On February 17, 2027, Google AdSense switches how it counts a display ad impression, moving from "count-on-download" to "begin-to-render." The practical effect for publishers: reported impressions, and likely earnings, will look lower after that date, even though nothing changed about how much traffic the site actually gets. No technical action is required to enable it, but every publisher running AdSense should understand the shift before the numbers move.
This is not a bug, a policy strike, or a sign of a traffic problem. It is a measurement change that publishers running display ads, especially anyone reporting AdSense numbers to a client, an investor, or a finance team, need to plan around now rather than explain after the fact.
What Actually Changes on February 17, 2027
Right now, AdSense counts an impression the moment an ad begins downloading to a user's device - a method called count-on-download (COD). Under begin-to-render, an impression only counts once the ad creative has actually loaded and started rendering on screen.
Google's own explanation is direct: impressions that started to download but never actually rendered - for example, because a user left the page before the ad appeared - will no longer be counted. The change applies to display web, mobile web, and CTV banner inventory starting on that date.
Why Your Numbers Will Look Lower, Not Your Traffic
Nothing about your site's actual audience changes on February 17, 2027. What changes is which of your ad requests get credited as a real impression. Any request where a visitor bounced, scrolled past too fast, or closed the tab before the ad rendered previously still counted under count-on-download - under begin-to-render, it will not.
For sites with a meaningful share of fast-bouncing traffic (thin content pages, slow-loading templates, aggressive pop-under placements), the drop could be more visible than for sites with strong dwell time and fast-loading ad slots.
Why This Methodology Change Is Actually a Good Thing
Begin-to-render is not a Google-only invention - it is the measurement standard backed by the Interactive Advertising Bureau (IAB) and the Media Rating Council (MRC), the two bodies most advertisers and agencies already trust for viewability standards. Aligning AdSense with that standard means your reported numbers will match what advertisers buying through IAB/MRC-compliant systems already expect, which can actually strengthen trust in your inventory over time even as the raw numbers drop.
What to Do Before February 17, 2027
Google says no technical action is required to enable begin-to-render - it applies automatically. That does not mean there is nothing to do:
- Set the expectation with stakeholders now. If you report AdSense revenue or impressions to a client, a manager, or an investor, flag the coming change in your next report rather than waiting for them to ask why the number dropped.
- Separate the metric from the story. Earnings could dip even with flat or growing traffic - have your explanation ready before the first post-change report goes out.
- Audit pages with fast bounce rates. If a template or placement is losing a large share of impressions to users leaving before render, that is now visible in a way it wasn't before - use it as a prompt to fix slow-loading ad slots, not just to explain the number.
- Check your own analytics against AdSense reporting. After February 17, 2027, expect a gap between raw pageviews and reported ad impressions to widen slightly - build that into how you read the numbers going forward.
For a broader look at how automated measurement changes are reshaping what publishers and advertisers can trust in reported data, see our analysis of Google Ads Terms of Service and AI automation changes from July 2026.
How This Compares to Past AdSense Measurement Changes
AdSense has adjusted measurement methodology before, and the pattern is consistent: Google tends to move toward whatever standard the broader ad industry already treats as the ground truth, then applies it automatically without asking publishers to opt in. Viewability reporting rolled out the same way years earlier - publishers who tracked their own analytics separately from AdSense noticed the shift immediately, while publishers relying only on the AdSense dashboard were caught off guard by a number that moved without an obvious cause.
The lesson carries over directly to February 17, 2027: if your AdSense numbers are the only measurement you look at, you have no independent way to tell a real traffic drop from a measurement-methodology drop when the change lands. Publishers who keep a second, independent view of their traffic (server logs, an analytics platform not tied to AdSense, or even a simple pageview counter) will be able to say with confidence "this is the begin-to-render change, not a traffic problem" on day one, instead of spending days investigating a drop that was never really there.
If You Run Auto Ads
Auto ads place units dynamically based on Google's own algorithm, which means a meaningful share of the ad slots on a page can sit below the fold or far down a long article - exactly the placements most likely to lose impressions under begin-to-render, since a visitor has to actually scroll that far for the ad to render at all. Sites leaning heavily on auto ads for monetization should expect a comparatively larger visible drop in impressions than sites running a small number of manually placed, above-the-fold units. This is not a reason to disable auto ads - it simply means the reported number will finally reflect what was actually seen, which is worth knowing before February 17, 2027 rather than after.
FAQ
Do I need to do anything to enable begin-to-render? No. Google has confirmed no technical action is required from publishers - the change applies automatically to display web, mobile web, and CTV banner inventory on February 17, 2027.
Will my earnings definitely go down? Reported impressions are expected to decrease for most publishers, since impressions that never actually rendered will no longer count. Whether earnings drop depends on your CPM structure and how much of your previous impression count came from non-rendered ad requests.
Is this the same as the count-on-download method other ad networks still use? No. Count-on-download counts an impression as soon as the ad starts downloading. Begin-to-render only counts it once the ad has actually started to render on the user's screen - a stricter, more advertiser-trusted standard.
Why is Google making this change? The shift aligns AdSense with the measurement standard already backed by the IAB and the Media Rating Council (MRC), the industry bodies that set viewability and impression standards most advertisers already expect.
Does this affect all AdSense inventory types? It applies to display web, mobile web, and CTV banner inventory starting February 17, 2027, per Google's current announcement.
Bottom Line
Nothing breaks on February 17, 2027, but your AdSense dashboard will likely show lower impressions and possibly lower earnings even with unchanged traffic. Get ahead of the conversation with anyone who sees those numbers, and use the more accurate measurement as a chance to find and fix pages that are quietly losing impressions to slow ad rendering.
