts (up 12.4%) — implying roughly 90% of fakes are caught, and exposing how much smaller the filed 'active' estimate is than the billions cited as removed. The last time Facebook broke out account-quality estimates this way, in 2017, it put duplicates at ~207 million and fakes at ~60 million, so both categories have grown sharply since.

First-order effects

  • Investors and the SEC now have two materially different Facebook numbers on the record for the same problem — a small active-fake estimate in filings versus multi-billion takedown counts — inviting questions about which figure the company actually manages to.
  • Facebook's own framing cuts against it: if ~90% of fakes are removed, the remaining 116 million still exceed the entire user base of most major platforms, and the 12.4% jump in duplicates shows the problem growing, not shrinking.

Second-order effects

  • Advertisers get a basis to challenge reach and engagement pricing built on monthly-active-user counts that include hundreds of millions of duplicate and fake accounts, pressuring Facebook's core revenue metric.
  • Rival platforms face the same disclosure logic: once one company publishes both stock and flow figures for fake accounts, peers' enforcement-only reporting looks evasive by comparison.

Third-order effects

  • If the pattern holds, platform-integrity reporting drifts from voluntary quarterly estimates toward standardized disclosure — the kind of metric definition regulators and auditors eventually step into when companies file inconsistent versions of the same number.
  • The takedown-arms-race dynamic Facebook describes — fakes created faster than they can be deleted — structurally favors measuring removal volume over prevention, meaning headline enforcement numbers can rise even as the underlying fake-account population grows.

The trend: Platform integrity is moving from opaque self-assessment toward dual-metric disclosure — stock (live fakes) versus flow (removals) — and the gap between the two is becoming the story regulators and advertisers watch.