What the filtered-traffic categories on your reporting mean, and why those impressions did not earn.
We record every ad we render, then run each one through invalid-traffic checks. An impression that fails a check is filtered: it still appears in your counts, but it does not earn, and the advertiser is not charged for it. This keeps both sides honest — you are never paid for traffic an advertiser would later dispute, and the same event log reconciles on both sides.
Your reporting shows filtered impressions grouped into a few plain categories. We show the category, not the exact rule or threshold — publishing the precise trigger would just teach bad actors how to pace under it. The categories are:
| Category | What it means |
|---|---|
| Automated | The request came from a known bot or automated user-agent, not a person. |
| Rate-limited | Too many impressions from the same network or the same viewer in a short span — a volume pattern real users don't produce. |
| Duplicate | The same viewer re-rendered the same ad in the same zone inside our de-duplication window. |
| Rate protection | Serving on the zone was automatically paused for a cooldown because its rate spiked; impressions during the pause don't earn. See Automated rate protection — it's temporary and not a penalty. |
| Other | A filter reason that doesn't fit the categories above — usually a newer check. |
A small, steady filtered share is normal — no ad network delivers 100% valid traffic. A category that suddenly climbs is worth a look: it usually points at a traffic source that changed. If Automated or Rate-limited jumps on one zone, check where that zone's traffic is coming from.
Country and network blocks are counted separately and are not included in these numbers — a blocked request never renders an ad, so it never becomes an impression here.
If a filtered figure looks wrong, contact us with the zone and date range and we'll walk through it. For how billing and counting work end to end, see How we count & filter traffic.