Quick Answer — Why the Same Campaign Has Four Different ROIs
A UTM parameter records where a visit came from. It does not record how much of the sale that visit deserves. Deciding that is what an attribution model does, and the model your dashboard defaults to — last click — hands 100% of the credit to whichever campaign happened to be the final touch before the order.
Feed the same tracked numbers into four models and the credit moves substantially. Take a campaign with $4,800 of spend, 12,500 sessions, 140 last-touch conversions, 96 first-touch conversions, 260 conversion journeys touched anywhere, and an average of 3.2 touchpoints per journey:
- Last click — 140.0 conversions credited
- Position-based (40/20/40) — 98.4 credited
- First click — 96.0 credited
- Linear — 81.3 credited
That is a 72% spread between the most and least generous model, from a single set of numbers. At an $85 average order value and a 62% gross margin, the last-click view returns 3.02x ROAS and 87% ROI. The linear view of the identical campaign returns 1.75x ROAS and 9% ROI — technically still profitable, but nowhere near the story the dashboard told.
The break-even ROAS is 1 divided by the gross margin. At 62% that is 1.61x, so both views clear it, which is the honest conclusion: the campaign works, but it is not the outlier last click makes it look.