Every influencer campaign starts with too many creators and not enough certainty. You've got a longlist of 50, 100, sometimes 200 KOLs that look plausible — and a client expecting you to back the handful you choose with more than "they have a big following." Cutting that longlist down to a confident shortlist is the highest-leverage work in influencer marketing, and follower counts are almost useless for it.
Here's the data-driven way to do it: use the comment section, in bulk, to filter creators on the things that actually predict campaign performance.
Why follower count is the wrong filter
Follower count tells you reach potential and nothing about quality. Two creators with 500K followers can be completely different bets:
- One has a real, engaged audience that comments, asks questions, and tags friends.
- The other has inflated followers and purchased engagement that converts to nothing.
You can't see the difference from the outside — until you read the comments. The comment section is the public, hard-to-fake signal that separates the two, and it's exactly what a shortlist should be built on.
The shortlist funnel
Think of it as three tightening stages, each cheaper to run than the manual alternative.
Stage 1 — Longlist (100 → 30): bulk export, fast filter
Pull a few top videos per creator and bulk-export every KOL's comments into one dataset. Then filter hard on cheap, obvious signals:
- Comment-to-view ratio — is anyone actually talking, or just scrolling past?
- Authenticity — duplicate phrasing, emoji-spam bursts, and generic bot text knock creators out immediately.
- On-topic conversation — are comments about the content, or unrelated noise?
This stage is about elimination. You're cutting the obviously weak creators before spending time on anyone.
Stage 2 — Mid-list (30 → 10): audience fit
Now the question shifts from "is the engagement real" to "is it your audience." For each surviving creator, run a Profile Audience analysis to estimate their commenters by age, gender, and country. Cut anyone whose real audience doesn't match your target customer — a country mismatch alone can disqualify an otherwise strong creator. This is where you find the expensive mistakes before making them.
Stage 3 — Shortlist (10 → final): depth and proof
For your finalists, go deep. Run AI Comment Analysis to pull sentiment, purchase intent, and the actual topics their audience cares about, and benchmark each creator's reach and consistency with a Profile Analysis. Now your recommendation comes with evidence: this creator, this real audience, this purchase intent — not "they felt right."
Why this wins you the business
Two things separate agencies that grow from ones that churn clients:
- You pick better creators. Filtering on engagement quality and audience fit beats follower-count selection on campaign performance, every time.
- You can prove your reasoning. When a client asks "why these ten?", you show comment data, authenticity flags, and audience breakdowns instead of a vibe. That's what justifies your fee and renews the retainer.
Running it as a repeatable process
Standardize the funnel so any analyst on your team produces comparable shortlists:
- Fix the inputs — same number of videos per creator, same sample size for audience analysis (500+ commenters for finalists).
- Keep every export and report on file, so post-campaign you can reconcile the audience you picked against the audience you reached.
- Because everything is pay/credit per use, your selection cost scales with how many creators you actually evaluate — not a flat platform subscription.
For the broader agency playbook, see our complete guide to bulk export and AI analysis for influencer agencies.
Frequently asked questions
How many creators should a longlist have?
Whatever your sourcing produces — the point of the funnel is that bulk export makes a 100-creator longlist no harder to process than a 10-creator one, so you don't have to pre-filter on weak signals like follower count.
What's the single best filter for stage 1?
Authenticity. A creator with impressive numbers but obviously purchased comments should be cut before any deeper analysis — fake engagement poisons every metric downstream.
How do I judge audience fit objectively?
Run Profile Audience on each creator and compare the estimated age, gender, and country distribution against your campaign's target customer. Large, consistent mismatches are disqualifiers.
Do I need accounts or subscriptions to do this?
No subscription is required to start — export and audience tools are pay/credit per use, so you can run a full shortlist process and only pay for the creators you actually evaluate.
Build your next shortlist on evidence
Stop defending creator picks with follower counts. Bulk-export your longlist's comments, filter on real engagement, check audience fit with Profile Audience, and hand your client a shortlist they can see the reasoning behind.