Since GA4 pruned its attribution menu, the real choice most teams face is down to two: data-driven attribution or last-click. They answer the same question in opposite spirits, and knowing which fits when saves a lot of misplaced budget. Here is how they compare.

What each one does

Last-click attribution gives 100% of the credit for a conversion to the final touchpoint before it. Data-driven attribution instead uses your account's own conversion patterns to distribute credit across multiple touchpoints, based on how much each appears to contribute. One is a simple fixed rule; the other is a model built from your data.

Point by point

The trade-offs line up clearly:

  • Credit logic – last-click: all to the final touch; data-driven: shared across touches by modeled contribution.
  • Transparency – last-click: perfectly simple and explainable; data-driven: a model you cannot fully see inside.
  • Requirements – last-click: works at any volume; data-driven: needs a conversion-volume threshold.
  • Bias – last-click: over-credits bottom-funnel channels; data-driven: fairer to assisting channels.
  • Consistency – last-click: stable and comparable over time; data-driven: can shift as the model updates.

Where last-click still earns its place

Last-click is crude but not useless, and its simplicity is a real virtue. It is perfectly transparent, works at any conversion volume, and gives a stable baseline that is easy to explain and to compare period over period. For low-volume accounts that cannot support a model, or for a sanity-check number everyone understands, last-click does an honest job of saying what closed the deal.

Why data-driven usually wins

For accounts that clear the threshold, data-driven is the better everyday model because it stops starving your upper-funnel channels of credit. Last-click systematically flatters whatever sits at the bottom of the funnel, like brand search, while ignoring the display, social, or content that created the demand. Data-driven spreads credit toward those assisting touches, which leads to less distorted budget decisions inside the platform.

What neither can do

The honest limit is that both are still correlational stories about credit, not proof of cause. Data-driven is a more sophisticated story than last-click, but it is modeled on observed paths, not on what would have happened otherwise. For questions of true incremental value – did this channel cause sales that would not have occurred – you need geo tests or lift studies, which sit above any attribution model rather than inside it.

A note on the 2026 landscape

It is worth being current about what you are actually choosing between. Because GA4 retired the rule-based models in late 2023, data-driven and last-click are effectively the whole in-platform menu now, not two options among six. That makes the decision cleaner but also raises the stakes of the data-driven threshold: an account that cannot support data-driven is not choosing last-click so much as being handed it by default. Knowing that prevents the common error of configuring a model the property cannot actually run.

How to actually decide

If you want a rule rather than a deliberation, use these:

  • Enough conversion volume and you want fairer cross-channel credit: data-driven.
  • Low volume, or you need a transparent number anyone can reproduce: last-click.
  • A budget decision across channels either way: use the model for guidance but confirm with an incrementality test before moving real money.

The pragmatic default

For most teams that clear the volume bar, the sensible default is to run data-driven as the working model while keeping a last-click view handy as a transparent cross-check. Data-driven guides day-to-day optimization with fairer credit, and the last-click number stays useful precisely because everyone understands it and it never shifts under you. Pairing the two, rather than treating them as rivals, gives you a nuanced model for decisions and a simple one for sanity-checking, which is more robust than betting everything on either alone.

With GA4's menu down to essentials, the choice is data-driven versus last-click, and it is mostly settled by volume and purpose. Use last-click for low-volume accounts and simple, stable baselines; prefer data-driven where you qualify, for fairer cross-channel credit; and lean on experiments for the causal questions neither can answer. Choose on those terms and the model serves the decision, instead of the decision bending to whatever model happened to be on.

Want a stronger data analyst role or a raise? Grab the FREE Product Analyst Playbook and get the exact roadmap to your next offer.