Marketing Attribution Explained: Models, Settings and Limits

How GA4 assigns credit for conversions with last click and data-driven models, which settings change your numbers, and what attribution can't tell you.

CTR Guide · Last reviewed October 22, 2026 · 6 min read

Free tool:UTM link builder →

Imagine a hypothetical customer of a small yoga studio. She sees a social post on Monday, searches for the studio by name on Wednesday, clicks a search ad on Friday, and books a class the next week after opening a newsletter email. Which of those deserves credit for the booking?

That question is what marketing attribution tries to answer. The answer you get depends on the model you choose, the settings in your analytics tools and, just as important, what those tools can't see.

This guide explains the attribution models available in Google Analytics 4 (GA4), what happened to the older multi-touch models, which settings change your numbers, and the limits that apply to every model. For the basics of events and reports, see our guide GA4 Basics for Marketers: Events, Key Events and Reports.

What Attribution Means

Google's Analytics Help defines attribution as "the act of assigning credit for important user actions to different ads, clicks, and factors along the user's path to completing the action." In GA4, those important actions are called key events, such as a purchase, a lead form or a booking.

An attribution model is, in Google's words, "a rule, a set of rules, or a data-driven algorithm" that decides how that credit is split. GA4's attribution reports offer three models:

Data-driven attribution
Paid and organic last click
Google paid channels last click

Google notes that all of these models exclude direct visits from receiving credit, unless the path to the key event consists entirely of direct visits.

Last Click: Simple and Lopsided

Paid and organic last click ignores direct traffic and gives 100% of the credit to the last channel the customer clicked through before converting (or, for YouTube, an engaged view). Google says it is the same model sometimes called last non-direct click. In Google's own examples, a path of Display, Social, Paid Search, then Email gives all credit to Email; a path ending in a direct visit after Paid Search gives all credit to Paid Search.

Google paid channels last click gives 100% of the credit to the last Google Ads channel the customer clicked. If there's no Google Ads click in the path, it falls back to paid and organic last click.

Last click is easy to explain and easy to check, which is why it remains popular. Its weakness is obvious from the yoga studio example: the social post that introduced the studio gets nothing. Last-click models tend to favor channels that catch people at the end of a decision, such as branded search and email, and to undervalue channels that start the journey.

What Happened to Multi-Touch Rules

Multi-touch models try to share credit across several touchpoints. For years, Google Ads and Google Analytics offered rules-based versions with names that describe their logic: first click, linear, time decay and position-based. Google describes rules-based models as assigning value to each touchpoint "based on predefined rules."

Google announced in April 2023 that these four models would be removed from Google Ads and GA4, saying they don't provide the flexibility needed to adapt to evolving consumer journeys. GA4's help pages state they have not been available since November 2023. Data-driven attribution, last click and external attribution were not affected.

If you read older articles recommending a linear or position-based model in Google's tools, that advice is now out of date for those products, although the ideas behind them still show up in other tools.

Data-Driven Attribution

Data-driven attribution, which Google made the default in Google Ads and GA4, is multi-touch attribution built from your own data rather than fixed rules. According to GA4's documentation, it:

Uses machine learning to evaluate both converting and non-converting paths.
Considers factors such as time from the key event, device type, number of ad interactions, the order of ad exposure and the type of creative assets.
Takes a counterfactual approach, comparing what happened with what could have happened to estimate which touchpoints most likely drove key events.
Builds a separate model for each advertiser and each key event.

Google's illustration: a path of four ad interactions leads to a 3% probability of a key event, and without the fourth interaction the probability drops to 2%, so that interaction gets credit for a 50% increase in probability. Credit is assigned this way across the path.

Two practical effects follow. You may see fractional credit, such as a channel credited with part of a conversion, in reports that use event-scoped dimensions. And Google notes that conversions can be reattributed for up to 7 days after they occur, so recent numbers can shift.

Settings That Change Your Numbers

In GA4, attribution settings are under Admin, then Events, then Attribution settings, and you need the Marketer role or above to change them. Three settings matter most:

Reporting attribution model. Changing it applies to both historical and future data and affects reports using event-scoped dimensions such as Source, Medium, Campaign and Default channel group. User- and session-scoped dimensions, such as Session source or First user medium, aren't affected.
Key event lookback window. This sets how far back a touchpoint can earn credit. For acquisition key events (first_open and first_visit), 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. Changes apply going forward only.
Channels that can receive credit. GA4 reports default to paid and organic channels; Google Ads reports default to Google paid channels.

These differences explain many "why don't my numbers match?" questions. Google Ads and GA4 can credit different channels by default, and if the two accounts use different time zones, Google notes you may see discrepancies.

The Limits of Any Attribution Model

Every model, including data-driven attribution, works only with what it can measure. Keep these limits in mind:

Untracked touchpoints get no credit. A friend's recommendation, a yard sign, a radio spot or a phone conversation usually leaves no click to attribute.
Direct visits are largely excluded from credit, as noted above.
Credit is not the same as a controlled test. Even data-driven attribution, with its counterfactual approach, divides credit for conversions that happened among the touchpoints it observed. It can't directly show how many of those customers would have bought anyway if you had switched a channel off.
Different tools give different answers. Each platform applies its own model, settings and tracking, so their totals rarely line up.
Small numbers are fragile. With only a handful of conversions a month, any split of credit rests on thin evidence.

Going Beyond Attribution Reports

Small businesses can get a fuller picture by combining attribution reports with a few simple practices:

Tag your campaign links consistently, as described in our guide UTM Parameters and Campaign Tracking: A Practical Guide, so traffic sources are recorded accurately.
Ask new customers "How did you hear about us?" and record the answers. Self-reported answers capture word of mouth and offline channels that analytics misses.
Run simple holdout tests. Pause a channel for a defined period or in one area, watch total leads and sales rather than attributed ones, and account for seasonality before drawing conclusions.
Know that marketing mix modeling exists. Google's open-source Meridian project, for example, is designed to estimate historical return on investment by channel, show how impact changes with spend, and suggest budget allocations, using historical sales, media and control data. It is a statistical modeling project rather than a report you switch on, so for most small businesses it's something to understand rather than start with.

Key takeaways

Attribution assigns credit for key events to touchpoints; the model decides how that credit is split.
GA4 offers data-driven, paid and organic last click, and Google paid channels last click; first click, linear, time decay and position-based were removed in 2023.
Data-driven attribution uses your own path data, gives fractional credit and can reattribute conversions for up to 7 days.
The reporting model, lookback window and eligible channels all change your numbers, and Google Ads and GA4 defaults differ.
No model sees offline touchpoints or proves cause; pair reports with UTM tagging, customer answers and simple tests.

Sources

Spotted something out of date? Let us know.