Which GA4 attribution model should a lead team use?

Use data-driven attribution when you want GA4 to distribute credit from your property's observed and modeled path data. Use paid and organic last click when you need a deterministic cross-channel baseline. Use Google paid channels last click only when the question is intentionally about the last eligible Google Ads interaction. None of these models proves that an inquiry became a contactable Lead or that a Delivery succeeded.

Practical rule: select the model for the decision, not for the number you prefer. Record the model, eligible channels, dimension scope, lookback window, date range, and Data Freshness with every lead report.

As of August 3, 2026, Google's current attribution overview lists three models in GA4 Attribution reports. Google also says direct visits do not receive credit unless the path consists entirely of direct visits. The table below summarizes Google's documented rules and adds Lucidity's reporting guidance. It does not rank one model as universally accurate.

GA4 model Documented credit rule Useful reporting question Main caution
Data-driven attribution Uses property data and machine learning to estimate the contribution of eligible touchpoints. Credit can be fractional. Which touchpoints contributed to key events across observed paths? Credit can change as GA4 processes and models more data.
Paid and organic last click Assigns all credit to the last eligible non-direct paid or organic channel before the key event. What was the final eligible channel before the key event? Earlier assistive touchpoints receive no credit.
Google paid channels last click Assigns credit to the last eligible Google Ads channel. If no Google Ads click exists, it falls back to paid and organic last click. Which last Google Ads interaction receives credit? It is intentionally narrower than a cross-channel view.

First click, linear, time decay, and position-based models are not current GA4 options. Google says they stopped being available in November 2023. A current implementation guide should not present those legacy models as selectable choices.

How can the same lead journey produce different credit?

Consider this synthetic path: a visitor first clicks a paid social campaign, later clicks a Google Search ad, returns through organic search, and finally visits directly before triggering a generate_lead key event. No personal data is used in this example.

Model Illustrative result What the result means
Data-driven attribution Eligible touchpoints may receive fractional credit. GA4 estimated that more than one interaction contributed. The exact fractions depend on property data and are not asserted here.
Paid and organic last click Organic Search receives all credit. The final direct visit is ignored because an earlier eligible channel exists.
Google paid channels last click Google Paid Search receives all credit. The model looks for the last eligible Google Ads interaction.
Example limit: this is a model illustration, not a reproduced GA4 report or a benchmark. Real data-driven credit depends on each property's data and configuration.

The three results can all follow their stated rules. The disagreement is not evidence that one report is broken. It shows why a report must name the model before anyone compares channel totals. For the event design that precedes this choice, see the guide to tracking qualified Leads in GA4.

Which reports does the attribution model affect?

GA4's reporting attribution model applies to key event reports and explorations that use event-scoped traffic dimensions. Google's attribution settings documentation names Source, Medium, Campaign, and Default channel group as examples. The same page says user-scoped and session-scoped traffic dimensions are unaffected by a reporting attribution model change.

Scope Example Question answered Affected by reporting model change?
User First user source Where was a user first acquired? No, according to Google's current settings documentation.
Session Session source Which source began the session? No, according to Google's current settings documentation.
Event Source with Key events Which source receives model-assigned key event credit? Yes.

Changing the reporting attribution model applies to historical and future data in the affected reports. That makes the model a report configuration, not a permanent property of the original website interaction. Preserve raw campaign and Source evidence separately when you need an audit trail. The Lead Source attribution guide explains this evidence boundary in detail.

How should you choose a key event lookback window?

A lookback window determines how far before a key event a touchpoint can remain eligible for credit. Match the window to the decision cycle you are measuring, then document it. Do not silently change the window to make a campaign look stronger.

Key event type Google-documented default Other documented choices Selection question
Acquisition key events 30 days for first_open and first_visit 7 days How far back should first-acquisition touchpoints remain eligible?
Other key events 90 days 30 or 60 days How long is the realistic path to this key event?

Google says lookback-window changes apply going forward and affect all reports in the property. It also says the selected key event lookback window applies to session attribution. Record the change date so reports from different periods do not appear comparable when their eligibility rules differ.

Why can GA4 attribution totals change later?

Two documented processes matter. Google's attribution overview says data-driven conversions can be reattributed for up to seven days after conversion. Its modeled key events documentation says attributed channel conversion data can be updated for up to 12 days after the conversion is recorded while Analytics processes data and trains its models.

Those windows describe different mechanisms and are not service guarantees. A lead report should identify when GA4 context was retrieved and whether recent dates may still be revised. It should not overwrite first-party Lead or Delivery timestamps to match a later attribution revision.

Recommended Data Freshness label: GA4 attribution context retrieved through [date] and subject to documented revisions. First-party Leads, Validation Outcomes, and Deliveries included through [timestamp]. Report generated in [timezone].

What should you ask before choosing the model?

What decision follows?

Budget allocation may need distributed credit. A simple reconciliation may need a deterministic last-click baseline.

Which channels count?

Decide whether the view must be cross-channel or intentionally limited to Google paid channels.

Which scope is shown?

Name user, session, or event scope. Do not compare unlike dimensions without explaining the difference.

How fresh must it be?

Decide whether stakeholders can wait for modeled revisions or need a clearly labeled provisional report.

What does attribution not prove about a Lead?

Attribution assigns credit to eligible touchpoints around a key event. It does not establish whether a form submission contained contactable information, passed the Business's validation rules, reached the correct Destination, or produced a successful Delivery. Those are first-party operational facts.

Record Question Source of truth Required evidence
Attribution credit Which eligible touchpoints receive credit for the key event? GA4 traffic context Model, scope, channel setting, lookback window, date range, freshness
Lead Was a first-party inquiry actually received? Lucidity or another Lead system Lead identifier, Intake, Business, received time, Source evidence
Validation Outcome Was the Lead contactable and safe to route under stated rules? Lucidity Validation Run Checks, reasons, outcome, provider context, evaluated time
Delivery Was the Lead sent to a Destination, retried, accepted, or failed? Lucidity Delivery record Destination, attempt identity, timestamps, terminal outcome
System boundary: GA4 is traffic and attribution context. Lucidity Leads, Validation Outcomes, Destinations, and Deliveries are first-party operational facts. A model can reassign channel credit without changing what happened to a Lead.

This separation also explains why GA4 key event counts may not match qualified or delivered Lead counts. The marketing attribution software guide expands the model into a full requirements matrix for lead teams.

A practical GA4 attribution configuration checklist

  1. Name the decision. State whether the report supports budget, campaign diagnosis, cross-channel explanation, or a deterministic reconciliation.
  2. Select the model. Use the current documented model whose credit rule matches that decision.
  3. Confirm eligible channels. Record whether paid and organic channels or Google paid channels can receive credit.
  4. Set and record the lookback window. Match it to a stated decision cycle and preserve the configuration date.
  5. Name every dimension's scope. Keep user, session, and event-scoped acquisition questions separate.
  6. Label Data Freshness. Flag recent GA4 context as provisional when documented attribution or modeling revisions can still occur.
  7. Reconcile outside GA4. Join safe campaign evidence to the first-party Lead, Validation Outcome, Destination, and Delivery records without sending personal contact data to analytics.

How does Lucidity fit?

Lucidity keeps the operational record that attribution models cannot provide. It captures first-party website inquiries as Leads, applies explicit Validation Outcomes, routes qualified Leads to configured Destinations, and records each Delivery attempt. GA4 remains useful traffic context and may revise how credit is distributed across channels.

Keep both views side by side. Use GA4 to explain channel credit under a named model. Use Lucidity to prove what happened from Intake through Delivery.

Pair attribution context with first-party Lead outcomes

See how Lucidity preserves Lead, Validation, Destination, and Delivery evidence beside GA4 traffic context for every Business.

Request a Lucidity lead reporting demo