Every agency knows the meeting. The campaign wrapped two weeks ago, the client's marketing director looks up from the deck and asks: so what was the ROI of influencer marketing for us this quarter? And every agency knows the night before that meeting: a campaign manager pulling reach from one platform, engagement from another, link clicks from a third, pasting it all into slides and hoping the numbers assemble themselves into an answer.
The awkward truth is that most published advice on the ROI of influencer marketing is written by companies selling measurement software, so it reads more confident than practitioners actually feel. Ask agency people privately and you hear something more honest: the value is real, the measurement is messy, and the client still deserves a straight answer. This piece is about giving them one: which KPIs to track for which goal, which costs to count (including the ones almost nobody counts), and what to put in the report.
What is the ROI of influencer marketing?
At its plainest: the value a campaign generated, minus what it cost, divided by what it cost. Positive means the campaign produced more than it consumed.
The formula is simple. Neither input is. "Value" is easy when a creator's tracked link drives a sale, and hard when a campaign shifts brand awareness that pays off months later. "Cost" looks easy (creator fees plus media spend) and is where most calculations quietly go wrong, because they leave out the hours it took to run the thing. We will come back to that.
Why measure at all, if it is messy? Two reasons. Budget defence: influencer spend competes with channels that arrive with dashboards attached, and a channel that reports nothing loses that argument by default. And learning: without measurement you cannot tell which creators to rebook and which to drop, and rebooking proven creators is the cheapest performance improvement available to any campaign.
How to measure influencer marketing: match the KPIs to the goal
Most measurement problems start before the campaign does, when nobody wrote down what it was for. Awareness, consideration and conversion campaigns need different influencer marketing KPIs, and a report that mixes them convinces nobody.
Awareness. You are buying attention. Measure reach, impressions, video views and cost per thousand people reached. Creator mix matters here: taking the follower bands most of the industry uses (nano at roughly 1–10k followers, micro at roughly 10–100k), an awareness campaign built on forty nano creators is measured very differently from one built on three large accounts.
Consideration. You are buying interest. Measure engagement rate, saves and shares, the substance of the comments (are people asking where to buy, or tagging friends?), and profile visits or branded search movement where the client can see it.
Conversion. You are buying action. Measure clicks on tracked links, promo code redemptions, and the sales or sign-ups the client can tie back to the campaign.
The commonest reporting mistake is running a conversion brief and reporting engagement because engagement looked better. Clients notice. The second commonest is failing to set measurement up in the brief itself: if tracked links, codes and deliverables were never agreed in writing, no report can recover them afterwards. We cover that end of the problem in How to write an influencer brief creators actually follow.
The denominator problem: the costs almost nobody counts
Here is the part the software-vendor guides skip. The cost side of the formula usually contains creator fees, paid amplification, perhaps content production and a tool subscription. It almost never contains the operational hours, and those are large.
The number we hear consistently from senior people at Benelux PR firms (not a study, not a benchmark, just the same answer whenever people count honestly) is roughly ten hours of manual operations per creator per campaign: sourcing, vetting, outreach, contracting, briefing, chasing, reporting. On a 30-creator campaign that is around 300 hours; at five campaigns a year, roughly 1,500. Take whatever a loaded hour of a campaign manager costs on your payroll (salary plus overheads) and multiply it out. The result is a cost line that often rivals the creator fees themselves. We unpack where those hours go in The 10-hours-per-creator problem.
If the hours it takes to run the campaign never enter the denominator, your ROI number is flattering you.
This matters in both directions. It matters for honesty: an ROI computed without the ops hours overstates every campaign you have ever run. And it matters for improvement: once the hours sit visibly in the calculation, cutting them becomes a legitimate way to raise ROI (the denominator shrinks) without touching creative, creators or media spend at all.
When attribution leaks, which it will
Tracked links undercount. Someone watches a story on Monday, searches the brand on Thursday, buys on Saturday: no click recorded, no credit given. Promo codes travel beyond the creator's audience. And last-touch attribution (giving all the credit to the last thing a buyer clicked) systematically short-changes a channel that works early in the journey.
Large advertisers answer this with marketing mix modelling: a statistical method that estimates each channel's contribution from top-level sales data rather than individual clicks. Most clients of a 10-to-80-person agency do not have that, and pretending otherwise helps nobody.
What works without it: pick two or three signals you can actually collect (tracked clicks, code redemptions, branded search movement), define them before the campaign, collect them the same way every time, and report direction rather than false precision. "Directionally positive, and here is the evidence" is a more credible sentence in front of a board than a suspiciously exact percentage that falls apart under one question.
What is a good ROI for influencer marketing?
You will find confident average figures online, repeated near-verbatim across sites that sell influencer software. We are not going to hand you one, because a generic benchmark misleads more than it informs. A supplement brand with repeat purchases, a retailer with thin margins and a B2B firm with an eighteen-month sales cycle have nothing useful in common on this number. Neither do an awareness campaign and a conversion campaign for the same client.
The useful comparison is the client against their own history. Measure three campaigns the same way (same KPIs, same cost definitions, hours included), and by the third report you can write "cost per engagement is down against your baseline" and defend it in the room. A baseline you built beats any benchmark you borrowed.
The influencer marketing report: what to send, and when
A cadence that holds up with clients:
Weekly, during the campaign. One page. What went live, what posts next, what needs a decision. No ROI claims mid-flight: the data is not in yet, and everyone knows it.
The wrap report, within two weeks of the last post. The goal restated in one sentence. The KPIs that were agreed for that goal, and only those. Full costs, hours included. What worked, what to drop, and a rebook-or-not call on every creator.
Quarterly, for retained clients. The baseline trend: this quarter's campaigns against the client's own history.
Now the uncomfortable part. Assembling that wrap report is exactly the work described above: numbers pulled from three platforms, stories screenshotted before they expire, everything reformatted into the client's template. Rule-following work, no judgement in it, and it is precisely the kind of drudgework automation handles well. It is where we point a first pilot, and we treat that pilot as failed unless manual time on the step drops by at least 70%. What the numbers mean (the analysis, the recommendation, the conversation where the client pushes back) stays with your team, because that is what the client is paying for. Who should build and run that layer is its own question, which we take apart in Who should own your agency's AI stack?.
The practical version of all this fits in four moves. Before the next campaign: write the goal in one sentence and pick three KPIs that match it. Put the tracking (links, codes, deliverables) into the brief, not into the post-mortem. Count all the costs, including the hours; ten per creator is the honest starting assumption until you have counted your own. Then report against the goal you wrote down, at a steady cadence, so that three campaigns from now your client has a baseline of their own instead of somebody else's benchmark. None of this makes influencer marketing ROI precise. It makes it honest. And honest, consistent measurement is what keeps budgets.