Attribution & Conversion Counting

Last-Click Attribution vs Position-Based Attribution

In short: Both models decide how much of a conversion each touchpoint gets to claim, but position-based tries to split the difference that last-click refuses to make. Last-click gives the entire conversion to the final click and nothing to anything earlier. Position-based gives 40 percent to the first touch, 40 percent to the last touch, and spreads the remaining 20 percent across whatever happened in between. Last-click is still selectable in Google Ads; position-based was retired from Google Ads and GA4 reporting in 2023. Default to last-click for simple, short paths, and reach for position-based only if you can compute it externally and specifically care about crediting both discovery and close.

By the AdFlint research team · Fact-checked against current Google and Meta platform behavior · Last reviewed July 2026

Last-Click Attribution

Gives all conversion credit to the final ad click before the conversion, ignoring every earlier touchpoint in the path.

All credit lands on the final click before the conversion, which makes reporting simple and reproducible but flattens upper-funnel work. Advertisers use it when paths are short or when they need an auditable baseline everyone can recompute. Google Ads still offers it while defaulting to data-driven attribution. The usual mistake is reading last click as what the ad caused rather than simply where the path ended.

Full definition

Position-Based Attribution

Assigns forty percent each to the first and last touchpoints and splits the remaining twenty percent across everything in between.

Also called U-shaped, this model rewards the two moments most advertisers care about, discovery and close, while acknowledging the middle without letting it dominate. It fits multi-step B2B paths where an early content touch and a final branded search both clearly matter. Google Ads and GA4 retired it in 2023. The weakness is that the weights are arbitrary conventions, not estimates derived from how your customers actually behave.

Full definition

Side by side.

The differences that actually change what happens in your account.

	Last-Click Attribution	Position-Based Attribution
What gets credited	Only the final click, full stop.	First touch and last touch each get a fixed 40 percent, the middle splits the leftover 20 percent.
Sensitivity to path length	None - always one click, however long the path.	The middle share gets thinner as more touches are added, but first and last stay fixed at 40 percent each.
Availability today	Still selectable inside Google Ads.	Removed as a selectable model from Google Ads and GA4 in 2023.
What it assumes matters	Nothing except proximity to the conversion event.	Discovery and closing moments matter most; everything between is secondary.
Where the weights come from	There are no weights to set.	Fixed convention, 40/20/40, not derived from your actual customer behavior.
Data requirement	Just the final click, already in standard conversion tracking.	A full timestamped, ordered path export since neither platform computes it anymore.
Failure mode	A campaign that introduced the customer gets zero credit even in a ten-touch path.	The 40/20/40 split can overstate the middle touches' collective role or understate a dominant single mid-path touch.

What actually separates them.

Last-click is winner-take-all with one recipient; position-based always splits credit across at least two touches, first and last, and often more, so the two models can never agree on how many winning campaigns a single conversion has.

Position-based fixes 40 percent to the first touch regardless of how long ago it happened, which means an old, possibly irrelevant first click can still outweigh several closer, more relevant mid-path touches; last-click would have discarded that first click entirely.

Last-click remains selectable and live inside Google Ads; position-based has to be reconstructed from an external multi-touch export since neither Google Ads nor GA4 computes it anymore.

Adding more mid-path touches shrinks each middle touch's individual share under position-based without changing the first or last touch's 40 percent at all, while last-click is completely unaffected by how many touches exist in between.

A campaign that only ever appears as the very first touch in long paths will show 0 percent under last-click and a consistent 40 percent under position-based, making position-based the better lens for proving upper-funnel value specifically at the discovery moment, not just generally.

Which one should you use?

Use Last-Click Attribution when

Your paths are typically short - one or two touches - so a first, middle, last split does not add much information.
You need a number pulled directly from Google Ads without exporting or recalculating anything.
You are feeding a bid strategy built around the platform's own conversion counting.
You want the simplest possible baseline that every stakeholder can independently verify.

Use Position-Based Attribution when

You run multi-step B2B or considered-purchase paths where both the discovery touch and the closing touch clearly matter.
You want to acknowledge middle-funnel touches without letting them dominate the discovery and close moments.
You have the ordered, timestamped path data to compute it externally, since it is no longer live in Google Ads or GA4.
You are specifically trying to justify keeping an awareness campaign that never appears as the last click but reliably starts the journey.

Common questions.

Why is it called U-shaped attribution?

Plotted as a bar chart across a path, the credit forms a U: high at the first touch, low across the middle, high again at the last touch. That is just a visual description of the fixed 40/20/40 split, not a different model.

Can I change the 40/40/20 weights in position-based attribution?

The classic position-based model uses those fixed percentages as a convention; it is not derived from your data and most tools that still offer it do not let you tune the split. If you need custom weighting, that is really a different model entirely, closer to a custom or algorithmic attribution setup than standard position-based.

Is position-based attribution still available in Google Ads?

No, Google Ads and GA4 removed it as a selectable model in 2023 along with the other rule-based comparison models, leaving data-driven attribution as the default and last-click as the simple alternative. You would need an external multi-touch data export to reconstruct a position-based view today.

How is position-based different from time-decay for a long sales cycle?

Time-decay weights every touch by how recently it happened relative to the conversion, so the shape of the credit curve depends on your actual timing gaps; position-based ignores timing entirely and just fixes 40 percent to whichever touch happened first and 40 percent to whichever happened last, regardless of the dates involved.

Or stop choosing between them.

AdFlint picks the setting, writes the ads, and keeps optimizing inside the Google and Meta accounts you already own.

Try AdFlint free See how it works

Related comparisons

All comparisons