How to vet an influencer before you propose them to a client
Classification criteria, real audience affinity, risk signals, and a simple matrix for putting the decision in writing.

How your agency selects influencers for a client with real data
For an agency or a freelancer, choosing influencers by follower count is the fastest way to burn through a client's budget. Useful selection combines audience, engagement, industry, content quality, risk, and the ability to drive results.
The goal isn't to hand the client famous profiles, but profiles that genuinely fit their campaign.
Classification criteria
Audience size
Nano, micro, mid, macro, or celebrity. Size matters, but it's no substitute for affinity or quality.
Useful engagement
Don't just look at the percentage. Review the quality of comments, saves, clicks, replies, and signals of intent.
Audience affinity
Check country, language, age, interests, and relationship to the category.
Content quality
Evaluate the script, editing, clarity, consistency, and ability to integrate a brand without sounding forced.
Risk
Review their history of controversies, purchased followers, oversaturation with collaborations, and legal compliance.
Rights and reuse
Define whether the brand will be able to reuse the content, for how long, and on which channels.
How to manage it during the campaign
Each selected profile should have a brief, deliverables, a schedule, review, approval, tracking, and reporting. Plus one rule almost nobody writes down: where the creator's deliverables land. If they end up in an expired transfer link, you've paid for content you can no longer use.
Recommended KPIs
- cost per creator,
- qualified reach,
- useful engagement,
- CTR,
- conversions,
- CAC,
- subsequent use of the content,
- comment quality.
Simple decision matrix
You can classify profiles into four groups:
- high affinity and high performance: prioritize,
- high affinity and low performance: test with a controlled budget,
- low affinity and high reach: use only for very specific awareness,
- low affinity and low performance: discard.
This matrix helps avoid emotional decisions. The team may really like a profile, but if it doesn't fit the audience, the message, and the objective, it shouldn't lead the campaign.
Size does not make the decision by itself. Compare profiles using a recent sample and consistent definitions of audience, views, cost, and conversion. The guide to interpreting influencer metrics explains how to avoid viral outliers and universal percentage rules. Then document deliverables, disclosure, exclusivity, and rights in the campaign brief.
You already paid for that content: use it more than once
The most expensive line in a creator campaign is the material, and it's also the worst-kept. You negotiate twelve months of reuse rights and three months in nobody can find the original video — all that's left is the post on the creator's profile. Paid content worth zero.
Polimake indexes whatever you upload: the AI generates the keywords as it ingests them — the literal content plus tone, concept, and context — and then you search in plain language down to the exact frame inside a video, so you find "unboxing with the blue box" without opening folder after folder. On top of that runs an AI layer that edits and produces from your own archive, today connected over MCP to ChatGPT and Claude, so you can cut new pieces from what you've already paid for.
If your creator campaigns evaporate the moment they end, park the deliverables in file management and try natural-language search on the last one.