You paid a creator to post about your product. The Reel got 50,000 views. Comments are positive. The creator tags you, their followers tap through, your Instagram gains 200 new followers.
Sounds like it worked. But did it? How many of those viewers actually bought something?
Most brands can't answer that question. They track vanity metrics because vanity metrics are easy to track. Views, likes, follows, impressions. These numbers go up after an influencer posts, and everyone feels good about it. But feeling good and making money are different things.
The only metric that matters
Revenue. Specifically: how much revenue did this influencer drive versus how much you paid them?
Everything else is a proxy. Views suggest reach. Engagement suggests interest. Follows suggest future potential. But none of these pay for inventory, shipping, or your next influencer deal.
If you spent £300 on a creator and they drove £900 in tracked sales, your return is 3x. That's working. If they drove £100, it's not. The calculation is simple. The hard part is getting the data.
Why most brands can't measure this
The typical influencer marketing workflow: send product, creator posts, check follower count and engagement rate, maybe look at website traffic that week, declare victory or failure based on gut feeling.
The problem is the gap between "someone saw the post" and "someone bought the product." Unless you connect those two events with actual tracking, you're guessing.
Google Analytics shows a traffic spike. Great. Was that from the influencer, or from the email campaign you sent the same week? Your Shopify dashboard shows 15 more orders than usual. Were those from the influencer's audience, or from the Instagram ad you're also running?
Without per-creator attribution, you can't isolate the effect. And if you can't isolate it, you can't optimise it.
Three ways to actually measure it
1. Tracked links (most reliable)
Give each influencer a unique URL. When someone clicks it, the click is recorded. When that person buys, the purchase is attributed to the creator who sent them.
This gives you hard data: Creator A drove 340 clicks and 12 purchases. Creator B drove 89 clicks and 1 purchase. No ambiguity.
The limitation: it only captures sales from people who clicked the link. If someone sees the post, remembers the brand name, and Googles it later, that sale won't be attributed. But the data you do get is reliable.
2. Unique discount codes (common but leaky)
Give each creator a code. SARAH10 gets 10% off. Count redemptions. Simple and familiar.
The problem: codes leak to coupon sites fast. Within a week, people who never saw the creator's post are using the code because they Googled "discount code" at checkout. You end up paying commission on sales the creator didn't actually drive.
Codes work as a rough indicator, but they overcount. Treat the numbers as directional, not precise.
3. Post-purchase surveys (soft data)
Add "How did you hear about us?" to your checkout or confirmation page. Options include each creator's name.
This is the least reliable method. People don't remember accurately. They say "Instagram" when they mean "TikTok." They pick the first option. Some skip the question entirely. But it's free and easy to set up, and it captures the cases that tracked links miss — people who saw the content but didn't click a link.
The best approach uses tracked links as the primary measurement, with surveys as a secondary signal to catch untracked conversions.
What "working" actually looks like
Forget benchmarks from marketing blogs claiming 11x ROI on influencer spend. Those numbers come from enterprise campaigns with large budgets and cherry-picked results.
For a small brand, here's a realistic framework:
Clearly working: 3x+ return. You spent £200 and made £600+ in attributed sales. Scale this creator up. Send more product. Increase the fee. Lock them in.
Marginal: 1-2x return. You broke even or made a small profit. The creator's audience is interested but not converting well. Try a different content format, a stronger offer, or a better call to action before cutting them.
Not working: Below 1x. You spent more than you made. Give it one more attempt with adjusted content, then move on. Some audiences just aren't buyers.
Can't tell: You have no attribution data. This is the most common situation, and it's the worst one. You're flying blind, making decisions based on vibes.
Red flags that it's not working
A few signals beyond the revenue numbers:
High clicks, low purchases. The creator drives traffic but visitors don't buy. This usually means the audience is curious but not your target customer, or your product page isn't converting. The creator might be fine — your landing page might be the problem.
Engagement but no clicks. People like and comment on the post but don't tap the link. The content is entertaining but the call to action is weak, or the audience treats the creator as entertainment rather than a shopping recommendation.
One-time spike, no repeat. You see sales on the day of the post and nothing after. This is normal for one-off partnerships. If you want sustained results, you need ongoing relationships — multiple posts over weeks, not a single mention.
The creator's other partnerships don't convert either. Look at their tagged posts. If they're promoting a new brand every week and none seem to stick, their audience has learned to tune out sponsored content.
Setting up proper measurement
The minimum viable tracking setup:
- Create a unique tracked link for each creator (LinkOwl does this)
- Connect your sales data (Shopify webhook, Stripe, or RevenueCat)
- Share the tracked link with the creator instead of a generic URL
- Wait for the campaign to run (give it at least a week)
- Check your attribution dashboard: clicks, conversions, revenue per creator
Cost: 5p per attributed sale with LinkOwl. No monthly fee. If a creator drives zero sales, you pay nothing for the tracking.
Time to set up: about 10 minutes.
The difference between "I think this creator is good" and "I know this creator drove £400 last month" is the difference between guessing and running a business. Get the data. Make decisions from it. Cut what doesn't work. Double down on what does.