A buyer may encounter a business through several interactions before submitting a form or making a purchase. An attribution model applies a set of rules or a data-driven algorithm to assign reporting credit across eligible touchpoints. It helps answer how a report is distributing credit, but it cannot fully observe every conversation, recommendation, offline event, or device change in a customer’s decision. Google Analytics describes attribution models as rules or algorithms that assign credit along a user’s path to an important action. See the current Google Analytics attribution model documentation for the options and scope in GA4.

Last-click gives all eligible credit to the last interaction before the key event, subject to the selected reporting scope and channel rules. It is simple to explain and can help answer which source most recently preceded an action. It can also understate earlier interactions that introduced the customer to the company or helped them compare options. The label “last click” does not mean the final interaction caused the decision. It means the model assigns credit that way. In GA4, the exact behavior depends on the reporting model, dimensions, and whether you are looking at paid-and-organic or Google Ads reporting.

Data-driven attribution distributes credit across eligible interactions using data associated with the account and key event. It is intended to estimate how interactions contribute to outcomes based on available data. The resulting allocation can differ from last-click, especially when customers use multiple channels. Because it depends on tracked interactions and account data, review whether the key events, consent settings, campaign tagging, and conversion paths are reliable enough for the report you are using. A data-driven label does not remove measurement gaps or turn observational data into a controlled experiment.

Use a model that matches the decision. If you need a simple view of the last recorded interaction, last-click may be easier to communicate. If the property has enough data and you want credit distributed across eligible interactions, data-driven attribution may add another perspective. Keep the chosen model and reporting period visible when comparing numbers. Do not treat two model outputs as competing “true” totals. They answer the same reporting question under different rules. If a budget decision matters, compare attribution with other evidence such as lead quality, sales outcomes, costs, and experiments where feasible.

Google notes that changing the reporting attribution model can affect historical and future key-event reports that use event-scoped traffic dimensions. User- and session-scoped traffic dimensions are not changed in the same way. First-click, linear, time-decay, and position-based attribution models are no longer available as GA4 reporting models as of November 2023; do not assume old screenshots or tutorials describe current settings. Record the property, key event, channel grouping, dimension scope, date range, and model used. If you compare reports across a settings change, annotate when the change occurred so readers understand why the view differs.

Use consistent campaign tags, verify that key events fire once at the right moment, and check whether form submissions and CRM outcomes can be reconciled. Keep a clear distinction between website enquiries, qualified leads, sales conversations, and revenue; they are different outcomes. Attribution is one lens on a broader measurement system. Link it with conversion tracking and a documented definition of what each key event means. For implementation or reporting support, see Marketing Attribution Services.