Attribution modelling is the methodology used to assign credit for a conversion to the marketing touchpoints a customer encountered before converting. A customer might have clicked a Google Ad, then come back via organic search, then clicked an email link before purchasing. Which touchpoint gets credit? Attribution models answer this differently: last-click (all credit to the email), first-click (all credit to the Google Ad), linear (equal credit to all three), or data-driven (GA4's default, which uses machine learning to distribute credit based on actual contribution patterns). The model you use significantly affects which marketing channels appear most valuable and therefore where you invest.
Most UK businesses using GA4's default settings are using data-driven attribution — which Google's machine learning applies across channels — without necessarily understanding it. This is generally appropriate for most use cases, but understanding attribution is essential when you are making budget decisions based on channel performance reports: the model underpinning those reports determines what you are actually measuring.
Common attribution models and their implications
- Last-click — 100% credit to the final touchpoint before conversion; tends to over-value bottom-of-funnel channels (branded search, email) and under-value top-of-funnel channels (display, social media) that build awareness earlier in the journey
- First-click — 100% credit to the first touchpoint; over-values awareness channels that introduced the customer, under-values channels that closed the sale
- Linear — equal credit split across all touchpoints; reasonable but ignores that some touchpoints have more influence than others
- Time-decay — more credit to touchpoints closer to conversion; logical for short sales cycles, less appropriate for long B2B consideration periods
- Data-driven (GA4 default) — machine learning distributes credit based on the probability of conversion with and without each touchpoint; most accurate but requires significant conversion volume to produce reliable results
- Position-based (40-20-40) — 40% to first touchpoint, 40% to last touchpoint, 20% split across middle touchpoints; a reasonable compromise for mixed funnel strategies
For most UK businesses with sufficient conversion volume (100+ conversions/month), GA4's data-driven attribution is the most accurate and is the recommended default. For businesses with lower conversion volume where data-driven models are unreliable, a position-based (40-20-40) model provides a reasonable balance between recognising awareness-driving channels and closing channels. Avoid last-click exclusively if you run any significant upper-funnel marketing (display, social media, SEO for informational queries), as it will systematically undervalue those channels and lead to underinvestment in acquisition-driving activities.
Universal Analytics defaulted to last-click attribution for all standard reports. GA4 defaults to data-driven attribution for Conversion and Advertising reports, while other reports (like Traffic Acquisition) use a last-click model for session-level attribution. This means conversion numbers in GA4 may differ from UA for the same period — not because conversions changed, but because the attribution model distributes credit differently across channels. When comparing GA4 data to historical UA data, be aware that this model difference affects channel-level conversion counts even when total conversions are the same.