Once a brand has a batch of candidates, influencer screening tends to turn into a spreadsheet contest: follower counts, average likes, and quoted rates all lined up neatly — and the people left at the end aren't necessarily right for this particular campaign. A list tells you who's available. A shortlist has to answer why this person, where the evidence came from, and what risks still need checking.
The screening logic is the same on Xiaohongshu (RedNote) and Douyin: filter in bulk first, break down the few accounts that survive, then validate against the comments.
The goal is a shortlist someone else can check
A shortlist ready for outreach has to let marketing, media, and content teams read the same story: how much relevant content the candidate has actually published, whether recent performance is steady, what content method they rely on, what the comment sections say back, and how all of that relates to this brief.
That's also why follower count can't carry the decision on its own. It's a useful proxy for account size, but it says nothing about whether a creator has sustained a point of view in your specific category, and it can't substitute for the public feedback users leave on the content.
Finding creators and screening creators are separate jobs, too. The first one answers where candidates come from — see finding benchmark accounts on Xiaohongshu and Douyin from category content. The second starts from a list you already have, and narrows it down to a small number of choices you can defend.
Filter in bulk first, then break down the few that are left
When the list is long, researching every candidate in depth spends your time on people you'll never work with. A steadier sequence:
- Test-run 5–10 candidate profiles first, to confirm your category keywords, what counts as relevant content, and whether the basic filters work.
- Bulk-collect public profile info and post lists, and cut the accounts with no relevant content or a recent direction that clearly doesn't fit.
- Only break down the small number of candidates still standing — comparing recurring topics, the gap between their best and worst performers, and how they express an idea.
- Open their representative posts and collect comments for validation, noting the real follow-up questions, reservations, and any unusual engagement patterns.
Use an account breakdown to see the content method
Once a candidate clears the first pass, move into Xiaohongshu account analysis or Douyin account analysis. The public AI workspace handles account research and single-post breakdowns, and keeps the findings and topic ideas in one place. When you read the report, don't just pull out an average — what's worth comparing is the difference between their consistent themes, their standout posts, and their weakest ones.
