Fill & Backfill
No-fill inventory earns nothing until it's filled. This page shows the no-fill rate per placement and the revenue that backfilling recovered — revenue that did not exist before.
Web no-fill is observed directly through the Prebid.js bid landscape in real time — every no-bid and timeout is visible. App no-fill comes from SDK waterfall callbacks. Neither source is estimated; the no-fill counts themselves are exact requests, not samples.
AdLynk running as a network adapter in your own mediation. Arrives on the same impression callback as every other network, so it appears in your mediation reporting line by line. This is the figure in the table below.
Weighted by no-fill requests across all placements.
Measured rows have user-level revenue to attribute recovery to. Modelled rows sit on aggregate-only demand.
Daily recovered revenue
14-day trend, all placements combined.
Two kinds of backfill
They differ in one way that matters: whether you have your own record of them.
Through your mediation
AdLynk network adapter (added by you, in your mediation). Revenue arrives on the same impression callback as every other network, so it shows up in your own reporting. Check it against Your AdMob / MAX / LevelPlay mediation reporting, by ad unit and network.
Direct via our SDK on no-fill
Never touches your mediation, so nothing in your stack records it. Reported on its own line below, never blended into totals, backed by a monthly statement, an impression-level log and a contractual audit right.
No-fill and recovery by placement
Sort any column. Modelled rows sit on aggregate-only demand (Meta AN, AdX) and cannot be attributed per user.
| Placement↕ | No-fill requests↕ | No-fill rate↕ | Backfilled requests↕ | Fill % after↕ | Backfill eCPM↕ | Recovered (30d)▼ | Basis |
|---|---|---|---|---|---|---|---|
Interstitial — session end app | 6.1M | 9.8% | 4.5M | 96.2% | $2.10 | $21,900 | measured |
Rewarded — level complete app | 1.8M | 4.4% | 1.4M | 99.1% | $3.80 | $18,640 | measured |
Web — article top web | 10.7M | 23.6% | 7.9M | 94.4% | $1.40 | $13,480 | measured |
Native — feed slot 3 app | 5.9M | 20.9% | 3.2M | 93.1% | $1.60 | $9,840 | measured |
Web — in-article web | 14.6M | 28.2% | 10.2M | 92.6% | $0.72 | $8,960 | measured |
App open app | 2.4M | 11.1% | 1.7M | 97.3% | $2.90 | $7,120 | measured |
Native — content feed (Meta AN) app | 2.2M | 12.4% | 1.3M | 91.8% | $1.10 | $4,920 | modelled |
Banner — home app | 3.3M | 1.8% | 2M | 99.4% | $0.31 | $4,260 | measured |
Web — sticky footer web | 7M | 15.8% | 4.6M | 96.8% | $0.48 | $3,540 | measured |
Web — footer (AdX aggregate) web | 3.9M | 18.7% | 2.3M | 90.4% | $0.39 | $2,180 | modelled |
Direct backfill via our SDK — reported separately
Served by our SDK when your waterfall returns no fill. These impressions never enter your mediation.
This is the one number on the platform you cannot verify independently. These impressions never touch your mediation, so your mediation dashboard has no record of them. We keep it on its own line and never blend it into the totals above, so nothing else on this page depends on you taking our word for it.
Excluded from recovered revenue above and from every blended total.
At $1.30 eCPM. All of it on requests your waterfall could not fill.
| Placement | Impressions | Revenue 30 d | Record in your mediation |
|---|---|---|---|
| Web — in-article | 6.4M | $7,920 | none |
| Web — article top | 4.2M | $6,140 | none |
| Native — feed slot 3 | 3.3M | $4,860 | none |
| Interstitial — session end | 2.9M | $3,410 | none |
| Web — sticky footer | 1.8M | $1,850 | none |
Audit right: You may audit these impressions at any time, with 10 business days' notice, against our raw logs. The right is in the contract, not a courtesy. Impression-level logs are available on request at any time, not only at audit.
Recovered revenue is measured directly wherever the backfilled demand carries user- or impression-level revenue (AdMob, Ad Manager, MAX, LevelPlay, Prebid). Where the backfilled demand is Meta Audience Network or AdX aggregate delivery, recovered revenue is modelled from placement x country day-over-day deltas and marked "modelled" above — treat those two rows as directional, not exact.