First-click, last-click, linear, position-based: the rule-based menu, ranked by honesty

First-click, last-click, linear, position-based: the rule-based menu, ranked by honesty

Before reaching for algorithmic attribution, most teams pick a heuristic from their analytics tool's dropdown. Which least distorts reality?

The contenders
Last-click credits the final touch. First-click credits the introducer. Linear splits credit evenly across all touches. Time-decay weights recent touches more. Position-based (U-shaped) hands 40% each to first and last, splitting 20% among the middle.

What each implicitly assumes
Last-click assumes the journey began the moment before purchase — it systematically over-credits branded search and retargeting, which sit near the finish line. First-click assumes the introducer deserves everything and ignores the closer. Linear assumes every touch is equal, which is rarely true but at least symmetric. Position-based bakes in a belief that discovery and closing matter most, the middle least — a defensible prior for considered purchases, an arbitrary one for impulse buys.

The honest framing
None of these are measured — they are assumptions wearing a number. Their only virtue over algorithmic models is transparency: everyone can see the rule. Their fatal flaw is that the rule is invented, not estimated, so it cannot be wrong in a way the data corrects.

Bottom line for practitioners: Rank by distortion, not by sophistication. Last-click is the most biased toward bottom-funnel; first-click the most biased toward top. Position-based is the least-bad default for considered purchases because it refuses to fully ignore either end. But treat all rule-based models as reporting conventions, never as causal claims — graduate to data-driven or experimental methods for budget decisions.