How to vet an influencer before they make the shortlist

Creator vetting starts by defining the decision the team must make. A polished profile cannot tell you whether a creator is right for a specific market, audience, product and risk tolerance.

Write the role, audience, market, required evidence and disqualifiers before reviewing candidates. This prevents attractive but irrelevant signals from quietly changing the brief.

Separate stored profile facts from calculations, estimates and human or AI interpretation. A past brand mention may be observed; category affinity may be inferred. Those are different kinds of answers and should be reviewed differently.

For every material claim, preserve the source, observation date and method used to reach it. Confidence should come from a traceable production path, not from a confident sentence.

Apply the same research questions to every candidate. Compare audience relevance, content behavior, commercial history and risk using the same definitions and time windows.

A useful shortlist does not flatten creators into one score. It makes the trade-offs legible so the team can understand why each creator advanced.

Review the content around the creator, not only the posts selected for a pitch. Look for repeated themes, prior partnerships, unresolved conflicts, audience mismatch and behavior that could change the campaign decision.

Risk review should be proportionate to the category and market. The questions for a beauty launch, a financial product and a family brand should not be identical.

Look for recent and recurring brand relationships, direct competitors, category saturation and whether commercial content is clearly distinguishable from ordinary posts. A prior partnership is not automatically a conflict, but it may change exclusivity, credibility or timing.

For work that can affect U.S. consumers, the Federal Trade Commission says material relationships should be disclosed clearly and where people can notice them. Other markets have their own rules. Treat disclosure review as one jurisdiction-specific question for counsel or the accountable campaign owner, not a universal legal verdict.

Illustrative teaching example, not real creator results or a product benchmark. Brief: creators who repeatedly explain beginner skincare routines, with a separately verified US or Canada audience. Use the same five-video sample window for each hypothetical candidate.

Candidate A: the bio mentions skincare, but the sampled videos are unrelated to beginner routines. Record a bio-only match and reject for this content requirement. Candidate B: three sampled videos explain beginner routines, but no audience report is available. Advance for content review and hold audience approval; the market remains unknown. Candidate C: one sponsored routine appears in the sample and the remaining videos discuss unrelated topics. Request more evidence before calling the subject a recurring theme.

In an actual review, attach real source URLs and time ranges to every observation. Do not copy these hypothetical outcomes into a creator record. The evidence worksheet below the free vetting checklist can capture the requirement, sample, source, audience evidence, reviewer and decision in a downloadable CSV.

Record why the creator advanced, which evidence mattered, what remains unknown and who approved the decision. This turns one campaign review into reusable decision memory.

The goal is not a risk-free creator. It is a decision the team can explain, revisit and improve.

Define whether you are reviewing a creator before a partnership, searching a collection of saved videos, or approving a commissioned draft. For pre-partnership vetting, use the same recent-content window and campaign criteria for every candidate. Record the size of the sample and sources you could not inspect.

For each relevant moment, retain the original content URL, time range where available, match field and surrounding context. Separate a bio claim, a caption mention, spoken discussion, on-screen text and visible product use. One match can be accidental; repeated relevant discussion needs support from the sampled content.

Write the observation separately from your interpretation. For example, a product mention can support topic relevance; it does not establish purchase intent, a paid partnership or campaign results. Record advance, reject or request more evidence with a reviewer and date.

Record time spent reviewing, false matches, missed relevant content and whether a second reviewer can verify the decision from its sources. Report the sample and method beside any result. These observations can evaluate the review process; they do not prove sales lift or predictive accuracy.