Google Ads: Why the Old Attribution Models Are Disappearing
Google Ads is going through one of the most noticeable changes to performance measurement in years. In July 2026, the platform permanently removes four attribution models while simultaneously expanding AI Max, a new automation layer for Search campaigns. This isn’t a cosmetic interface update: it directly affects how advertisers see each click’s contribution to conversions, and what automated bidding strategies will optimize toward.
What happened
As of mid-July 2026, the “First click,” “Linear,” “Time decay,” and “Position-based” models can no longer be selected for new conversion actions in Google Ads. By September 2026, all four models will disappear entirely: accounts that were still using them for existing conversion actions will be switched automatically to Data-Driven Attribution (DDA). Google confirms this in its help center article on attribution models, explaining the decision by noting that these four models together accounted for less than 3% of all web conversions on the platform, since the vast majority of advertisers had already moved to DDA or “Last click” long ago.
At the same time, Google officially confirms that AI Max, the expanded automation layer for Search campaigns, is exiting beta with stronger targeting and creative controls. Starting in September, legacy tools such as Dynamic Search Ads will begin automatically upgrading to AI Max. Another signal pointing in the same direction: updated terms of service, which Google applied to all accounts on July 1, 2026, explicitly allow automated systems to “form, select, or generate targets, ads, or landing pages” on the advertiser’s behalf.
Why it matters
Attribution isn’t an abstract interface setting; it’s the foundation for how an advertiser and their bidding strategies see the value of each channel and each click. If your account has used, say, a linear model for years to evaluate long B2B deal cycles, the forced switch to DDA will change how value is distributed across touchpoints, and along with it, your reported ROMI figures by campaign may shift, even if the actual number of conversions and spend stay the same.
At the same time, the expansion of AI Max marks another step toward automated system decisions replacing manual control over Search campaign structure. For accounts used to working with precise control over keywords and ads, this calls for a rethink of the optimization approach.
How it works
| Before | Now | What to do |
|---|---|---|
| 5 attribution models to choose from, including first click, linear, time decay, position-based | Only “Data-driven” (DDA, the default) and “Last click” remain | Check which model each conversion action in the account uses before September 2026 |
| Dynamic Search Ads and similar tools worked separately from AI Max | Starting September 2026, legacy tools automatically upgrade to AI Max | Test AI Max on one campaign ahead of time rather than waiting for a forced transition |
| Terms of service didn’t directly mention automated generation of targets and ads | Updated terms (effective July 1) explicitly permit automated systems to do this | Review the automated assets and advertising goals settings in your account |
For most accounts, there’s technically nothing to “break”: the transition happens automatically. The key difference is that reporting on existing campaigns may start showing different conversion-distribution figures across channels and keywords after the switch to DDA, and that’s not a configuration error.
Who this affects
The attribution change touches every single Google Ads advertiser without exception; there’s no opting out, since the models are being removed by force. How significantly your reported figures will change, though, depends on which model your account used before. Advertisers with a long deal cycle (B2B, expensive services) who deliberately chose a linear or time-decay model precisely because last click undervalued early touchpoints will feel this shift the most.
The expansion of AI Max is especially relevant for accounts with medium to large Search budgets, where there’s enough data to train automated strategies. Small accounts with limited conversion volume should test more cautiously and compare results against a manual approach for longer before fully relying on automation.
Worth calling out separately are accounts where attribution historically served not just reporting but part of a commercial arrangement, for example when a specific value-distribution model between paid channels and other traffic sources was agreed with a client. In these cases, the forced switch to DDA is not only a technical task but a communication one: you need to explain in advance to the client or leadership why the reporting methodology is changing, not on the agency’s or specialist’s initiative, but because of a platform decision.
Editorial analysis
The mass forced transition is still rolling out. At the time this article was published, we had not yet tested the consequences on our own client accounts, so we won’t pass off assumptions as verified results. But the DDA mechanism itself isn’t new: the data-driven model has been available in Google Ads since 2016, and the accumulated industry experience with it is substantial enough to forecast the main consequences with high confidence, even without our own test of this specific forced transition. We’ll update this piece with concrete figures once we’ve gathered enough data across client accounts after September 2026.
A practical example
Imagine a services account with a 3-4 week B2B deal cycle that has been using linear attribution for the “Consultation request” conversion action. After the forced switch to DDA, the report may show that branded queries and upper-funnel search campaigns receive less “credit” for the conversion than before, even though actual user behavior hasn’t changed. If the bidding strategy is set to a specific target CPA, this could cause temporary bid fluctuations while the system adapts to the new data distribution.
By contrast, an e-commerce account with a short purchase cycle (a single session, one click to conversion) will notice the change far less: when the conversion path already consists of just one or two touchpoints, the difference between a linear model and DDA is minimal. This is an important filter: before worrying about this update, it’s worth looking at the actual conversion path length in your own account (the “Conversion paths” report under Tools → Attribution) rather than reacting to the news the same way regardless of business type.
What to do right now
- Check under Settings → Conversions which attribution model each conversion action in the account uses.
- Record current reported figures by channel and keyword before September 2026, to have a baseline for comparison after the forced transition.
- Don’t adjust target CPA/ROAS immediately after switching to DDA; give the system 1-2 weeks to stabilize before adjusting bids.
- Test AI Max on one non-critical campaign ahead of time instead of waiting for the automatic upgrade in September.
- Review whether your reporting strategy relies on “before and after” comparisons of different periods without accounting for the attribution model change.
Expert tip. If a campaign runs on target CPA or ROAS, don’t panic over the first fluctuations after switching to DDA; the bidding algorithm needs time to relearn the new conversion value distribution. Sharp manual bid corrections right after the switch usually just prolong the instability period.
Something to watch for. If you report to a client or leadership on a monthly basis, the attribution model change can create a “break” in month-over-month comparisons: the numbers by channel will change not because the advertising got worse, but because the counting methodology changed. It’s worth explaining this to stakeholders in advance, not after the numbers have already surprised someone in a report.
FAQ
Do I need to manually configure anything for the new attribution? No, the switch to data-driven attribution for conversion actions that used legacy models will happen automatically by September 2026.
Can I keep the “Last click” model? Yes, it’s the only model besides DDA that remains available after the other four are removed.
Will actual ad spend change because of this? No, the change affects how conversions are distributed across touchpoints in reporting, not the spend itself or the number of conversions.
Is switching to AI Max mandatory? Currently, AI Max is an expanded option for Search campaigns, not a mandatory replacement. But Dynamic Search Ads and similar tools will start automatically upgrading to AI Max starting September 2026.
Where can I check which attribution model an account currently uses? Under “Settings → Conversions” for each individual conversion action in Google Ads.
Will this affect comparisons between Google Ads and other channels, like Meta Ads? Yes, since the conversion-distribution methodology is changing only in Google Ads, direct ROMI comparisons between channels should be made more cautiously during the transition period.
Conclusion
The attribution changes in Google Ads amount to a forced but technically invisible transition: there’s nothing to manually reconfigure, aside from carefully reviewing your reporting after September 2026. The main risk isn’t the model change itself, but the possibility that fluctuating numbers get misread as declining ad performance. The parallel expansion of AI Max is a signal of the broader direction: Google keeps shifting control from manual settings to automated systems, and it’s worth getting used to working within that logic in advance, rather than trying to hold on to full manual control where the platform is gradually removing it.
If you want to understand how these changes will specifically affect your account and reporting, let’s discuss it in an audit. More on how we run Google Ads campaigns is on the Google Ads page, and in our article on how to reduce CPA in Google Ads.