Agency Marketing Influencer
Core topic

Measuring an influencer campaign

Every campaign report contains numbers that sound like results and numbers that are results. Telling them apart is the single most useful skill a brand-side buyer can have, because the first kind is always larger and always arrives first.

Signal strength

How much each number is worth

Ranked by how strongly the metric supports a claim that the campaign worked. The bars are an editorial judgement, not a measurement — but the ordering is not controversial, and the metrics at the top of a typical agency report are usually at the bottom of this list.

Revenue attributedHard Sales traced to a code or link. The only number that settles an argument.
ConversionsHard Sign-ups, installs, bookings. Countable, and tied to an action you asked for.
Cost per resultHard What each action cost. The number that lets you compare against every other channel.
ClicksMedium Real intent, but only counts the traffic that used your link.
Engagement rateMedium Best used before you buy, to judge whether an audience is real.
Saves and sharesMedium Weak, but the most honest of the engagement signals — both cost the viewer something.
ImpressionsSoft Counts screens the post appeared on, not people who noticed it.
Earned media valueSoft A model of what equivalent ads would have cost. Every vendor computes it differently.

On earned media value

EMV converts attention into a dollar figure by estimating what buying the same reach would have cost. It is not a fabrication and it is not a measurement — it is a model, and the multiplier that drives it is chosen by whoever is reporting.

Two agencies can report EMV for identical campaigns and differ by a factor of three, both honestly. Ask what multiplier was used and what it was benchmarked against. If nobody can answer, the number is decoration. This site quotes EMV once, on the Fenty Beauty page, and labels it as modelled.

Set up attribution before the brief

Almost all unprovable influencer ROI is an attribution failure rather than a campaign failure. Four things, all cheap, all much harder to add afterwards:

  1. Give every creator their own code or link

    Not one campaign code. One per creator, so the report is a ranked list rather than a total.

  2. Agree the attribution window in writing

    Influencer purchases lag. A 24-hour window will show you a failure; 30 days is more honest for considered purchases.

  3. Capture a baseline

    Two weeks of normal traffic and sales before launch. Without it, every subsequent number is uninterpretable.

    If the agency does not ask for a baseline, ask why they expect to prove anything.
  4. Ask new customers where they heard about you

    A single post-purchase question catches the demand that never touched a tracked link — which on a working campaign is most of it.

What good looks like

There is no universal benchmark, and any agency offering one is selling. What is portable is the comparison: measure influencer against your existing paid social on cost per result, using the same window and the same definition of a result. That question — "cheaper or dearer than the channel I already run?" — is answerable, and it is the one your finance team is going to ask anyway.

A quick test

Ask for these

  • Cost per acquisition, next to your other channels
  • Revenue by individual creator
  • Results against the pre-campaign baseline
  • Which creators underdelivered, and what happened next

Treat these as context only

  • Total impressions
  • Combined follower reach across the roster
  • Earned media value with no stated multiplier
  • Engagement rate quoted after the campaign rather than before