When you set up conversion tracking in Google Ads, you pick an attribution model. Most advertisers leave it on the default and never think about it again. That default is quietly misrepresenting where your conversions are coming from — and costing you budget.
Here is what each Google Ads attribution model actually does, how to tell which one is hurting your account, and what to use instead.
What attribution models do in Google Ads
When a customer clicks more than one of your ads before converting, Google Ads has to decide how to distribute credit for that conversion. The attribution model is the rule it uses to make that decision.
The model you choose changes which campaigns, ad groups, and keywords appear to drive conversions — which changes where your budget goes. Getting it wrong means you are systematically underfunding the campaigns that work and overfunding the ones that do not.
The six attribution models Google Ads offers
Last click
All conversion credit goes to the last ad the customer clicked before converting. If a customer clicked a brand keyword ad after already clicking a generic keyword ad earlier in the week, the brand keyword gets 100% of the credit.
This was the default for years. It biases heavily toward brand terms and bottom-of-funnel keywords, which appear at the end of the path by design. Non-brand discovery keywords — the ones that introduced you to the customer — get zero credit for the sales they ultimately drive.
First click
All credit goes to the first ad the customer clicked. The opposite problem from last click: it overcredits discovery and awareness keywords, undervalues the ads that close the deal.
Useful for understanding which keywords introduce new customers. Not useful as your primary model if you are optimizing toward conversions.
Linear
Credit is split equally across all ad clicks in the conversion path. A customer who clicked four ads before converting gives each click 25% of the credit.
Treats every touchpoint as equally important. This is rarely true, but it is less distorted than first or last click and is a reasonable starting point if you have no data to suggest otherwise. More on the linear attribution model here.
Time decay
Ad clicks closer to the conversion receive more credit. A click that happened an hour before the conversion gets more credit than one that happened a week before.
This model assumes recency = relevance. For short sales cycles (same-day purchases) that can be reasonable. For B2B accounts with a two-week or longer buying process, it still overcredits the close and undercredits the research phase. Full breakdown in the time decay attribution guide.
Position-based (U-shaped)
40% of credit goes to the first click, 40% to the last click, and the remaining 20% is split evenly across any middle touchpoints.
This model values both discovery and conversion equally, which matches many B2B buying journeys where both matter. It is more nuanced than first or last click and does not ignore the middle of the funnel the way position-based first-touch or last-touch models do.
Data-driven attribution
Google’s machine learning analyzes your account’s historical conversion data to determine how much credit each touchpoint in your actual conversion paths deserves, compared to similar paths that did not convert.
This is the most accurate model in theory — and Google made it the new default in 2021 for conversion actions with enough data. It requires at least 300 conversions in a 30-day period to activate. Accounts below that threshold fall back to last click.
Full explanation of how the model works and when to use it: data-driven attribution.
Which model should you use?
For most Google Ads accounts, the answer depends on your conversion volume:
- 300+ conversions/month: Data-driven attribution. Let the algorithm use actual path data. It is more accurate than any rule-based model you could choose manually.
- 50–300 conversions/month: Position-based or linear. Both are more balanced than first or last click. Position-based is a reasonable proxy if you believe both discovery and close matter.
- Under 50 conversions/month: Linear. At low volume, data-driven is unavailable and any model is a rough approximation. Linear avoids the extreme distortions of first or last click.
What to avoid: last click for any account running non-brand keywords alongside brand terms. It will systematically undercredit non-brand keywords that drive real demand and cause you to cut the campaigns that are actually doing the work.
The model does not solve the offline problem
Attribution models in Google Ads only assign credit across ad clicks. They do not change what conversions you are tracking. If your conversion action is a form submission and 60% of your actual customers call instead of filling out a form, no model fixes that gap — you are still measuring the wrong thing.
The same problem applies to B2B accounts where a lead to a closed deal takes weeks. Your Google Ads conversion data is reporting lead volume. The attribution model determines which clicks get credit for those leads. But if you are not tracking which leads became revenue, you are optimizing for lead volume, not sales.
Offline conversion tracking is how you fix that — importing actual closed revenue back into Google Ads so Smart Bidding optimizes toward deals, not just form fills. The attribution model then works on top of real revenue data, not just top-of-funnel lead proxies.
How to change your attribution model in Google Ads
Attribution model settings live at the conversion action level, not the campaign level.
- In Google Ads, click Goals in the left navigation, then Conversions and Summary
- Click the name of the conversion action you want to update
- Click Edit settings
- Under Attribution model, select the new model from the dropdown
- Click Save
The change applies going forward. Historical data is not retroactively recalculated. It can take a week or two for your performance metrics to stabilize after the change as the new model populates.
If you are switching from last click to data-driven and your account does not have enough conversion volume, Google will tell you. In that case, position-based or linear are the next best options while you work toward the volume threshold.
Attribution model vs. attribution window
These are two different settings that work together. The attribution model decides how credit is distributed across clicks. The attribution window decides how far back Google looks for ad clicks to include in that calculation.
A 30-day click attribution window means Google will count any ad click from the last 30 days when assigning credit for a conversion. A 7-day window only counts clicks from the last week. For B2B accounts with longer sales cycles, a short window will miss the early touchpoints that influenced the deal — even if your model is set to distribute credit fairly across the path.
Choosing a balanced attribution model and the right window together is what gives you an accurate picture of which ads are actually driving revenue.
If you are on WordPress and want to connect those Google Ads clicks all the way through to actual revenue in your CRM, that is exactly what Sales Provenance is built to do — tracking every touchpoint and tying the closed deal back to the campaigns that drove it.