Why attribution matters
Almost every paid-media decision starts with an attribution number: which channel deserves credit for that conversion, which campaign delivered the revenue, which ad was the catalyst. Get attribution wrong and you scale the wrong channels, kill the right ones, and reward the platforms that are loudest about claiming credit.
Attribution is also where most marketing teams quietly self-deceive. Meta’s reported ROAS plus Google’s reported ROAS plus TikTok’s reported ROAS routinely exceeds 150% of actual revenue, because each platform claims any conversion it touched. Without an independent attribution view, you can’t tell which credit is real.
How attribution works
An attribution system needs three things:
- Touchpoint data. Every ad view, click, email open, organic visit attached to a user identity.
- Identity resolution. Stitching anonymous sessions on different devices into a single user journey.
- A credit-assignment model. A rule (or learned function) that splits the conversion revenue across touchpoints.
The first two are tracking problems. The third is the attribution model, which is where most of the debate lives, but where the data quality below it matters more than the choice between options.
Attribution models
Common shapes:
- Last-click, 100% credit to the final touchpoint. Simple, biased toward bottom-funnel channels.
- First-click, 100% credit to the first touchpoint. Inverted bias, favours top-funnel.
- Linear, equal credit across all touchpoints. Easy to explain, ignores actual influence.
- Time-decay, credit shifts toward more recent touchpoints. A reasonable compromise.
- U-shaped / position-based, heavy credit on first and last touchpoints, lighter on the middle. Useful when those two roles are most important.
- Data-driven, model learns credit weights from conversion data (typically Shapley values or similar). Requires sufficient volume.
- MMM (Marketing Mix Modeling), top-down regression on aggregate spend and outcomes. Operates above the user-journey level entirely.
No model is “right”, the right model is the one whose credit distribution matches your actual sales motion.
Why attribution breaks
The single biggest cause of bad attribution is bad tracking. ITP, ad blockers, server-side pixel failures, missed events, and identity collisions all silently strip data out of the touchpoint log. When 30-40% of touchpoints are missing, no attribution model can compensate.
Fix tracking first. Then choose a model.
FAQ about Attribution
What is marketing attribution?
Marketing attribution is the discipline of assigning credit for conversions to the marketing touchpoints that influenced them. It is what decides which channels deserve budget.
Why is attribution so hard?
Two reasons: customers touch many channels before converting, and modern tracking misses 30–40% of those touches due to ITP, ad blockers, and cross-device gaps. Both are usually present.
What is the difference between in-platform attribution and independent attribution?
In-platform attribution is what each ad platform reports about itself. Independent attribution uses one consistent model across all channels and adds up to 100% of actual revenue.