Attribution Model: What It Is and How GA4 Uses It

An attribution model is the method an analytics or ad platform uses to decide which marketing touchpoints get credit for a conversion. When someone finds you through a blog post, comes back from an email and finally buys after clicking an ad, the attribution model decides how much each of those steps counts.

For a business, the model shapes the story your reports tell. Switch models and the same sales can look like they came from different channels, which changes where budget goes next. Knowing how your model works, and what it can't see, keeps you from cutting a channel that's quietly doing its job.

What is an attribution model?

Google's Get started with attribution page describes attribution as assigning credit for key events, like purchases, to the ads, clicks and other touchpoints a user interacts with on the way. A model is the specific rule set or algorithm that does the assigning.

Models fall into two broad groups. Rules-based models follow a fixed formula, like giving all the credit to the last click. Data-driven models use your account's data to estimate how much each touchpoint actually contributed.

Why attribution models matter

They drive budget decisions. Channels that get more credit tend to get more money. A model that undervalues early touchpoints can starve the channels that bring new people in.

They change how SEO looks. Organic search often introduces people to a brand long before they convert. Under a last-click view, that early role disappears.

They shape ad bidding. In Google Ads, the attribution model feeds automated bidding, so the model you choose affects which clicks the system pursues.

They miss what they can't track. Recommendations from AI assistants, word of mouth and podcasts often influence buyers without a trackable click. Any model only splits credit among the touchpoints it can see.

How attribution models work

Classic rules-based models

Marketers have long used a handful of standard models. Last click gives all credit to the final touchpoint. First click gives it all to the first. Linear splits it evenly. Time decay gives more to touchpoints closer to the conversion. Position-based gives most credit to the first and last touchpoints and spreads the rest across the middle.

What changed in Google's tools

Google retired four of those models. Its Google Ads help page says first click, linear, time decay and position-based models were removed from both Google Ads and GA4, while data-driven, last click and external attribution weren't affected. Google's Analytics help says those four models are no longer available in GA4 as of November 2023. Search Engine Land reported that a Google spokesperson cited low adoption, with fewer than 3% of Google Ads conversions using those models.

The models GA4 offers today

In Google Analytics 4, three models remain for reporting:

Data-driven attribution. Google says it uses machine learning to evaluate both converting and non-converting paths, then assigns credit based on how likely each touchpoint is to drive a key event. Credit can show up as fractions, such as 0.6 and 0.4.
Paid and organic last click. All credit goes to the last channel the person clicked before converting, ignoring direct traffic.
Google paid channels last click. All credit goes to the last Google Ads channel clicked. If there was no Google Ads click, it falls back to paid and organic last click.

Lookback windows

Google's attribution settings documentation explains that the lookback window limits how far back GA4 looks for touchpoints. For acquisition key events like first visits, the default is 30 days, with 7 days as an option. For all other key events, the default is 90 days, with 30 or 60 days as options. Changing the reporting attribution model applies to historical and future data, while changes to the lookback window only apply going forward.

How to choose and use an attribution model

Define your key events first. Attribution is only as good as the key events it credits. Track the actions that represent real business value.
Tag your campaigns. Consistent UTM parameters on email, social and partner links help GA4 assign traffic to the right channel.
Compare models. GA4's attribution models report lets you compare key event metrics for different models side by side. Channels that gain credit under data-driven attribution often play an earlier role, earning assisted conversions.
Look at paths. The key event attribution paths report shows which channels start, assist and close conversions.
Keep your sales cycle in mind. If buyers take months to decide, many early touchpoints may fall outside the lookback window.
Add what tracking can't see. Ask new customers how they heard about you. Answers like "ChatGPT recommended you" won't show up in any model.

Common mistakes and misconceptions

"Attribution shows what caused the sale." Attribution splits credit among recorded touchpoints. It shows correlation along a path, not proof that a channel caused the purchase.

"Last click is the truth." Last click is simple, but it ignores everything that happened earlier. Content and organic search that introduce buyers usually look weaker than they are.

"Data-driven is a black box I can ignore." It's built from your own data, so tracking gaps and low volume affect it. Check its results against the paths report and common sense.

"Direct traffic is a channel." Direct often means the source couldn't be identified, which can include people who heard about you from an AI assistant and typed in your URL. GA4's last click models ignore direct traffic when another channel is available.

"GA4 still has linear and first click." Those models were retired. Older guides that mention them describe options that no longer exist in GA4.

How Butter approaches attribution models

At Butter, we start with measurement before optimization: agree on which key events matter, confirm they're tracked accurately and segment by channel so search, AI assistants and ads can be compared fairly. Rankings and AI citations are leading signals, but we want to see whether they turn into engaged visits and key events.

On our GEO+SEO plan, the Butter Analytics dashboard pulls in Google Analytics and Search Console, among 85+ integrations, so you can see organic search, AI referrals and other channels side by side. We don't replace your analytics setup, but we help you read it in the context of your SEO and GEO work. Learn more on our analytics page.

Frequently asked questions

What attribution models does GA4 offer?

GA4 offers data-driven attribution, paid and organic last click, and Google paid channels last click. First click, linear, time decay and position-based models are no longer available as of November 2023.

What is data-driven attribution?

It is a model that uses machine learning to compare converting and non-converting paths, then assigns credit to each touchpoint based on how likely it was to drive the key event. Credit can be split into fractions across several touchpoints.

Which attribution model is best?

There is no single best model. Data-driven attribution suits accounts with enough data, while last click is simpler to explain. Compare models in GA4 to see how much credit shifts between channels before making budget decisions.

Does attribution capture AI search?

Only partly. Visits that click through from an AI assistant can be tracked as referrals, but recommendations that lead someone to search your brand later or type your URL directly are hard to attribute to the AI tool.

Sources and further reading

KEEP LEARNING

100 SEO and GEO terms explained in plain English.

Browse the full glossary →