DSA Transparency Reports — VLOPs & VLOSEs

Reading the data

• Notices (Table 4) are submissions from users or trusted flaggers reporting allegedly illegal content. A single notice can reference multiple items.
• Own-initiative actions (Tables 5 & 6) are measures taken proactively by a platform without a notice — Table 5 covers illegal content, Table 6 covers policy violations.
• Government orders (Table 3) are orders from EU member states to act against content or provide user information.
• Appeals (Table 7) covers complaints submitted to a platform's internal mechanism and out-of-court dispute settlement bodies.
• Automated means (Table 8) reports volume of moderation actions taken by automated means versus human review, along with accuracy, precision, and recall metrics per service and language.
• Human resources (Table 9) reports the number of internal and external content moderators employed by each platform, including a breakdown by language.
• User reach (Table 10) reports the average monthly active recipients (AMAR) per EU member state. Some platforms report approximate figures: Booking.com uses ranges (midpoint shown); Temu uses "M" notation; Google services link to a separate MAR report and are omitted. Values below 1,000 in the source are treated as millions.
• TOTAL rows aggregate all content categories and are shown when "All categories" is selected.
• Surface (Tables 6, 7 & 8) — Google reports these as several disjoint sub-reports per service (organic "Core", "Ads", and for Search a breakdown by action level: URL-, domain-, host-level, etc.). They are summed by default; filter to Google (or one of its services) on those tabs and a Surface control appears to isolate a single breakdown.
• Empty cells in the source data are treated as zero or not applicable. Contextual notes from individual platforms are not included in this view.