Fill rate is one of those metrics that looks simple on the surface — a percentage — but hides a lot of diagnostic information about the health of your ad setup. Publishers who don't monitor it regularly often leave significant revenue on the table without realizing it.
What is fill rate?
Fill rate is the percentage of ad requests that result in an ad being served. If your page sends 1,000 ad requests and 820 of them result in a displayed ad, your fill rate is 82%. The remaining 18% are unfilled impressions — slots that loaded but showed no ad.
Ad requests are generated whenever an ad slot on your page makes a call to the ad network. Not every request results in a fill because demand isn't infinite and not every impression is appealing to every advertiser. Some requests come back with no qualifying bid. Others are rejected by the ad network for policy reasons. Others time out before an ad loads.
Fill rate vs. viewability
Why fill rate matters
Every unfilled impression is a missed revenue opportunity. If you have 100,000 daily page views with two ad slots each, that's 200,000 ad requests per day. At an 80% fill rate, 40,000 impressions go unsold. If your average CPM is $3.00, that's $120 per day — or roughly $44,000 per year — in revenue that never materializes.
Fill rate also affects ad network relationships. Networks track fill rates across publishers as a signal of inventory quality and demand alignment. Chronic low fill rates can result in reduced priority or reduced demand allocation from some networks.
Causes of low fill rate
Floor price set too high
The most common cause of low fill rate is a floor price that exceeds what the market will pay for your inventory. If your floor is $5.00 CPM and the genuine demand for your audience is $2.00–$3.00, most requests will come back empty. The fix is to test lower floors and find the level where demand becomes available without sacrificing too much per-impression value.
Geographic audience mismatch
Advertiser demand is heavily concentrated in a handful of markets — primarily the US, UK, Australia, and Canada. If a significant share of your traffic comes from markets with thin programmatic demand (parts of Southeast Asia, Africa, or Latin America), those impressions may frequently go unfilled simply because few advertisers are bidding there.
This is less a problem to fix and more a factor to account for. Segment your fill rate analysis by geography before drawing conclusions about your overall setup.
Too few demand sources
If you're running a single ad network, you're leaving demand on the table. Each network has a different pool of advertisers, different category focus, and different geographic strength. Running multiple demand sources — ideally through a header bidding setup — means more bids per impression and higher fill rates because more advertisers are competing for your inventory.
Slow ad loading and timeouts
Many ad requests are configured with a timeout: if the ad network doesn't respond within a set window (often 1–2 seconds), the request is abandoned and the slot goes unfilled. Slow auction infrastructure — common with client-side header bidding under load — causes impressions to time out before a bid comes back.
Server-side header bidding eliminates most timeout-related fill loss because the auction runs on dedicated infrastructure rather than in the user's browser.
Brand safety and contextual filters
Advertisers apply brand safety filters that exclude content categories, keywords, and topics they don't want their ads adjacent to. Publishers covering news, politics, or controversy-adjacent topics often see lower fill rates because more advertisers' filters exclude their content.
How to improve fill rate
- Add demand sources.If you're running one network, add two or three more through a header bidding setup. Each additional demand source increases the probability that someone bids on any given impression.
- Test your floor price. If fill rate is the primary concern, lower your floor incrementally and monitor whether RPM holds up. Often a modest floor reduction (10–20%) recovers a disproportionate amount of fill.
- Increase timeout limits.If you're seeing a significant number of timeouts in your ad network reporting, increasing your request timeout from 1s to 1.5s can recover impressions without noticeably affecting page performance.
- Use house ads.Configure house ads or promotional content to fill slots that don't clear from paid demand. A house ad that promotes your newsletter or a popular article is better than a blank space.
- Audit your placements. Slots deep in the page — below 5,000 pixels — see significantly lower demand because viewability is lower. Consider removing or consolidating placements that consistently underperform.
Fill rate vs. RPM: which to optimize for
Fill rate is a means, not an end. The goal is revenue per thousand page views (RPM), and fill rate is one of the inputs. Optimizing fill rate at the expense of CPM — for example, removing your floor to hit 100% fill — almost always reduces total revenue.
The right target is a fill rate that maximizes RPM given your audience, content, and demand mix. For most publishers on a well-configured header bidding setup, this lands somewhere between 85–95%. Below 75%, something structural is usually wrong. Above 98%, your floor is probably too low.
Fill rate is a diagnostic metric — it tells you whether your ad setup is functioning as intended, not whether it's optimized. Start by benchmarking your current fill rate, segment by geography and placement, identify the dominant cause of unfilled impressions, and work through the fixes methodically. Most publishers can recover 10–20 percentage points of fill with a few weeks of focused attention.