What to Measure in Adgora Publisher Analytics to Know if Fill Rate Is Healthy

Start with the right health question, not the fill rate number

Fill rate is not one question. It is several. If you ask the wrong one, the number will still look neat, and you will still make the wrong call.

Start by naming the unit you are checking: overall inventory health, one ad unit, one country, one device type, or a 7-day window. A publisher who checks the whole account can miss a broken mobile placement in Brazil. A publisher who checks only one placement can panic over a normal weekend dip.

That is why the first step is scope. Write down the exact slice before you open Adgora analytics. Otherwise, you will compare a Tuesday desktop audience to a Sunday mobile audience and call it “fill rate trouble,” which helps nobody.

If you need a practical filter, choose one question per review. “Is the homepage healthy?” is different from “Is the sidebar healthy?” and different again from “Is inventory in Germany healthy?” Three questions, three answers, three possible fixes.

Check match rate, not just fill rate

Fill rate tells you how many requests became filled impressions. Match rate tells you how many requests found a buyer or demand source in the first place. Those are not twins. They often move together, but not always.

A healthy fill rate usually starts with a healthy match process. If requests are arriving and getting matched, the rest of the path has a chance. If requests are not matched, the fill rate number can still look acceptable for a while, especially in a small sample, while the actual supply path is already weak.

In practice, this means you should look for the gap between total requests and matched requests. If that gap is large, the problem may sit before delivery. If the gap is small but fill is still soft, the problem may be downstream, such as demand selection or ad rendering.

This is where the phrase “what to measure in Adgora publisher analytics to know if fill rate is healthy” becomes a real reporting question, not a title. The answer starts with requests, matched requests, and filled requests in the same view, because one number by itself does not tell you where the leak is.

Look at time-to-fill and delay patterns

A fill rate can look fine and still feel bad. If ads arrive after 2 seconds, or after 5 seconds, the page may already have lost the user. You have a technical fill rate, but not a useful one.

Check time-to-fill, response delay, and any timeout behavior shown in Adgora analytics. A placement that fills 90% of the time but only after a long wait is not healthy in the same way as a placement that fills quickly. The browser does not care about your dashboard, and neither does the reader.

Watch for patterns by hour. Morning traffic can fill quickly, while late-night traffic stalls because demand sources respond slowly. That kind of pattern matters because it is invisible if you only look at a weekly average.

One simple test helps: compare fast fills and slow fills in the same placement. If the slow group is much larger, you may have a time-based demand issue rather than a pure fill issue. Small delay, big effect.

Separate first-load behavior from repeat behavior

First-load requests are often different from refreshes, auto-refreshes, or repeated pageviews. Treat them differently. A homepage can fill well on the first visit and badly on the second request, or the other way around.

Why does this happen? Demand sources sometimes prefer fresh inventory, and some setups reward repeated requests more than the first one. A publisher who mixes both into one average can miss the real pattern.

Look at new pageviews, refreshes, and repeated requests as separate buckets if Adgora shows them. A first-load fill rate of 78% and a refresh fill rate of 96% do not mean the placement is perfect; they may mean the setup is leaning too hard on repeat behavior. That matters if your site has a lot of single-page visits.

Do not ignore the practical side. If a page only looks healthy because a later refresh rescues it, your user experience may still be weak. One load should not need a second chance.

Compare fill rate across placements and formats

One weak placement can drag down confidence in an otherwise healthy account. Start with the placements, not the account average. Then compare format by format: banner, native, push, or whatever is active in your setup.

Look for the placement that is always underfilling. If the footer fills normally, but the in-content unit keeps missing, you are not looking at a site-wide problem. You are looking at an inventory or setup problem tied to that spot.

This is also where format choice matters. A placement that works for display may not work for a native unit, and a display unit can behave differently from a push flow. If you run multiple setups, you can compare them with the same logic used in a क्रिप्टोक्यूरेंसी बैनर विज्ञापन गाइड, then ask which format is actually getting matched.

Do not average away the trouble. A 92% site-wide number can hide one placement at 54%. That is not a “small issue.” That is a placement worth checking line by line.

Segment by geography, device, and traffic source

Healthy fill is not always healthy everywhere. A strong US desktop audience can cover a weak mobile audience in India and make the account look better than it is. The average smiles. The segment does not.

Break the data by geography, device, and traffic source. If Adgora shows enough detail, compare at least three cuts: country, mobile versus desktop, and source type such as search, direct, social, or referral. One segment can drag down the whole account without meaning the full account is weak.

A publisher who sees poor fill on social traffic might have a quality mismatch, while a publisher who sees poor fill on one country might have demand limitations there. The fix is different. So is the budget impact.

Do not assume every low segment is a problem. Some segments are supposed to be less profitable. The useful question is whether the low segment is isolated or spreading. That is the difference between a known pattern and a real warning.

Watch for demand concentration and dependency

A fill rate can look healthy and still be fragile. If one buyer, one deal, or one demand source is doing most of the work, the account depends on that source staying active. When it slips, the fill rate slips with it.

Check whether a single demand source is carrying the majority of filled impressions. If the answer is yes, the headline number is less stable than it appears. That is not theory. It becomes obvious the day one buyer pauses spending.

This is the kind of issue that matters in crypto-heavy traffic, too. If one source is propping up the account, the setup may look fine today and shaky next week. For publishers running varied traffic, a related read like क्रिप्टो विज्ञापन नेटवर्क लक्ष्यीकरण विकल्प गाइड can help frame how demand concentration happens in the first place.

Dependency also hides in deals. A private deal can fill well for two weeks, then go quiet. If the dashboard only shows the combined result, you may think the account softened for no reason. There was a reason. It was just sitting in one source.

Use a simple healthy-vs-warning checklist

Keep the review simple. Use the same checklist every time, and check it by placement, country, device, and source. If you change the checklist every week, you will confuse the chart before the chart confuses you.

First, confirm the scope: account, placement, geography, device, or traffic source. Second, compare requests, matched requests, and filled requests. Third, look at time-to-fill and delays. Fourth, split first-load behavior from repeat behavior. Fifth, compare placements and formats. Sixth, check concentration in one buyer or one demand source.

Now classify the result. If fill is stable across segments, fills arrive quickly, and no single source carries the load, the account is probably healthy. If one placement underfills, one segment underperforms, or one demand source does too much work, it needs investigation. Three warning signs are enough.

Use the checklist the same way each time. That consistency matters more than a flashy number. A fill rate of 85% can be healthy in one setup and poor in another, depending on delay, match quality, and demand spread. The number alone does not decide.

For publishers who also run other formats, the same logic applies across ad types. A solid fill pattern in display does not mean push or native is equally steady, which is why a cross-format reference like क्रिप्टो ऑफ़र के लिए नेटिव विज्ञापन can be useful when you compare the pattern across units.

If you want a quick final pass, ask four things in order: Did requests match? Did matches fill quickly? Did the result hold across placements and segments? Did one source carry too much weight? If any answer looks off, the fill rate is not healthy yet, even if the account average still looks tidy.