Build Your Own TikTok Creator Library (Discovery Plans + Similar Creators)

The one-off-search trap

Most people open KOLens, type a keyword, get 200 creators back, and treat that result as the answer. It is not the answer. It is the first drop.

The mental model mistake is treating a KOL discovery tool like Google when it should be treated like a CRM. Google gives you the page you asked for and then you leave. A CRM is something you fill over time, and the value lives in what you have accumulated, not in any single query.

If you keep running fresh one-off searches, three predictable complaints follow. The first is that 200 creators is not enough for a real campaign roster. The second is that your filters — country, niche, audience country, recent views — are so narrow that almost no one survives a single pass. The third is that managed agencies seem to have deeper rosters than you do, because they have been quietly submitting creators into their own database for months or years.

All three complaints are the same complaint: you are searching when you should be collecting. The fix is not a bigger search. It is a workflow that compounds.

The three levers that compound

KOLens has exactly three tools that turn a single search into a growing library. None of them is hidden. Most users just never connect them.

1. Discovery Plans — recurring scrapes that fill themselves

A Discovery Plan is a subscription-style scrape. You configure a keyword, the platforms you care about, and the filters you would normally apply manually — country, niche, follower range, audience country. KOLens then re-runs that discovery every day and adds new creators to your workspace, deduplicated against what you already have.

The numbers matter here. A single search caps at 200 creators. A Discovery Plan running daily for 30 days does not give you 30 times 200 — most of those overlap — but in practice it surfaces 400 to 800 net-new creators per month per active plan, because TikTok is producing fresh content and new accounts are matching your keyword every day.

This is the lever that turns the 200-creator ceiling from a complaint into a non-issue. The ceiling never moved, but the floor keeps rising.

2. Find similar creators — lateral expansion

Discovery Plans answer the question 'who matches this keyword today'. They do not answer the question 'who is adjacent to my best creator and would never have shown up under my keyword in the first place'.

That is what find_similar_creators is for. You pick a creator you already trust — someone who posted a strong organic mention, someone who converted in a past campaign, someone with the audience country you actually want — and KOLens returns others with overlapping audience and content profile.

This matters most when your filters are narrow. If you require US-only audience plus a sub-niche like 'minimalist skincare for sensitive skin', keyword search saturates fast — there are only so many people posting under that exact phrase. Lateral expansion from one good anchor pulls in creators who never used your keyword but share the audience.

3. Watchlist + Lists — your library itself

Discovery Plans and similar-creator expansion produce raw flow. The library is what you do with that flow.

Watchlist is for monitoring. You add creators you are interested in but not yet ready to contact, and KOLens tracks growth, engagement shifts, and silent-quit signals over time. The result is a change digest instead of a fresh search every week. (We wrote a full piece on this in Watchlist monitoring signals.)

KOL Lists are the pipeline itself: pending, contacted, replied, negotiating, signed, rejected. This is where the library converts into deals. The account-role filter quietly does work here too — it separates real creators from brand and reseller accounts so your pipeline does not bloat with non-creator entries.

A 90-day playbook from 200 to 2,000 creators

This is the timeline we see from teams that actually run the workflow. Numbers are typical, not guaranteed — niche depth varies.

1. 1

   Week 1: Seed with three to five searches

   Run three to five keyword searches across your top niches. Expect roughly 200 creators each, so 600 to 1,000 raw entries. Apply the account-role filter, drop anything that is obviously a reseller, and import the survivors into a single KOL List called something like 'cohort-q2-raw'. This is your starting drop and it is the only manual day in the workflow.
2. 2

   Weeks 2-4: Turn the best searches into Discovery Plans

   For each keyword that produced strong results in week one, create a Discovery Plan with the same filters. Set the schedule to daily. By the end of week four you should have three to five active plans, each adding 100 to 200 net-new creators per week. Library size at this point is usually 1,200 to 1,600.
3. 3

   Months 2-3: Lateral expansion and Watchlist

   Identify the 20 to 30 strongest creators from the library so far — the ones who responded well, or whose audience matches your target country tightly. Run find_similar_creators on each. Push promising results into Watchlist instead of straight into the contact pipeline; you want to see 30 days of growth signal before reaching out. By day 90 the combined library is typically 2,000 to 2,500, with 200 to 400 of those actively monitored and 50 to 150 in active outreach.

The compounding shape is what matters. Week one feels like one-off search. Week four feels like an automated pipeline. Month three feels like an asset.

The agency comparison — what KOLens actually replaces

This section is for the cross-border (出海) audience in particular, because the comparison to MCN agencies comes up in almost every sales conversation. We will not pretend agencies are worse than they are.

Agencies are more accurate per-creator. They have humans who watch videos, verify the audience, check past brand deals, and decide whether to submit. That work produces a tighter list — typically 50 to 100 submissions per month per account manager. For a single high-budget campaign with five anchor creators, an agency roster is often the right answer.

What agencies are not is scalable. The human bottleneck is real. You also do not own the library — when the contract ends, the database goes with them, and you start over with the next agency. Cost-per-creator-discovered usually lands somewhere between $50 and $200 in agency models, versus pennies in self-serve scrape models.

The honest pitch is this: if you have $10K a month and want five vetted creators, hire the agency. If you have $300 a month and want two thousand creators you actually own, build the library yourself in KOLens. Most cross-border sellers we talk to need the second one, because the campaigns they run are scale plays — fifty to two hundred creators per launch, not five.

When narrow filters kill your funnel — and what to do

The complaint behind 'almost no one survives my filters' is usually a funnel design problem, not a filter problem.

The wrong shape is narrow entry and narrow exit. If your keyword is already specific ('US sensitive-skin minimalist skincare creators with 50K to 200K followers and 70% female audience'), and you apply every filter at the search step, the 200 cap shrinks to 8. There is nowhere for the funnel to do work.

The right shape for library-building is wide entry and narrow exit. Use a broader keyword at the search step — 'skincare review' or 'sensitive skin' — and let the 200-creator pool fill. Run a Discovery Plan on that broader keyword so the pool keeps growing. Then apply your tight filters at the Watchlist and List stage, not at discovery. The audience-country snapshot is a Watchlist-stage filter, not a search-stage one — it works much better after a creator has 30 days of data in your workspace.

The reframe is that the filters are not too narrow. They are applied at the wrong step. Search should be permissive; the library should be selective.

Why ownership of the library is the real moat

The agency comparison usually ends on accuracy, but the more important axis is ownership. An agency-built roster lives in the agency's CRM. A KOLens library lives in your workspace, exports to CSV, and is queryable through our MCP for any agent or BI tool you connect.

This matters most for cross-border sellers who run multi-year programs. The creators who ignored you in 2024 might be perfect in 2026 — bigger audience, better fit, slower growth that finally matches your price point. If you have the history in your own library, you can re-approach them. If the history lives in an agency database you no longer pay for, that signal is gone.

Library ownership also lets you experiment with positioning. The same 2,000 creators can be re-sliced for a different product launch, a different country, or a different campaign type. A managed roster gets reused for the brand that paid for it; a self-serve library gets reused indefinitely.

What to do next

If you are a cross-border seller comparing KOLens to an MCN, run the workflow side by side for 60 days. Pay the agency for their 100 submissions; pay KOLens for one Discovery Plan and a similar-creator expansion budget. At day 60 compare library size, cost-per-creator-in-pipeline, and reply rate. We will not pretend the answer is universal — for some niches the agency wins on quality. For most, the scale and ownership difference is large enough that the decision is obvious.

If you are an SMM agency, the same logic applies inverted: KOLens becomes the engine that feeds your human submitters. Your team stops being a discovery bottleneck and starts being a vetting layer on top of an already-large library.

If you are a DTC brand running one campaign per quarter, start with a single Discovery Plan on your strongest keyword, two find_similar_creators expansions per month from your best performers, and let the library compound quietly between launches. The next campaign will start from a 500-creator base instead of a 200-creator search. That is the entire pitch.

For pricing details see /pricing — plans start at $29/mo Starter and credits never expire. To run a first live search and feel the workflow, start at /find-tiktok-influencers. For the deeper companion piece on automation see automated TikTok KOL Discovery Plans, and for the cross-border lens see TikTok influencer marketing for cross-border sellers.

Frequently asked

Why do I only get 200 creators per search?

200 is the per-run ceiling for one keyword on one platform — it is enough to test a niche but not enough to staff a campaign. The product is designed around running many searches over time, not one giant search. Discovery Plans automate that repetition for you so the library grows in the background.

How is a Discovery Plan different from just re-running a search?

A Discovery Plan is a subscription-style automation: you pick the keyword, filters, and platforms once, and KOLens re-runs the discovery on a recurring daily schedule. New creators are deduplicated against your workspace and added automatically, so you spend zero clicks per day to keep the library current.

What does find_similar_creators actually look at?

It takes a creator you already trust and returns others with overlapping audience country, content profile, and engagement shape. It is lateral expansion from a known-good anchor — useful when keyword search has saturated and you want adjacent niches without inventing new keywords.

Can an MCN agency really build a more accurate list than KOLens?

Per-creator, yes — a human submitter can verify niche fit in ways an algorithm cannot. But agencies submit on the order of 50 to 100 creators per month, while a single Discovery Plan plus similar-creator expansion can add hundreds per week. The honest tradeoff is human accuracy versus automated scale plus ownership of the data.

Does the library work for non-TikTok platforms?

Yes. TikTok is the primary platform, but Discovery Plans also run on Douyin, Instagram, X, and Xiaohongshu, with YouTube available behind a feature flag. Cross-border sellers usually run TikTok plus one other platform per market — for example TikTok plus Xiaohongshu for a US-to-China lane.