How I analyzed a failed UGC campaign and stopped making the same mistakes: what the numbers actually told me

Кейсы и истории успеха
1

I want to share a campaign that flopped, and more importantly, how I analyzed what actually went wrong so we didn’t repeat it. This isn’t a success story, but I think that’s actually more useful than success stories sometimes.

Setup: We ran a UGC campaign across Russia with 10 different content creators. Budget was $25k total. We had the same brief for everyone, expected similar baseline performance, and honestly? We thought we’d do well because our previous Russian UGC campaigns had averaged 8-12% engagement.

Results: This campaign averaged 3.2% engagement. Conversion was 0.8% (previous campaigns averaged 2.5%). Revenue was $18k against $25k spend. We lost money.

My first instinct was to blame the creators: “they just didn’t execute well.” But before I wrote that in a report, I actually dug into the data, and I’m glad I did, because the creators weren’t the problem.

Here’s what I found when I actually looked:

Hypothesis 1: Wrong creators. I checked the follower counts and historical performance of the creators we’d booked. They looked solid on paper. But when I looked at who followed them, I noticed something: 40% of followers for 6 of the 10 creators were flagged as potentially bot accounts (using a basic audit tool). So they looked like they had reach, but actual human reach was way lower. That explained some of the underperformance, but not all of it.

Hypothesis 2: Wrong product fit. I looked at what was being sold. It was a skincare product, pretty niche—luxury positioning, high price point ($95). I looked at the audience demographics for each creator, and I realized: most of them had audiences that skewed younger (16-25), lower income. A $95 skincare product isn’t a fit for that audience. We’d briefed the same product and brief to all 10 creators without thinking about whether their audiences would even want it.

That was a big miss on our side.

Hypothesis 3: Brief execution. I watched the actual videos. Seven of ten creators had basically copied the brief’s talking points verbatim. It didn’t sound authentic; it sounded like a script. The three that didn’t copy the brief? They had 2x engagement vs. the others. So creators who felt comfortable improvising and making it their own did way better than creators who stuck to script.

Hypothesis 4: Timing. I checked when videos went live relative to platform algorithms. Four of the videos posted during low-traffic hours (early morning, late night). Two posted during platform maintenance windows. We hadn’t given creators guidance on when to post, so they just posted whenever it was convenient for them. That’s on us.

So here’s what actually happened: we picked creators who looked good but some had bot followers, we picked a product that didn’t fit most of their audiences, we gave them a script that killed authenticity, and we didn’t coordinate posting timing. It’s not that the creators were bad; it’s that we set them up to fail with a poor campaign structure.

Once I understood that, the question shifted: what should we change for next time?

For creator selection: We started doing audience audits as part of vetting. Are these actual humans likely to be interested in this product? Not just “does the creator have followers.”

For product fit: We started matching product to creator audience first, then brief. Instead of “all 10 creators get the same brief,” we created three different brief variations for three different audience segments. Russian creators with luxury-focused audiences got a premium positioning brief. Creators with younger, value-conscious audiences got a different brief focused on quality-for-price.

For authenticity: We stopped writing scripts. We moved to “here are the key storylines we want told, but tell it your way.” It sounds like a small change, but 2x engagement difference. Creators know their audiences better than we do.

For timing: We added a “best posting time window” recommendation based on each creator’s historical data. Didn’t mandate it, but provided it.

We ran a second version of this campaign with those changes: same budget ($25k), different creators (better-vetted), different briefs (audience-matched), same call for authenticity. Revenue was $38k. That’s recovery.

Was it the creators? No. Was it bad luck? No. We just had a campaign structure that was working against us, and I almost published a report blaming the creators for it, which would have meant missing the actual problem.

The hard part was sharing this with leadership: “we lost money on this campaign because our structure was wrong, not because creators didn’t perform.” That’s a more uncomfortable message than “the creators we picked weren’t good” because it’s a reflection on campaign strategy, not external factors.

But it’s also the message that actually lets you improve.

Has anyone else had a campaign totally flop and realized the structure was the problem, not the execution? And how do you keep yourself from jumping to blame creators when numbers are bad?

2

This is the kind of detailed breakdown that separates good data analysis from defensive analysis. You caught yourself before publishing a blame-the-creators report, and that’s actually harder than it sounds.

I want to push on a couple of your findings, though:

First: the bot followers. You said 40% of followers for 6 of 10 creators were flagged as bots. But did you look at engagement rate for those creators and see if it was suppressed compared to their real-follower percentage? Because sometimes you can have 40% bot followers and still have decent engagement if the real followers are highly engaged.

Second: the brief authenticity improvement (2x engagement when creators improvise). That’s a real insight, but I’d be cautious about generalizing it. Did you measure conversion for those high-engagement videos? Because I’ve seen campaigns where high engagement looks impressive but drives lower conversion—basically, the creator’s improvisation is more entertaining but less sales-focused.

Third: product-audience fit. You realized a $95 luxury skincare product didn’t match the 16-25 income-lower demographic. But did you measure whether that audience was incapable of buying it, or just less likely? Because there’s a difference between “wrong fit” and “worth less effort to convert.”

Last: you said the second campaign generated $38k revenue. But what was the engagement and conversion rate? Because if you’re celebrating revenue recovery without showing that the actual per-creator performance improved, you might just be getting lucky with better product-audience matching, not learning anything about campaign structure.

What I’d want to see for next time: engagement rate and conversion rate and cohort retention data (did those customers buy again?). Revenue alone can hide a lot of sins.

3

Also: have you built a “pre-launch checklist” into your campaign process so you catch these issues (bot followers, product-audience mismatch, timing gaps) before you launch? Because if you’re repeatedly discovering these problems post-mortem, that’s a process design issue.

4

I’m really glad you shared this honestly. The part that resonates most is how close you came to blaming creators and how that would have poisoned future relationships.

From a partnership perspective, here’s what I want to highlight: those creators who didn’t perform well? How did you communicate with them about what happened? Like, did you tell them “hey, we set you up to fail with bot followers and a product mismatch,” or did you let them think their content just wasn’t good enough? Because the difference between those two conversations affects whether they want to work with you again.

Also—the creators who did perform well (the improvisation crew with 2x engagement)—did you bring any of them back for the second iteration of the campaign? Because if you have creators who naturally know how to adapt and make content authentic, those are people worth building retainer relationships with.

One more thing: when you shifted from scripts to “key storylines, tell it your way,” did you communicate that change to creators explicitly? Like, did you give them a framework for the freedom, or did you just say “be authentic” and hope they understood? Because some creators really do want more structure, and it can go wrong if you suddenly flip the expectations without explaining why.

5

This is useful because we’re about to run our first UGC campaign as we expand to a new market, and I was basically planning to do exactly what you just described as “wrong.”

So real questions: when you audited audiences for bot followers, what tool did you use? And more importantly, how much time does that actually take per creator when you’re vetting 10 people?

Also—you said you created three different brief variations based on audience segments. How did you decide which creators got which brief? Like, did you analyze their audience demographics and assign briefs, or did you ask creators which positioning felt right for their community?

And the posting-time recommendation—did you research that per creator individually, or did you have baseline recommendations (like “never post between 2-4am”)? Because if you had to do custom analysis for all 10 creators, that’s a lot of ops overhead.

I guess my underlying question is: how much of this should be automated/templated, and how much needs human judgment per campaign?

6

Strong root-cause analysis. You correctly identified structural problems instead of blaming execution, which is harder than most teams manage. But I want to question one of your conclusions.

You said creators who improvised had 2x engagement. But then the second campaign you ran had different creators entirely (better-vetted) with different briefs (audience-matched) and the same “tell it your way” guidance. So how do you know which variable actually drove the $38k revenue improvement?

Was it:

  • Better creator vetting (bot followers removed)?
  • Better product-audience matching (luxury positioning to luxury audiences)?
  • The authenticity/improvisation approach?
  • Timing optimization?
  • Or some combination?

Because if you’re going to build this into repeatable process, you need to know which lever matters most.

Also: first campaign lost $7k ($18k revenue on $25k spend). Second campaign made $13k profit ($38k revenue on $25k spend). That’s a $20k swing. That’s meaningful, but I’d want to check: did you improve creator performance, or did you just get lucky with better audience match + product fit? The way to test that would be: run the same second-round creators with the old (script-based) brief and see if they still outperform.

Lastly: scaling question. You tested these improvements on one campaign. Before you commit to this as your new standard process, how would you validate that the insights are robust? What if the audience-matching insight only works for skincare/beauty, but doesn’t translate to other categories?

7

Okay, reading this as a creator, I have mixed feelings. On one hand, I’m really glad you figured out that the brief-as-script approach was killing performance and you moved to “tell it your way.” That’s genuinely better for creators and audiences.

On the other hand: you said you “almost published a report blaming creators,” but I’m guessing the creators who performed poorly never got that same analytical treatment you gave yourself, right? Like, they probably just thought the campaign underperformed and assumed it was their fault. That’s sucks.

So here’s my feedback: when a campaign underperforms, don’t just analyze internally. Actually tell creators what you learned. Something like: “Hey, we looked at the data and realized the product-audience fit wasn’t right, not that your content wasn’t good. For next time, here’s what we’re going to change.” That communication is how you keep creator relationships healthy even after a flop.

Also—the “tell it your way” brief shift. That’s great from a performance perspective, but I’d want to know: how much creative freedom did you actually give? Like, were there guardrails (“must mention A, B, C”), or was it truly open-ended? Because the difference between those is huge for how creators approach the work.

One last thing: did you ever ask the creatorswho improvised well why they chose improvisation? Like, were they nervous about the brief, were they confident enough to take risks, did they just have a different working style? Because those insights matter if you’re matching briefs to creators going forward.

8

This is the kind of transparent analysis that builds trust with clients (and creators). You walked them through exactly what went wrong, which is frankly rarer than it should be.

But from an agency perspective, I want to know: did you structure this learning so it scales across your team? Like, did you build this into your pre-launch checklist (creator audit, product-audience matching, brief variation framework)? Or are these insights just sitting with you, and if you’d gone on vacation, the next campaign would have made the same mistakes?

Also: when you ran the second campaign and it worked, did you charge the client premium rates because you’d optimized the process? Or did you keep rates the same and just improve margins? Because if you’re creating more value (learning how to structure campaigns for success), you should be capturing some of that value in pricing.

Lastly: did you share these insights with any of your creator partners, and did that become part of your value-add when pitching future campaigns? Like, “we’ve built a vetting and briefing process that’s proven to improve UGC performance.” That’s a differentiator.