Measuring Influencer ROAS, Honestly

Ask five agencies for the ROAS of the same campaign and you'll get five different numbers — sometimes off by 3–4x. The reason isn't dishonesty so much as measurement choices made before the campaign started. Last-click attribution punishes creators because their job happens early in the journey: a viewer sees a Reel on Tuesday, searches the brand on Friday, and buys through a Google ad on Sunday. Google's ad gets the credit; the creator who manufactured the demand gets zero.

The opposite failure is just as common. Campaigns measured on 'earned media value' or platform-reported view counts produce numbers that look spectacular and mean almost nothing to a CFO. EMV answers the question 'what would this exposure have cost as paid media?' — which is not the question anyone paying the invoice is asking. If a metric can't survive a finance review, it shouldn't headline your campaign report.

First lie: creator content that drives branded search 'didn't convert.' Across our own campaign data, branded search volume routinely lifts 20–60% during a well-run creator flight — and those searchers convert at rates paid social can't touch. Last-click books all of it as 'organic search.'

Second lie: the discount code tells the whole story. Codes and affiliate links capture only the buyers who remember them at checkout — typically a minority of creator-driven purchases. They're a floor, never the total. Treat coded revenue as a directional signal and calibrate a multiplier against holdout data, not as the campaign's final grade.

Third lie: the conversion window matches the buying cycle. A 7-day click window makes sense for retargeting; it makes no sense for a ₹40,000 appliance or a skincare routine someone researches for three weeks. When the measurement window is shorter than the consideration cycle, creator campaigns will always look worse than they are.

Layer one is deterministic tracking: UTM-tagged links, unique codes, and landing pages per creator. Cheap, precise, and understated — use it to rank creators against each other, not to size the total effect. Relative performance between creators is one thing last-click gets roughly right.

Layer two is lift-based: geo holdouts, pre/post branded-search analysis, and matched-market tests. Run the campaign in some cities and not others, then compare. This is the only method that credibly answers 'what happened because of this campaign that wouldn't have happened anyway?' It requires planning before launch — you can't reconstruct a holdout afterwards.

Layer three is survey-based: a one-question 'how did you hear about us?' at checkout. It's unfashionable and imperfect, and it consistently surfaces 2–3x more creator-attributed purchases than click tracking does. When all three layers point the same direction, you can defend the number in any room.

Set the target before choosing creators, because the honest benchmark depends on the job. Direct-response campaigns with strong offers and conversion-optimized creators can be held to blended ROAS targets in the same way paid social is — just with the longer window and the search-lift adjustment applied. Awareness campaigns should be held to cost-per-incremental-reach and branded-search lift, and it should be agreed in writing that nobody will judge them on 7-day revenue.

The most expensive mistake we see is a hybrid brief graded on a pure-DR scorecard. If the brief asked creators for storytelling and brand-building, grading the output on next-week ROAS guarantees disappointment — and usually kills a channel that was actually working.

Report three numbers, clearly labelled: tracked revenue (the deterministic floor), modelled incremental revenue (with the methodology stated), and cost efficiency versus the channel it displaces. Never blend them into one 'ROAS' figure — the blend is where credibility goes to die.

Then add the operational metrics finance rarely asks about but should: content production cost per usable asset versus your studio baseline, and the paid-media performance of whitelisted creator content versus brand-shot creative. In many of our campaigns the licensing value of the content alone justifies a meaningful share of the creator fees — before a single organic view is counted.