AURAInfluencer Contact
Issue Nº 07 · Measurement

How to track influencer ROI per creator.

Campaign-level ROI tells you whether to run influencer marketing again. Per-creator ROI tells you who to re-book, who to drop and who to pay more. Here is how to measure it without fooling yourself.

By the AURA desk9 min read

The short answer

Give every creator a unique tracked link and a unique discount code, add a post-purchase "where did you hear about us?" question, then attribute revenue per creator and divide by what that creator cost you. No single method is complete: links miss delayed and cross-device buyers, codes leak to deal sites, surveys are self-reported. Layering all three and accepting a margin of error beats trusting one number that looks precise and is wrong.

Campaign ROI hides the thing you need to know

A campaign that returns 4x looks like a success. Underneath that average, three creators returned 9x, four broke even and two lost money. Next campaign you re-book all nine, because the only number anyone wrote down was 4x.

This is the practical case for per-creator measurement. Campaign ROI answers a question you only ask once a quarter - do we keep doing influencer marketing? Per-creator ROI answers the question you face every time you build a roster: who is worth paying more, who is worth keeping, and who was an expensive way to make a nice video.

5.2–5.8×the commonly cited average return per unit of influencer spend, with top campaigns reported between 11× and 18× and weak ones below 3×. The spread is the point: averages hide everything.

Why the click and the purchase come apart

Influencer attribution is hard for one structural reason: the moment of discovery and the moment of purchase are usually not the same moment, on the same device, in the same session.

Someone watches a Reel on their phone on Tuesday. On Friday they type your brand name into Google on a laptop and buy. Under last-click attribution, that sale belongs to organic search. The creator who created the demand gets nothing, and next quarter you cut their budget and increase your search spend - optimising your way into paying more for demand you already generated.

On top of that structural problem sit three practical leaks. UTM parameters get stripped by apps, link shorteners and privacy browsers. Discount codes get screenshotted onto deal forums, where they collect orders from people who never saw the creator. And affiliate platforms frequently credit whichever link was touched last, not the one that started the journey.

If you treat any single tracking method as ground truth, you will systematically undercount what influencer marketing actually did.

The four methods, and what each one misses

There is no perfect method. There are four imperfect ones, and they fail in different directions, which is precisely why you run several.

  • Tracked links, one per creator. The backbone. A unique URL per creator means every click and every sale that follows it in-session is attributed exactly. Misses: anyone who buys later, on a different device, or who never clicks because they search your brand instead.
  • Unique discount codes. Catches the delayed buyer that links lose, because the code travels with the customer rather than the session. Misses: nothing, until the code leaks - then it over-credits, quietly, and keeps doing so.
  • Post-purchase survey. One question at checkout: where did you hear about us? Crude, self-reported, and the only method that catches the person who saw a Story, remembered the brand and typed it in a week later. Misses: accuracy - people misremember, and response rates are partial.
  • Holdout and lift tests. Run a region or audience segment without influencer activity and compare. The most honest answer available, and the most expensive to get. Misses: nothing methodologically; it just needs scale and patience most campaigns do not have.

Run the first three on every campaign. Keep the fourth for the moment someone senior asks whether any of this is incremental - because sooner or later, someone will.

The formula

Per creator, over a fixed window that starts when they post and ends at least fourteen days later:

  • Attributed revenue = link-attributed + code-attributed + survey-attributed, de-duplicated so one order is counted once.
  • True cost = fee + product cost + shipping + the hours your team spent briefing and chasing.
  • ROI = attributed revenue ÷ true cost. Cost per acquisition = true cost ÷ orders.

The de-duplication matters more than it sounds. If a customer clicks a tracked link and uses the code, naive addition counts the order twice and hands you a number that flatters everyone. Pick a precedence order - link first, then code, then survey - and apply it consistently.

The second detail people skip is the hours. A creator who needs four rounds of revisions and three reminders can cost more in team time than their fee. It rarely changes the ranking, but it does change who you want to work with again.

What to look at while the campaign runs

Measuring after the fact turns reporting into an autopsy. The value of per-creator tracking is that it is actionable during the campaign:

  • Day 1–3: are the links firing at all? A creator who posted without the link, or with a broken one, is a fixable problem on day one and an unanswerable question on day thirty.
  • Day 3–7: which formats convert? If Reels are outperforming static posts three to one, the briefs for the second half of the roster should change now, not next campaign.
  • Day 7–14: who deserves more budget? The creators converting above the roster median are where incremental spend belongs - a second post from a proven creator generally beats a first post from an unproven one.

A note on earned media value

Most reported influencer ROI is not revenue. It is earned media value: an estimate of what you would have paid in ads to reach the same people. Around 83 percent of marketers treat EMV as a legitimate measure of return, and published EMV benchmarks sit near 4.87 per unit spent.

EMV is a reasonable way to value awareness. It is not money. A report that presents EMV next to a euro sign, without saying so, is not lying exactly - but it is letting a finance team believe something that is not true. If you use EMV, label it. Keep it in a different column from attributed revenue, and never sum the two.

Key takeaways

  • Measure per creator, not per campaign - the average hides both your best and your worst performer.
  • Run tracked links, unique codes and a post-purchase survey together; each covers a different blind spot.
  • De-duplicate before you add, or you will double-count the customers who used both.
  • Judge nothing before day fourteen. Discovery and purchase are days apart.
  • Never present earned media value as revenue, and never add the two together.

Frequently asked

How do you calculate influencer marketing ROI per creator?

Take the revenue attributed to that creator, subtract what you paid them plus any product, gifting and shipping cost, and divide by that total cost. A creator paid 800 euro who drove 3,400 euro in attributed revenue returned 4.25x. Do it per creator, not per campaign, or your best and worst performers cancel each other out into a meaningless average.

What is a good ROI for influencer marketing?

Published benchmarks put the average return somewhere between 5.20 and 5.78 euro or dollar per unit spent, with top-performing campaigns reported between 11x and 18x and weak ones under 3x. Treat those as a sanity check, not a target: methodologies differ wildly and most of them count earned media value rather than money that actually arrived.

Why is per-creator attribution so hard?

Because the purchase rarely happens where the click did. Someone sees a Reel on their phone, searches your brand three days later on a laptop and buys from an email. Last-click gives the credit to email. UTM parameters get stripped, discount codes get shared to deal sites, and affiliate platforms often credit the last touch rather than the one that created the demand.

Are discount codes or tracked links better for attribution?

Neither on its own. Tracked links catch the people who click through immediately but miss anyone who buys later or on another device. Codes catch the delayed buyers but leak once someone posts the code publicly. Run both, and add a post-purchase survey question for the gap neither of them sees.

How long should you wait before judging a creator's performance?

Give it at least fourteen days after the post. Influencer-driven purchases are not instant: discovery and purchase are often days apart, so a creator judged on day two looks like a failure and the same creator judged on day fourteen can be your best performer.

Sources

See it on real campaigns

AURA does the layering described above for you: a tracked link per creator, revenue attributed per post, and ROI per creator as a live number while the campaign runs.

Book a demo →