Attribution decides which marketing channel gets credit for conversions. Here’s when each model applies — and why it matters.
Why Attribution Matters
Attribution determines how budget gets allocated. Last-click attribution says Google Ads drove the sale — so allocate more to Google Ads. First-click says social media drove it — so more to social. Different models tell different stories. Get attribution wrong and you’ll systematically under-fund channels that drive demand and over-fund channels that capture it. The choice of model matters as much as the data inside it.
Last Click: The Default (and Its Flaws)
Gives 100% credit to the last marketing touchpoint before conversion. Simple, easy to explain, but unfair to upper-funnel channels. Customer sees your Instagram post, reads your blog, clicks a Google Ad, and converts. Last-click gives all credit to Google Ads — ignoring the social and content that created awareness. Last-click over-invests in bottom-funnel and starves the channels that produce future demand.
First Click
Opposite of last-click. Gives credit to the first touchpoint. Favors discovery channels (social, SEO, display). Flawed in different ways — it ignores the closing work. Real customer journeys involve 5–20 touchpoints; crediting only the first or last is obviously incomplete. Some brands use first-click for awareness reporting and last-click for conversion bidding — both views at once.
Linear and Time-Decay
Linear: equal credit to every touchpoint. Fair but unhelpful — every channel looks equally important. Time-decay: weights recent touchpoints more heavily. Generally more useful than last or first click. Example: social touchpoint 14 days before conversion gets 10%; search touchpoint 1 day before gets 60%. Time-decay reflects how humans actually forget earlier touches and respond to recent ones.
Position-Based (U-Shape)
Gives 40% to first touchpoint, 40% to last, 20% split among middle touchpoints. Credits both discovery and conversion channels. Good default for B2B where both matter significantly. Less accurate than data-driven but more intuitive than ‘data-driven algorithm.’ For companies without enough conversion data for data-driven, position-based is the best compromise model.
Data-Driven Attribution (GA4 Default)
Machine learning-based. GA4 compares conversion paths to non-conversion paths and assigns credit based on which touchpoints actually move conversions. Most accurate but requires 3,000+ ad clicks and 300+ conversions over 30 days — many accounts don’t qualify. When data-driven is available, use it. Its dynamic nature means it adapts as customer behavior shifts.
Choosing a Model
Small accounts: last-click (simple, sufficient). Medium accounts with some brand awareness: time-decay. Mature accounts with B2B long cycles: position-based. Large accounts with volume: data-driven. Test in GA4’s Attribution report — compare channel performance across models. The ‘right’ attribution is whichever lets you make better budget decisions. There’s no perfect model — there’s the least-wrong model for your data volume.
Need help with analytics?
Get a free audit of your analytics setup. We’ll show you exactly where the opportunities are.
Get Free Audit →