Attribution modeling is the framework you use to answer "which marketing channels caused this revenue?" The choice of model isn't neutral — different models will tell you different stories about the same customer journey, and budget decisions follow the model.
The five common attribution models
Last-click attribution
100% of conversion credit goes to the last touchpoint before purchase.
- Strengths: Simple, easy to explain, default in most tools historically
- Weakness: Undervalues awareness channels (display, social, video) that initiated the journey
- Best for: Direct-response, short purchase cycles, transactional decisions
First-click attribution
100% of conversion credit goes to the first touchpoint that brought the customer.
- Strengths: Highlights what's creating initial awareness
- Weakness: Ignores the closing channels and remarketing that drove the final decision
- Best for: Understanding awareness investments, brand-building analysis
Linear attribution
Equal credit distributed across every touchpoint in the journey.
- Strengths: Honors that every touch matters; democratic
- Weakness: Doesn't reflect that some touches actually influence more than others
- Best for: When you want a balanced view across all channels
Time-decay attribution
Touchpoints closer to conversion get exponentially more credit; earliest touches get the least.
- Strengths: Reflects that recency often matters; weighted toward what actually closed the deal
- Weakness: Still arbitrary; the decay curve is a guess
- Best for: Longer sales cycles where late-stage touches drive decisions
Data-driven attribution (DDA)
Algorithmic model that learns from your actual conversion path data which touchpoints predict conversion.
- Strengths: Adapts to your specific business; less arbitrary than fixed models
- Weakness: Black-box (hard to explain), requires sufficient data volume, varies between platforms
- Best for: Mature businesses with substantial conversion volume; GA4's default
Position-based / U-shaped attribution
40% credit to first touch, 40% to last touch, 20% distributed across middle touches.
- Strengths: Honors both discovery and closing while accounting for middle nurture
- Best for: Lead-gen B2B businesses where awareness and close both matter materially
How to choose
There's no single right answer. The right approach for most businesses is running multiple models in parallel:
- Primary model: Data-driven (default in GA4) for ongoing optimization
- Secondary model: First-click to understand awareness investments
- Tertiary model: Last-click for comparing against historical UA data
When models agree on a channel's value, you can be confident. When they disagree dramatically, that's a signal to investigate the channel's role more carefully.
What attribution can't do
- Capture every touch — offline conversations, podcast listens, peer recommendations don't show up
- Measure brand impact — ads that build memory without immediate clicks are systematically undervalued
- Handle long latency — touches outside typical attribution windows (30-90 days) get dropped even when they influenced the eventual decision
- Account for view-through impact — ads seen but not clicked still affect behavior; some platforms try to model this, none do it perfectly
For a more complete picture, attribution modeling pairs with brand surveys, marketing mix modeling (MMM), and incrementality testing. Attribution is necessary but rarely sufficient.