Skip to content
Cover illustration for GA4 multi-touch attribution without BigQuery: what the built-in reports can and can’t do

GA4 multi-touch attribution without BigQuery: what the built-in reports can and can’t do

GA4’s built-in views cover model comparison, conversion paths, and lookback windows. Here is what they answer well, where they stop, and when the raw export becomes necessary.

Jamie IsabelBy Published Updated 8 min read

Founder of MetricMaven. Previously managed Google Tag Manager and Google Analytics across 800+ sites at HCA Healthcare, and built attribution and eCommerce reporting at RYZE Superfoods and Mars.

You can investigate attribution in GA4 without building a warehouse. The limit is how much control you need over the underlying events and model. Use the built-in reports for the questions they answer, and record where their rules differ from your business definitions.

What the attribution settings control

The property’s attribution settings (found under Admin, in the Data display section; menus move, so look for the attribution settings page rather than a fixed click path) control two things: the reporting attribution model and the lookback window for each key event. Those two settings decide what every attribution report in the property is able to show.

Available lookback windows can differ by key-event type and property configuration. Check the configured setting in your property before comparing reports, and note the date of any change because attribution reporting can be affected by the selected window.

The reporting models you still have

GA4’s current attribution reports offer data-driven attribution and last-click variants, including paid and organic channels last click and Google paid channels last click. The available models and report options can change, so check the model selector in the property when reproducing an older analysis. There is no built-in first-touch or linear report to switch to, so any other credit rule has to be built from exported data.

Which reports use the model, and which don’t

GA4 does not apply attribution uniformly. The attribution reports in the Advertising section use the reporting model, while Acquisition reports work differently: First user and Session dimensions are rule-based.

First user credits the first acquisition, and session source follows last non-direct click logic. Neither uses the reporting model. Only some key-event metrics in acquisition reports are attributed with it, which can produce different channel totals across reports.

Three report windows examine one event trail through different lenses, alongside a separate archive for deeper raw-data access.
QuestionGA4 report to useLimitation
Does the reporting model change channel credit?Model comparison in the Advertising attribution reportsCompares the models GA4 offers, which are data-driven and last-click variants. It does not show how credit was assigned to each path.
Which touchpoints came before a key event, and how many?Conversion paths (early, mid, and late touchpoints)Shows touchpoints before a key event, not a full audit of each journey. You cannot change the journey rules.
Which source gets credit for a new user?Acquisition reports with First user dimensionsRule-based, not driven by the reporting model. Totals can differ from attribution reports.
What sequence did users follow across pages or events?Path explorationShows sequences, not attribution credit. Explorations may sample data on large date ranges.
Where do people drop off between steps?Funnel explorationShows step-by-step completion, not channel credit. Sampling can apply on large ranges.
Which CRM deals or offline revenue belong to each path?None in GA4Needs the raw export joined to CRM or order data, which means BigQuery or a warehouse.
Use the built-in view that matches the question, and treat the limitation column as the point where you stop.

A practical recipe

Use the built-in views in this order so each step answers one question and the results are easy to explain later.

  1. Decide the business question first, such as which channel drives qualified demand or what caused a shift in conversions.
  2. Pick one to three key events. A longer list splits the signal and makes model comparisons harder to read.
  3. Set the lookback window to fit your real sales cycle, and write down the date you set it.
  4. Use model comparison to see whether channel credit moves between data-driven and last-click. Large shifts need a check before they become a conclusion.
  5. Use conversion paths to count the touchpoints before each key event, and note which differences you cannot explain yet.
  6. Export or record the differences, then reconcile the key event counts with your CRM or order system.

Google says that if there is not enough data to inform modeled key events, those events may not be reported or may be attributed to Direct. Treat a model comparison with little movement as a reason to check available data and settings, not as proof that two models agree.

What you cannot do without raw data

Some questions sit outside the built-in reports. GA4 does not let you define custom journey rules or join CRM revenue and offline deals to web paths in these views.

Questions about non-Google impressions or the exact credit assigned to an individual conversion also need another data source. Those are the points where the built-in views may stop being enough.

When one of those gaps blocks a decision, move to the raw export. The BigQuery companion guide covers journey stitching and model logic built from the GA4 export, and it is the right next read if you need custom credit. For the model definitions behind these choices, start with multi-touch attribution and models explained.

If the GA4 views disagree with ad platforms, the Google Ads and GA4 conversions guide covers the usual causes. For a team that needs the warehouse build scoped and documented, conversion tracking and attribution is the service that covers it.

If the GA4 views disagree with your CRM and you cannot tell which one to trust, book a tracking audit. Bring the report and your key event list, and we’ll map which built-in view answers each question and where the raw export is needed. For how an attribution review is scoped, see the example 30/60-day attribution audit scope in the tracking audit checklist.

Get the next one before it hits.

Short emails when Meta, Google, GA4 or LinkedIn change something that moves your numbers.

Frequently asked questions

Can you do multi-touch attribution in GA4 without BigQuery?

Partly. GA4 offers a reporting attribution model, model comparison, and conversion paths, which cover channel credit and touchpoint counts. It does not let you define custom credit rules, join CRM revenue, or audit individual journeys without exported data.

Which attribution models are still available in GA4?

GA4’s current model choices include data-driven attribution and last-click variants. Check the current model selector and Google’s help documentation when reproducing an older analysis, because options and labels can change.

Why do GA4 acquisition reports and attribution reports show different channel totals?

Acquisition reports rely on rule-based First user and Session dimensions, which do not use the reporting model. Only some key-event metrics in those reports are attributed with the reporting model, so the totals can differ.

Can MetricMaven help with GA4 attribution?

Yes. We confirm the key events and lookback windows, check which built-in view answers each question, and document where the raw export is needed so the team can explain differences without rebuilding reports by hand.

Sources and references

Updated October 16, 2026

Check what your site sends GA4 before you model it.

The free GA4 checker flags more than one GA4 property, repeated page views and retired Universal Analytics tags on your homepage. About 10 seconds.