What Is Attribution Modeling? Models, Examples and How to Choose

After its November sales campaigns, an e-commerce company’s marketing team reviews the channel report and sees that most sales came from branded search ads and direct traffic. Video and social ads account for only a small share, so cutting their budgets looks like a sensible decision at first glance. Yet the split shown in the report reflects the attribution model in use far more than the channels’ actual contribution. Change the model and the same sales are credited to different channels and the basis for the budget decision changes with them.

What Is an Attribution Model?

An attribution model is the rule, set of rules or data-driven algorithm that determines how the value of a conversion is divided among the touchpoints a user interacted with before converting. The model itself is called an attribution model, while the process of building and comparing these models is called attribution modeling.

The concept rests on three core terms:

• Attribution: the process of assigning conversion value to touchpoints.
• Touchpoint: any single interaction between a user and a brand, such as an ad click, an organic search visit or an email newsletter.
• Conversion path: the ordered sequence of touchpoints a user passes through before converting.

How Does the Choice of Model Change Reported Performance?

An attribution model does not change the total number of conversions measured; it determines which channel each conversion is credited to. Because users rarely decide to buy on a single visit, and usually interact with a brand several times across different channels and devices, this split directly affects three areas:

• Budget allocation: channel-level investment decisions rely on the conversion and revenue data the model produces.
• Performance metrics: the return on ad spend (ROAS) and cost per conversion of the same campaign differ depending on the model.
• Upper-funnel investment: channels that build awareness can look underperforming in models that only reward the final interaction.

What Are the Types of Attribution Models?

Attribution models fall into three groups, based on how they distribute conversion value.

Single-Touch Models

Single-touch models assign all of the conversion value to one touchpoint. The last click model credits the final interaction before the conversion, while the first click model credits the very first interaction with the brand. These models are easy to interpret, but the rest of the conversion path is left out of the assessment.

Multi-Touch Rule-Based Models

Multi-touch rule-based models divide conversion value across several touchpoints according to a predefined rule. The linear model distributes value equally across all touchpoints. The time decay model gives more weight to interactions closer to the conversion. The position-based model concentrates weight on the first and last interactions and splits the remainder among the touchpoints in between.

The Data-Driven Attribution Model

Data-driven attribution is a model that learns from the account’s own data rather than following a fixed rule. According to Google Analytics Help, the model uses machine learning to evaluate both converting and non-converting paths; it calculates how much the presence of a touchpoint on a path increases the probability of conversion and distributes value according to that contribution. Factors such as time to conversion, device type, number of ad interactions and the order of ad exposures are taken into account.

The difference between models is easiest to see on a single conversion path. A user planning to buy a dishwasher first clicks a home appliance brand’s YouTube ad. A few days later, they search for “quietest dishwasher” and reach the brand’s blog post through the organic results, then click a remarketing ad, and finally search for the brand name and complete the order through a search ad. Here is how the same order is shared out under each model:

Model	YouTube Ad	Organic Search	Remarketing	Branded Search Ad
Last click	0%	0%	0%	100%
First click	100%	0%	0%	0%
Linear	25%	25%	25%	25%
Time decay	Lowest	Low	High	Highest
Position-based	40%	10%	10%	40%
Data-driven	Data-based	Data-based	Data-based	Data-based

 

Under last click, the entire order goes to the branded search ad; under first click, to the YouTube ad. The organic visit and the remarketing ad only receive credit in multi-touch models. The position-based row uses the common 40-20-40 split; in time decay and data-driven models the percentages are not fixed.

Which Attribution Models Are Available in GA4?

Google Analytics 4 (GA4) offers three models in its attribution reports:

• Data-driven attribution: distributes value across touchpoints based on the account’s own data.
• Paid and organic last click: ignores direct traffic and gives all the value to the last channel clicked before the conversion.
• Google paid channels last click: gives all the value to the last Google Ads click on the path; if there is no Google Ads click on the path, it falls back to paid and organic last click.

As stated in Google Analytics Help, the first click, linear, time decay and position-based models were retired in November 2023. These models remain useful as concepts, but in GA4 the practical choice today is between the data-driven model and last click.

Three details matter for reading the reports correctly:

• Direct traffic: all models exclude direct visits unless the conversion path consists entirely of direct visits.
• Lookback window: determines how long before a conversion a touchpoint can occur and still receive credit. According to the attribution settings documentation, the default is 30 days for acquisition events such as first visit and 90 days for all other key events.
• Scope of the setting: the reporting attribution model is changed under Admin > Data display > Events > Attribution settings. The change also applies to historical data, while session-scoped and user-scoped traffic dimensions are not affected by it.

How Do You Choose the Right Model?

There is no single attribution model that suits every business; the right one depends on the business model, channel mix and reporting needs. Four criteria stand out:

• Length of the sales cycle: the longer the decision process, the more touchpoints there are and the more single-touch models narrow the picture.
• Channel diversity: businesses investing in many channels need a multi-touch view to see how those channels influence one another.
• Data volume: the data-driven model learns from the account’s own path data. In accounts with low conversion volume, reading its results alongside last click is the cautious approach.
• Need for transparency: the logic of rule-based models is easy to explain to stakeholders; with the data-driven model, the calculation method is less visible.

Rather than committing to a single model, you can compare the results of two models side by side in the model comparison report in the Advertising section of GA4. Channels whose share rises noticeably when the model changes are the ones being undervalued in last click reports.

What Are the Limits of Attribution Data?

Attribution models only work on measurable interactions; a touchpoint that cannot be measured receives no credit in any model. The main limits are:

• Interactions from users who decline cookie consent may fall outside measurement.
• Movement between devices and browsers cannot always be matched to a single user.
• GA4 models are based on clicks and engaged views on YouTube; ads that are only viewed, and offline touchpoints such as TV or physical stores, receive no credit.

For this reason, attribution data should be read alongside methods such as incrementality testing, which measures a channel’s impact under controlled conditions, and marketing mix modeling.

Why Does Attribution Modeling Matter for SEO and Marketing Performance?

Attribution modeling is the measurement framework that makes the roles of SEO and paid channels on the same conversion path visible together. Organic search is a channel users turn to frequently during the research stage; a visitor who arrives through an informational query often does not buy on that first visit. A report that rewards only the final interaction can therefore understate the contribution of content and SEO investment. In the example above, the blog post receives no credit under last click yet that visit is the moment the user started considering the brand.