- The company (NASDAQ:FB) notes it's focused on incorporating greater third-party validation measures and also updating internal review processes. However, in an effort to provide clarity on the matter, cites various further miscalculations across certain products and features.
- On one Page Insights dashboard, 7-day and 28-day organic page reach was miscalculated by reporting sum of daily reach as opposed to de-duplicating repeat users over the given timeframes. De-duplicated summaries for 7-day and 28-day options will be, on average, 33% and 55% lower, respectively. This issue has been live since May. Reach counts going forward will be based on viewable impressions (similar to paid reach), resulting in stricter definition, and consequently, 20% lower reported reach on average.
- The "video watches at 100%" metric has also been impacted, attributed to audio and video tracks not matching up and differences between various devices. The fix to this issue, however, will result in an approximate 35% increase in count for the metric.
- For Instant Articles, average time spent per article was over-inflated by 7%-8% on average since August 2015. This issue is claimed to be fixed.
- Within the company's Facebook Analytics for Apps dashboard, "Referrals" are on average for top users considered to have been overrepresented by 6%. A fix is ongoing.
- Interest lists and associated follower metrics are to be discontinued. Facebook estimates a 5% or less drop in followers for profiles utilizing the feature.
- Metrics FYI (new blog similar to the Facebook's News Feed FYI)
- Improvements to Facebook Ads Reporting
- Previously (September 23): Facebook inflated video metrics for two years
- Previously (September 23): Facebook -1.5%; executive issues statement on metric miscalculation
Original article