I almost made a huge mistake last month. I was about to lock in a campaign with an influencer who had 200K followers, solid engagement metrics, and seemed credible on paper. But something felt off, so I did some digging.
Turns out, about 40% of their followers were bots or inactive accounts. Their recent engagement was artificially inflated through engagement pods. And their previous brand partnerships? Barely any actual conversions, but they’d been paid anyway.
This could have tanked our brand’s credibility if we’d launched without catching it. So I’ve been reckoning with: How do you actually spot fraud before it becomes a problem?
Here’s what I’ve learned by digging into my mistakes and successes:
Obvious red flags:
- Sudden follower spikes (buybot activity)
- Comments filled with random emoji or generic praise
- Geographic mismatch (50K followers but 80% from a country the creator doesn’t target)
- Engagement drops sharply in recent posts
- Zero verifiable brand partnerships or PR mentions
Subtler red flags:
- Engagement rate suddenly too consistent across all posts (weird for real accounts)
- Comments from people with no followers or suspicious profiles
- Heavy reliance on hashtags but low organic reach
- Claims of “influencer status” but no actual conversion evidence
- Pricing that’s way below or way above market for their size
What actually works to verify:
- Check historical posts (6+ months back) for consistency
- Look at comment quality, not just volume
- Cross-reference with other platforms
- Ask for proof of past brand partnerships
- Run a micro-test campaign with lower spend
- Talk to other brands who’ve worked with them
But here’s my real challenge: I can spot obvious fraud, but I miss the middle ground. Creators who aren’t totally fake, but have inflated engagement or exaggerate results. How do you calibrate your fraud detection to catch those?
And how much time should you actually spend on this? I’ve got 20+ creators to vet for this quarter. I can’t audit each one for 3 hours.
What’s your process? Are you using tools, manual checks, or a combination? And what’s actually changed your ability to spot fraud?
Ой, это такая проблема! Я даву контакты людей, которых я уже проверила. Репутация—это мой капитал, поэтому я беру это серьёзно.
Одна вещь, которую я сделала—я начала собирать отзывы от других маркетологов о криэйторах. Вроде, неформальной базы репутации. Если кто-то говорит “Этот креатор мошенник” или “Обещал результаты и не доставил”—я это помню.
Мне кажется, в сообществе было бы полезно иметь канал, где люди могут анонимно делиться этой информацией. “Будьте осторожны с человеком X” без доказательств, просто как warning. Это помогло бы всем.
Я разработала скоринг-систему для этого. Может, она полезна:
Tier 1 Red Flags (Critical):
- Follower spike > 30% в месяц
- Engagement rate < 0.3% (для реальных аккаунтов)
-
80% фолловеров из одной страны при глобальном позиционировании
- 0 верифицируемых партнёрств
Tier 2 Red Flags (Caution):
- Engagement rate > 15% (скорее всего, они используют боты)
- Комментарии повторяются или кажутся генерированными
- Резко упала активность в последние 3 месяца
- Цена за пост не совпадает с примерным рынком (слишком低 или слишком高)
Tier 3 Yellow Flags (Investigate):
- Большая разница между ER на разных постах
- Некоторые посты получили заметно больше ER после публикации
- Профиль выглядит новым для своего размера
Мой процесс:
- Автоматическая проверка Tier 1—если есть, отсеиваю
- Ручную проверку Tier 2 и 3
- Разговор с креатором о результатах
- Если что-то смущает—микро-тест на 5% от основного бюджета
По точности: я ловлю ~95% явных фейков и ~60% скрытых 
У меня было два случая, когда я ошибся.
В первом случае—я верил цифрам. 100K фолловеров выглядело хорошо. Я не проверил комментарии. Половина из них была на каком-то восточноевропейском языке от аккаунтов без картинок.
Во втором случае—я видел, что креатор работал с известной компанией. Я предположил, что они check сделали. Они не сделали. Оказалось, что компания не знала о его фейковых метриках.
Теперь я:
- Точечная проверка: случайно выбираю 50 комментариев и смотрю на них вручную
- Проверка истории: смотрю его посты на 6-12 месяцев назад
- Проверка других платформ: если в TikTok 100K реальные, то в Instagram должно быть пропорционально
- Разговор лично: я звоню и расспрашиваю о его бизнесе, откуда ноги растут, что за стратегия
Это занимает двадцать минут на человека. Но уходит мне гораздо меньше, чем потом разбираться с неудачной кампанией.
Here’s my system:
Fast Track (5 minutes):
- Check follower growth curve (HypeAuditor or similar)
- Spot sample 30 comments (are they real people?)
- Compare engagement rates across last 20 posts
If any red flag, move to next tier. If clean, move to contract.
Deep Dive (30 minutes):
- Pull detailed audience breakdown
- Check audience overlap with similar creators (bot networks overlap)
- Cross-reference with Instagram, TikTok, YouTube (metrics should be proportional)
- Ask for recent brand partnership receipts
Micro Test (only for questionable cases):
- $2-3K campaign
- Track actual conversions
- If it underperforms by >50% compared to prediction, they’re likely inflated
I skip the micro test for obvious green accounts (established, clean metrics, verified partnerships). Not worth the time.
The key insight? Fraud isn’t random. Fraudsters follow patterns. Once you’ve seen a few, you spot them faster.
I’ve now rejected about 200 creators over 2 years for fraud. I can usually tell in the first 5 minutes now.
One more thing—pricing is a huge signal. If someone with 50K followers is charging the same as someone with 500K, that’s fraud. They boosted followers. Real creators price proportionally to reach.
Also—I’ve never regretted turning down a creator due to fraud concerns. But I’ve definitely regretted not being cautious enough. The cost of one bad partner is way higher than the cost of being thorough upfront.
From my side, I see creators actively boosting followers all the time. And honestly? It makes it harder for legit creators like me. Brands get burned by fake influencers and then they’re suspicious of everyone.
What I wish brands knew: Real engagement is variable. Some of my posts get 8% engagement, some get 2%. That’s normal. If someone has 12% engagement on every single post, that’s fake.
Also—I’ll always provide proof of past results. If a creator won’t share that, they have something to hide.
I get annoyed when brands are too skeptical, but honestly? Better to be thorough. I’d rather spend 30 minutes proving I’m legit than work with a brand, underdeliver secretly, and tank my reputation.
Real talk: I’ve been approached by services promising fake growth. I turned them all down because I built my following for real and I’m not throwing that away. My engagement is lower than it could be, but it’s real.
Brands who take time to vet me properly? Those are the people I work hardest for. There’s a mutual respect there.
I’ve spent a lot of time on this problem. Here’s my framework:
Automated Layer (using tools):
- Follower growth velocity (>25% YoY is suspect)
- Engagement rate distribution (real accounts vary, fake accounts are consistent)
- Audience demographic overlap with similar creators (botnet signal)
- Comment sentiment analysis (bots don’t vary in tone)
Manual Layer:
- Spot check 100 random comments (read them, not just count them)
- Historical content deep dive (6-12 months vs. recent)
- Cross-platform consistency (followers proportional across all platforms)
- Track record verification (actual brand partnerships, not just claims)
Field Test:
- For high-risk but potentially valuable creators, 10% budget test
- Track conversion, not just clicks
- Compare actual performance to their claims
Database Building:
- I track all creators I’ve worked with and their performance
- Over time, I know which creators deliver, which inflate, which are honest
- This becomes my best signal
Key insight I’ve learned: Fraud is usually not malicious fraud. It’s creators who bought followers early, then realized they’re stuck with inflated metrics. They feel committed to the lie. Some will inflate results to justify the deception.
The ones I trust most? Transparent about limitations. They’ll say, “My audience is mostly US-based, so if you’re targeting EU, this won’t convert as well.” That’s honesty.
The suspicious ones? They oversell. Every metric is perfect. No creator is perfect.
I’ve probably caught 95+ cases of inflated creators by now. About 60% were accidentally inflated (they bought followers years ago and thought it was fine). About 40% were intentional fraud. Either way, I don’t work with them.
One last thought: Community is your best fraud detector. If you work with 50 creators, you know which ones are real. That knowledge compounds. I’d recommend building relationships, not just transacting. Over time, you’ll know who to trust.
