Google is about to make your Display impressions more honest

Starting February 17, 2027, a Display ad that never actually loaded will stop counting as an impression on Google AdSense and Ad Manager. That might sound like a technicality. It's actually a meaningful upgrade to how Display gets measured, and it's worth understanding now so you can see it coming, rather than wondering later why your numbers moved.

What's actually changing

Today, AdSense and Ad Manager count a Display impression the moment an ad starts downloading to a user's device. That's it. The user doesn't have to see it. They don't even need to stick around long enough for it to finish loading. If someone clicks away from a page a split second after it starts to load, that still shows up in your impression count.

From February 17, 2027, Google is switching to what it calls Begin-to-Render (BTR): an impression only counts once the ad has actually loaded and started rendering on screen. Google has already been using this standard for native, app, and video inventory. This change simply brings Display in line with the rest.

In plain terms: Google is closing a gap between "we sent the ad" and "the ad was actually shown." On the AdSense and Ad Manager side, Display is the last format to make that jump. (Google’s buy-side tools, DV360 and Campaign Manager 360, already switched in 2025.)

Why your numbers will move, and why that's a good thing

Once this rolls out, expect total impressions to drop, and that's the point. Google is no longer counting impressions that started downloading but never rendered, because the user bounced, the tab lost focus, or the connection dropped. Those were never real impressions in any meaningful sense; they were just noise in the count.

The upside shows up fastest in CTR: if impressions fall while clicks hold steady (people who click were, by definition, actually shown the ad), CTR should climb. That's not noise; it's your existing performance measured against a denominator that finally reflects reality. The same logic likely applies to eCPMs, since publishers are now being paid only for impressions that genuinely rendered, which should push ad quality and inventory value up over time.

Bottom line: nothing about how your ads perform is changing. What's changing is that the number finally matches what actually happened.

Don't forget about Performance Max

This is easy to file away as "a Display thing" and move on, but Display inventory is also what powers a chunk of Performance Max delivery. If your PMax campaigns lean on Display placements (and most do, to some degree), you'll likely see the same pattern show up there: impressions down, CTR up, and reporting that looks different even though nothing about your targeting or creative changed.

If you're not expecting that shift inside PMax specifically, it's easy to misread it as a sudden change in campaign health.

What this means for optimization on Mixilo

This is a data-quality upgrade, not a delivery change, so it shouldn't require you to rebuild anything. But it will affect any process that leans on historical impression and CTR baselines, which is most optimization logic, automated or not.

A couple of things worth building in ahead of February 17, 2027:

• Treat the date as a hard break in your time series. Any model, alert, or benchmark comparing impressions or CTR before and after that date should account for the methodology change, not just the raw trend line.
• Re-baseline CTR thresholds for Display and PMax. If any tool flags anomalies based on historical CTR ranges, those ranges will look artificially "too good" right after the switch unless they're recalibrated.

None of this is urgent today. It is worth putting on the roadmap for Q1 2027 so nobody's caught off guard by a "performance spike" that's actually just better counting.

The takeaway

Google is tightening its definition of what counts as an impression, and Display is finally catching up to native, app, and video. For advertisers, that means cleaner data: you're paying for and measuring ads that people actually had a chance to see. The one thing to plan for is a one-time reset in historical comparisons around that date: impressions down, CTR up, so nothing built on old baselines misreads better data as a performance swing.

FAQ

When does this change take effect? February 17, 2027.

What counted as an impression before this change? An impression counted the moment an ad started downloading to a user's device, even if the user left the page before it fully loaded or rendered.

What counts as an impression after this change? An impression only counts once the ad has successfully loaded and begun rendering on the user's screen. This is the Begin-to-Render (BTR) methodology. Note that rendering isn’t the same as being viewed: an ad can render below the fold without the user scrolling to it.

Will my impression numbers go up or down? Down. Impressions that started loading but never rendered will no longer be counted.

Does this mean my ads are performing worse? No. This is a measurement change, not a performance change. Fewer impressions are being counted because Google is no longer counting ads that were never actually shown to anyone.

Will my CTR change? Likely yes. It should go up, assuming clicks stay stable while impressions fall. That's expected and not a sign your creative suddenly improved.

Does this affect Performance Max? Yes. Display inventory feeds into PMax delivery, so campaigns with Display placements should expect the same pattern: lower impressions, higher CTR. Google hasn’t detailed how the change will show up in Google Ads reporting, so treat any shift around February 17 as a measurement change before reading it as a performance one.

Is this new for Google overall, or just Display? Just Display  on AdSense and Ad Manager. Google already applies Begin-to-Render to native, app, and video inventory. (DV360 and Campaign Manager 360) moved Display to Begin-to-Render in 2025. This change brings AdSense and Ad Manager Display in line with the rest.

What should I do before February 17, 2027? Flag the date as a break point in any historical reporting, re-baseline CTR and impression benchmarks for Display and PMax, and make sure any automated alerts or optimization rules account for the shift rather than misreading it as a performance change.