Google Ads now has two attribution models, not six. Last click gives all the credit to the final ad click. Data-driven splits the credit based on your own past conversion data. Google has retired first click, linear, time decay and position-based, and conversion actions that used them were moved to data-driven automatically.
Most articles about Google Ads attribution still list six models with neat diagrams. Those articles are out of date, and following them will have you looking for settings that are no longer there.
The change is bigger than a menu tidy up. It removes the option to impose your own theory about how credit should be shared, and replaces it with one that is computed from your account.
What attribution models does Google Ads still support?
Google’s documentation is explicit. It states that “the first click, linear, time decay, and position-based attribution models are no longer supported by Google.”
What remains:
| Model | What Google says it does | When to use it |
|---|---|---|
| Data-driven | “Distributes credit for the conversion based on your past data for this conversion action” | The default choice for almost every account |
| Last click | “Gives all credit for the conversion to the last-clicked ad and corresponding keyword” | When you need a simple, predictable number that matches other systems |
| First click | Removed | No longer available |
| Linear | Removed | No longer available |
| Time decay | Removed | No longer available |
| Position-based | Removed | No longer available |
What this means for you: if you inherited an account and someone tells you it is on time decay, check. It is not. It was upgraded to data-driven, and whoever told you that has not looked recently.
What is data-driven attribution actually doing?
It compares the paths of people who converted against the paths of people who did not, and works out which touchpoints made a difference. Credit ends up distributed across the clicks that mattered, in fractions rather than whole numbers.
That is why conversion counts in a data-driven account show decimals. A campaign with 12.4 conversions is not a rounding error. It received partial credit for several conversions where it was one step in a longer path.
The practical consequence is that assisting campaigns stop looking worthless. Under last click, a campaign that introduces people to your brand and never closes them shows zero conversions and gets cut. Under data-driven it shows the share it earned.
The honest limitation: it is a model, not a measurement. It is Google’s estimate based on Google’s view of the journey, which excludes everything that happened outside Google.
Should you use data-driven or last click?
Use data-driven unless you have a specific reason not to. It is the default, it is what Google’s own bidding uses, and it reflects multi step journeys that last click cannot see.
There are two reasonable reasons to choose last click.
The first is reconciliation. If your finance team compares Google Ads conversions to orders in a system that uses last click, fractional credit will cause arguments every month. Sometimes the simpler number is worth more than the more accurate one.
The second is very low volume. Data-driven needs enough conversion paths to model anything. If your account produces a handful of conversions a month, the model has little to work from and last click is at least honest about what it is.
For most accounts, neither applies. Use data-driven.
Does attribution affect bidding?
Yes, and this is the part that actually costs or saves money. Smart Bidding optimizes toward the conversions it is credited with, so the attribution model changes which clicks Google decides are worth buying.
Under last click, upper funnel keywords look poor, so bids on them fall, so they deliver less, so they look poorer. Under data-driven, those same keywords receive partial credit and the bidding treats them accordingly.
Switching models will move your numbers. Expect campaign level conversion counts to shift around even though total conversions stay roughly the same. Give it thirty days and judge the account total, not individual campaigns, in the first few weeks. Our post on bidding strategies covers how the model and the strategy interact.
How does this relate to GA4 attribution?
They are separate systems and they will not match. Google Ads attributes conversions to the ad click. GA4 attributes to sessions and includes channels Google Ads cannot see, like organic search and direct traffic.
Three differences cause most of the confusion:
- Date assignment. Google Ads reports a conversion on the date of the click. GA4 reports it on the date of the conversion. For a long sales cycle, that alone puts the numbers in different months.
- Scope. GA4 can give credit to non paid channels. Google Ads cannot.
- Counting. The two platforms have different rules about repeat conversions.
Pick one system as the source of truth for each decision. Use Google Ads numbers to manage Google Ads, because that is what the bidding uses. Use GA4 to understand the whole picture across channels. Trying to force them to match is a project with no end.
What attribution cannot fix
Attribution decides how to share credit for conversions you recorded. It cannot help with conversions you never recorded at all.
If your sales happen on the phone or close weeks later in a CRM, no attribution model will surface them. That is a job for offline conversion tracking and enhanced conversions.
Equally, if your conversion action counts the wrong thing, attribution will carefully allocate credit for the wrong thing. Get conversion tracking right first, following Google’s setup documentation. Attribution is the last layer, not the first.
Common questions
Why does my conversion count have decimals?
Because data-driven attribution assigns fractional credit. A conversion with three contributing clicks might give 0.5 to one campaign and 0.25 to two others. The totals still add up.
Can I still choose first click attribution anywhere?
Not in Google Ads. Google removed it. GA4 has its own attribution settings which differ, but within Google Ads the choice is data-driven or last click.
Will switching models change my total conversions?
The account total stays roughly the same. What changes is how the credit is spread across campaigns, keywords and ad groups. Watch the total to check nothing broke, then look at the distribution.
Does attribution apply to view through conversions?
View through conversions are counted separately from click based attribution and are reported in their own column. Do not add them to your click conversions when calculating cost per acquisition unless you have a deliberate reason to.
Does Performance Max use data-driven attribution?
Performance Max uses whatever model is set on the conversion action, same as other campaign types. Because Performance Max spans several channels, data-driven usually describes what happened more fairly than last click does.
How long does a data-driven model take to build?
It needs a period of conversion data before it can produce a stable model, and the more conversion paths it has the better it gets. New accounts effectively behave like last click until there is enough history.
Where to go from here
Open your conversion actions and check which model each one is set to. Accounts that have been running for years often have a mix, because different actions were created at different times, and that inconsistency makes reporting harder than it needs to be.
Setting them deliberately takes ten minutes. If you want a second opinion on what the numbers are telling you, our Google Ads management page explains how we handle measurement, and you can get in touch for a look at your account. Our post on what ROAS is also covers how attribution changes the return figure you report.
