Cambridge Analytica denies Facebook's claim that it accessed data on 87M Facebook users, claims it “licensed data for no more than 30M people”
Context & Ripple Effects
The number at the center of the scandal keeps moving. The original reporting traced how Facebook's old terms of service let a quiz app with just 270K users hand over data on 50M people; hours before this piece, Facebook itself raised its own estimate to as many as 87M, mostly in the US — its formal disclosure of the larger figure.
Cambridge Analytica's response reframes the dispute from 'did misuse happen' to 'how big was it': it denies touching 87M profiles and says it licensed data for no more than 30M. That gap matters because every downstream reckoning — regulatory and legal — keys off the scale Facebook asserts.
First-order effects
- Facebook and Cambridge Analytica are now on the record with irreconcilable counts (87M accessed vs. no more than 30M licensed), which weakens Facebook's disclosure as an authoritative account and hands skeptics a concrete inconsistency to press.
Second-order effects
- The contested timeline becomes evidence in legal proceedings: DC AG filings later allege Facebook knew about the misuse months before it was first reported — the court filing Facebook disputes — so the size disagreement feeds directly into arguments about what Facebook knew and when.
Third-order effects
- Even a smaller licensed-data count may not cap the blast radius: reporting suggests Cambridge Analytica claimed it wiped raw Facebook data in 2016 yet retained models derived from that data into 2017, pointing to a structural problem — 'deletion' claims are unverifiable when the value lives in trained models rather than the underlying records.
The trend: Platform data scandals are converging on an accountability fight over numbers and provenance — who counts the affected users, who verifies deletion, and whether platform disclosures can be trusted without independent audit.