Multi-Touch Attribution Models

Last-click attribution assigns 100% of conversion credit to the final marketing touchpoint before purchase, completely ignoring every earlier interaction that influenced the customer's decision.

This creates a systematic bias toward bottom-funnel channels (branded search, retargeting) and against top-funnel channels (display, social awareness, content) — causing brands to underinvest in awareness and over-invest in conversion tactics.

First-click attribution assigns all credit to the first touchpoint a customer interacted with before converting, focusing entirely on what drove initial awareness and discovery.

This model is most useful for brands focused on acquisition — understanding which channels introduce new customers — but it ignores everything that happened between discovery and purchase, making it similarly one-dimensional to last-click.

The linear attribution model distributes conversion credit equally across all touchpoints in a customer's journey — if there were five touchpoints, each receives 20% of the credit.

Linear attribution acknowledges that multiple channels contribute to conversion and is easy to communicate to stakeholders, but its equal weighting ignores the different roles and persuasive weights of different touchpoints.

The time-decay model assigns more credit to touchpoints that occurred closer to the conversion event, with credit decreasing exponentially as touchpoints get further back in time.

This model reflects the intuition that recent interactions are more causally connected to the purchase decision than early ones, making it useful for short sales cycles. For long B2B sales cycles, it can still undervalue top-funnel awareness.

The U-shaped (position-based) attribution model assigns 40% credit to the first touch, 40% to the lead conversion touch (or last touch), and distributes the remaining 20% equally among middle touchpoints.

This model recognizes the importance of both acquisition (first touch) and close (last touch) while still giving some credit to nurturing interactions — making it popular for B2B marketers who care about both acquisition channels and conversion channels.

The W-shaped model adds a third key position to the U-shape: the opportunity creation touch (when a sales opportunity is formally opened). Credit is divided 30/30/30 across first touch, opportunity creation touch, and close touch, with 10% distributed across all others.

This three-milestone model is particularly suited to B2B organizations with distinct marketing-to-sales handoff stages and named accounts where multiple marketing interactions occur before a formal opportunity is created.

Data-driven attribution (also called algorithmic attribution) uses machine learning to analyze thousands of actual customer journeys and assign fractional credit to each touchpoint based on its calculated contribution to conversion probability.

Unlike rules-based models, data-driven attribution is empirically derived from your own data — it learns which touchpoint sequences and channel combinations are associated with higher conversion rates — but requires sufficient conversion volume (at least 3,000+ conversions per month) to be statistically reliable.

Use last-click for performance budgeting decisions on conversion campaigns. Use first-click for acquisition channel evaluation. Use linear or time-decay for simple funnel visualization.

Use U-shaped or W-shaped for B2B multi-stage funnels. Use data-driven as your primary model when you have sufficient conversion volume — it is the most accurate reflection of each channel's true contribution across all journey patterns.

GA4 defaults to data-driven attribution for conversions when sufficient data is available, falling back to last-click when data is insufficient. You can change the attribution model in Admin > Attribution Settings.

In Google Ads, each conversion action can have its own attribution model — shifting from last-click to data-driven in Google Ads typically increases reported value of upper-funnel search terms and shifts Smart Bidding toward a broader audience.

GA4's Attribution Model Comparison report allows you to see how different attribution models would credit each channel — side by side — using the same historical conversion data.

This report is invaluable for educating stakeholders about attribution methodology choices and for quantifying how much budget reallocation would be justified if you switched from last-click to data-driven attribution.

An attribution lookback window defines how far back in time from a conversion an attribution model will look for contributing touchpoints — common windows are 7-day click, 30-day click, and 1-day view for view-through.

Shorter windows undercount top-of-funnel contributions and favor channels used close to conversion; longer windows may credit channels from so far back that their actual influence on the conversion decision is questionable. Choosing the right window requires understanding your actual buying cycle length.

Test your understanding of multi-touch attribution models.

Attribution models range from single-touch (first-click, last-click) to multi-touch (linear, time-decay, U-shaped, W-shaped) to algorithmic (data-driven) — each with different strengths and appropriate use cases.

Data-driven attribution is the most accurate model when you have sufficient conversion volume; GA4's model comparison report helps quantify the business impact of changing your attribution methodology.