How to Track Influencer Marketing ROI (Properly)

Ask ten brands what their influencer marketing returned and nine will show you reach. Reach is not a return. If you can't say what a creator sold, you can't decide whether to pay them again — and you'll spend another quarter renewing on vibes.

Here's the tracking stack we put behind every creator campaign, in the order you should build it.

Layer 1: unique codes, one per creator

The simplest thing that works. Every creator gets their own discount code — RITIKA10, not SAVE10.

• It's self-reporting: the buyer types the code, so attribution needs no cookie
• It survives iOS privacy changes, ad blockers and cross-device journeys
• It gives the creator a reason to say it out loud, which improves recall

Its limits are real too: codes under-count (people forget them, or buy later without one) and get scraped onto coupon sites, which steals credit from other channels. Treat codes as your floor, not your ceiling — and keep them off the aggregator sites by rotating them.

Layer 2: tracked links with disciplined UTMs

Every link a creator posts carries UTMs, and they are consistent or they are useless:

utm_source=instagram
utm_medium=influencer
utm_campaign=spring_launch
utm_content=ritika_skin

The rule that matters: utm_content is the creator identifier, always spelled the same way. One creator spelled three ways is three creators in your reports and zero useful conclusions.

Use short branded links so they're typeable from a bio or a spoken URL, and never let a creator shorten your link themselves — you lose the parameters.

Layer 3: server-side events

Browser pixels miss a large share of conversions — ad blockers, privacy settings, in-app browsers, delayed purchases. If your only data source is the pixel, you're under-reporting creator performance and quietly starving your best creators of budget.

Send purchase events from your server (Conversions API, server-side GTM) with the creator identifier attached. Now the platforms can optimise on real outcomes when you amplify the post as an ad, and your own reporting stops disagreeing with your bank account.

Layer 4: the incrementality check

Codes and links tell you what got attributed. They don't tell you what was incremental — whether those buyers would have purchased anyway.

You don't need a research team for this. Two cheap methods:

• Geo holdout: run the campaign in some regions and not others, compare like for like
• Time-boxed pulse: measure baseline revenue for two weeks, run creators for two weeks, compare — accounting for anything else that changed

Do this once a quarter, not per campaign. It calibrates how much to trust your day-to-day attribution.

The report that actually gets used

One page, one row per creator, and exactly seven columns: creator handle, fee paid, orders, revenue, cost per acquisition, return on creator spend, and owned assets delivered.

That's it. No reach column, no engagement rate, no screenshots. When a creator returns 6.1× and another returns 0.7×, the page has done its job — and it fits on a phone screen, which is where your founder will read it.

Then three decisions, every month: re-sign, re-brief, or cut. That's the entire management layer, and it's why programs that track compound while programs that don't just repeat.

Four mistakes that quietly wreck the data

1. Shared codes across creators. You've merged your two most important rows.
2. UTMs invented per post. Inconsistent spelling destroys aggregation.
3. No baseline. Without a pre-campaign revenue baseline, every number is unanchored.
4. Counting the post, not the asset. If you have usage rights, the content keeps earning as a paid ad — measure its whole life, not its first 48 hours.

What to build first

If you're starting from zero: unique codes this week, UTM discipline next week, server-side events within the month. That order gets you a defensible number fast, then makes it accurate.

The goal isn't a prettier dashboard. It's being able to say, out loud and with evidence, "these five creators make us money — give them more budget."

---

We wire codes, tracked links and server-side attribution into every creator campaign, and report revenue per creator each month. See the service, or read how we do the same for performance marketing and analytics.