Google is preparing a significant measurement change for AdSense publishers, and the first impact may show up as a visible drop in reported display ad impressions. Beginning February 17, 2027, Google says AdSense will move display ad impression counting from the current count-on-download method to a begin-to-render approach, meaning an impression will be recorded only when an ad has successfully loaded and started to render on the user’s device.
The change was documented in a new Google AdSense help page and reported by Search Engine Land. Google’s stated reason is consistency: display ad measurement will be brought closer to the begin-to-render methodology already used, or supported, across native, app and video inventory in AdSense and Google Ad Manager. For publishers, however, the practical result is that some impressions previously counted will disappear from reports because the ad started downloading but never actually appeared on screen.
What is changing in AdSense reporting
Under the current count-on-download model, AdSense can register a display impression when an ad begins downloading to a user’s device. That is an earlier point in the ad delivery chain, and it can include cases where the visitor leaves the page, closes a tab or navigates elsewhere before the creative has a chance to render. Under begin-to-render measurement, that gap becomes important because Google will wait until the ad has loaded and begun rendering before counting the impression.
This is not a policy penalty, an account-level enforcement action or a signal that ads will stop serving. It is a reporting methodology change. The same page may continue to request and serve ads, but some requests that previously ended up as counted impressions may no longer qualify. That distinction matters because publishers often use impression volume to assess RPM, inventory performance, layout experiments, fill behavior and the relationship between traffic and revenue.
Why publisher numbers may fall
The most immediate effect is likely to be cleaner but lower impression totals for display inventory. If a reader lands on a page and quickly leaves, if a slow connection delays ad rendering, or if a layout causes an ad to begin downloading late in the user session, those events may no longer appear as impressions unless the ad actually starts to render. Google has not provided a universal percentage estimate, which means the size of the change will depend heavily on site speed, page structure, ad placement, audience behavior and device mix.
Publishers with fast-loading pages and ad slots that render reliably near the user’s active viewport may see a relatively modest reporting shift. Sites with heavier templates, slow scripts, aggressive lazy loading, unstable ad containers or high bounce rates could see a more noticeable drop in reported impressions. The change also means year-over-year comparisons after February 17, 2027 may become harder to interpret unless publishers annotate their analytics and reporting dashboards around the transition date.
Revenue impact is less straightforward than impression impact. Fewer counted impressions can make CPM-style metrics appear different, but the underlying value of impressions that actually render may be stronger than inventory that merely started to download. In practice, publishers should watch not only total impressions but also revenue, page RPM, impression RPM, active view metrics, ad request trends and user engagement after the switch. A decline in impressions alone will not necessarily tell the whole story.
A measurement shift toward real delivery
The move fits a broader advertising trend away from counting technical ad calls and toward counting events that better reflect whether an ad had a real chance to appear. Begin-to-render is still not the same as a fully viewable impression, since rendering does not guarantee that the user looked at the ad or that it met viewability thresholds. Even so, it is a stricter standard than counting the moment an ad begins downloading, and it removes a category of impressions that were vulnerable to disappearing before the creative became visible.
For advertisers, the change should make reported display impressions slightly more meaningful because the count moves closer to actual creative delivery. For publishers, it may create a short-term perception problem if impression totals fall while traffic remains stable. The important point is that lower numbers after the transition may reflect a measurement reset rather than a sudden deterioration in audience demand or ad serving health.
How publishers should prepare
Publishers have several months before the February 2027 transition, and the best preparation is to create a clean baseline now. Teams should document current impression trends, page RPM, ad request volume, viewability, Core Web Vitals, bounce behavior and template-level performance before the methodology changes. After the switch, those same reports should be reviewed with a clear note that AdSense display impressions are being counted differently.
Technical preparation should focus on reducing the gap between an ad request and a rendered ad. Faster pages, stable ad containers, fewer blocking scripts and better lazy-loading logic can all help ensure that ads that are requested have a better chance of rendering before a user leaves. Publishers should also review whether important ad units are positioned or loaded in ways that make them unusually vulnerable to abandonment before render.
The reporting change is ultimately a reminder that not every ad call has the same commercial value. From February 17, 2027, AdSense display reporting will place more weight on impressions that actually begin to appear, even if that means some dashboards look weaker at first glance. Publishers who understand the methodology shift in advance will be better positioned to explain the numbers, benchmark performance fairly and optimize for ad experiences that load quickly enough to count.