Google Just Killed Four Attribution Models: The Last-Click Attribution Problem This Creates by September
Google confirmed in a mid-July 2026 update that first-click, linear, time-decay, and position-based attribution are no longer selectable in Google Ads.
Google’s own announcement frames it as a follow-up to an earlier notice, and by September every conversion action still running on one of those four models gets force-migrated to data-driven attribution, whether the account owner asked for it or not.
What follows is a Google Ads last-click attribution problem that has been building since 2023 and is about to land on advertisers who never got a vote in the decision.
If your reports look different this fall and nothing changed in your account, this is why.
What Google Actually Changed, and When
Two dates matter here. Starting mid-July 2026, first-click, linear, time-decay, and position-based models stopped being available for any new conversion action. Trade coverage of the rollout confirms the second date: by September 2026, the four models disappear entirely, including for conversion actions that were still using them.
Anything not already switched gets migrated to data-driven attribution automatically.
Once that migration finishes, exactly two attribution models remain in Google Ads: data-driven attribution and last-click. That is not new information on its own.
Google first floated removing these four models back in 2023, citing adoption under 3%. What is new is that 2026 closes the loophole. Accounts that quietly held onto one of the four models for three years are the ones about to feel this.
The Real Problem Nobody Is Naming: Data-Driven Attribution Needs Volume You Might Not Have
Data-driven attribution sounds like the obvious upgrade. It is not automatically the right fit for every account. Google generally wants a conversion action to clear roughly 300 conversions in the trailing 30 days before the model has enough data to produce dependable output, and that threshold makes data-driven attribution effectively inaccessible for smaller accounts and niche advertisers with lower monthly conversion counts.
Below that line, the model does not just work worse, it can quietly mislead. As one analysis of low-volume accounts puts it, the output can look sophisticated but actually be misleading when there is not enough data behind it.
That is the actual Google Ads last-click attribution problem worth paying attention to.
Smaller e-commerce brands and leaner agency accounts, the exact profile most likely to be running lower monthly conversion volume, are also the accounts least likely to clear the data-driven threshold. In practice, that pushes their reporting toward last-click-shaped results by default, which concentrates credit on the bottom of the funnel and quietly erases the contribution of prospecting, top-of-funnel video, and early-journey touch points from the numbers leadership sees.
This Fits a Pattern You Have Already Seen This Year
Google is not the only platform narrowing how credit gets assigned in 2026. In January, Meta permanently removed the 7-day and 28-day view windows from its Ads Insights API, and industry analysis put the resulting conversion drop at 15 to 30 percent for accounts that leaned on those longer windows.
In March, Meta narrowed what counts as a click-through conversion so that likes, shares, and saves no longer trigger the 7-day click attribution window, only link clicks do.
None of these changes are dishonest. Each one comes with a public announcement and a stated rationale. But stack Google’s September migration on top of Meta’s January and March changes, and the shape of the year becomes clear: the platforms are simplifying measurement on their own schedule, reported numbers move without any change in real performance, and the advertiser has no visibility into the model doing the deciding.
It is a version of the same dynamic AdBeacon has written about before: every layer of platform-reported credit sits inside a system the advertiser cannot independently audit.
What to Actually Do Before September
This is not a five-alarm fire, but it is worth an afternoon before the migration completes.
Start in the Attribution tab under Tools in Google Ads and check which model every conversion action is currently running.
Pull the trailing 30-day conversion volume for each one and compare it against that rough 300-conversion bar.
That tells you which conversion actions are genuinely ready for data-driven attribution and which ones are about to get migrated into a model they do not have the data to support well.
When the September numbers come in, do not compare them straight against August. Annotate the change for your team or your clients the same way you would any platform update: note the date, note that the model changed, and give performance a few weeks to settle before drawing conclusions.
This is also a good moment to build, or lean harder on, an attribution number that does not depend on which model a platform prefers this quarter, whether that is a blended marketing efficiency ratio or independent, click-based first-party measurement sitting outside the platform entirely.
An attribution model change inside Google Ads should not be able to move your reported performance overnight if you already have an independent number sitting next to it.
If you want to see what your Google and Meta spend actually produced, measured the same way regardless of what either platform changes next quarter, book a live AdBeacon demo and we will show you the gap on your own account.
——-
FAQ
What did Google actually remove from Google Ads in 2026?
Google removed first-click, linear, time-decay, and position-based attribution as selectable models. Only data-driven attribution and last-click remain.
When do these changes take full effect?
Selectability for new conversion actions ended in mid-July 2026. By September 2026, any conversion action still using one of the four deprecated models is automatically migrated to data-driven attribution.
What happens to conversion actions still running a deprecated model?
They are migrated to data-driven attribution automatically. There is no opt-out and no action required to trigger the migration, it happens on Google’s timeline.
Do I need a minimum number of conversions for data-driven attribution to work well?
Roughly, yes. Google generally wants about 300 conversions in the trailing 30 days for a conversion action before data-driven attribution has enough data to be dependable. Accounts below that threshold may not see reliable results.
What should I do before September?
Check which model each conversion action is currently using, compare trailing 30-day conversion volume against the roughly 300-conversion threshold, and build an independent, platform-agnostic performance number so a model change alone cannot swing your reported results.
Sources
- Google Ads Help: New features & announcements
- Boostify: Google Ads Attribution Changes in July 2026
- Kreativa Group: Paid Search Attribution Models, A Complete 2026 Guide
- Attributy: Data-Driven Attribution Guide
- Databox: The Ad Attribution Problem
- Digital Applied: Meta Ads Click-Through Attribution Change 2026