Why portfolio planning starts before outreach
A campaign brief might say “find 10 creators,” but a raw list of 10 profiles rarely produces 10 viable conversations. Some creators have no recorded TikTok Shop GMV, some are a poor category fit, and some will not respond or accept the proposed collaboration. Our dataset can quantify only the first of those filters: whether a creator shows a recorded commerce footprint in the snapshot.
That limited measure is still useful. It prevents teams from treating every follower-qualified profile as equally ready for a sales-oriented campaign. Across the July 2026 snapshot, active-seller rates decline as follower tier increases: 54.1% for nano, 49.4% for micro, 38.8% for mid-tier, and 28.0% for macro creators.
Expected active sellers at common shortlist sizes
The table below applies each tier's observed active-seller rate to three common starting-pool sizes. Values are mathematical expectations—pool size multiplied by the observed rate—not guarantees for any individual shortlist.
| Tier | Observed active rate | From 10 profiles | From 25 profiles | From 50 profiles |
|---|---|---|---|---|
| Nano (<10K) | 54.1% | 5.4 | 13.5 | 27.1 |
| Micro (10K–100K) | 49.4% | 4.9 | 12.4 | 24.7 |
| Mid (100K–500K) | 38.8% | 3.9 | 9.7 | 19.4 |
| Macro (500K+) | 28.0% | 2.8 | 7.0 | 14.0 |
This does not mean nano creators are automatically better. It means a follower-only macro list is more likely to contain profiles without recorded commerce activity. Macro creators still show the highest absolute average GMV among active sellers, while nano creators lead on GMV efficiency per 1,000 followers. Those are different campaign objectives, so the right tier mix depends on whether you prioritize broad reach, commerce efficiency, creative volume, or a blend.
How large should the sourcing pool be?
Dividing the desired number of active sellers by the observed active rate gives a simple first-pass sourcing requirement:
Round upward, not downward. Then add a separate operational buffer for category relevance, contactability, response, pricing and product acceptance. We do not publish a universal multiplier for those later stages because our dataset does not reliably measure response or acceptance rates. Inventing one would make the model look more precise than the evidence allows.
A practical mixed-tier starting portfolio
For a conversion-oriented beauty campaign that still needs some reach, a reasonable research workflow is to start with a larger nano/micro pool and a smaller mid-tier layer:
- Source 25 nano profiles. The category average suggests about 13–14 will show recorded GMV.
- Source 15 micro profiles. The category average suggests about 7–8 will show recorded GMV.
- Source 5 mid-tier profiles. The category average suggests about 2 will show recorded GMV.
- Verify every profile individually. Check current GMV, category overlap, recent content, contact details and partnership status before outreach.
- Do not sum average GMV into a revenue promise. Tier averages describe a highly skewed historical snapshot; individual outcomes vary by orders of magnitude.
This 45-profile research pool would be expected to contain roughly 23 commercially active profiles before the qualitative and outreach filters. It is not a prescription for every brand. It is a transparent way to turn category-level evidence into a starting workload.
What the model does not tell you
Recorded GMV is not the same as audience fit, content quality, brand safety, response probability or incremental sales caused by a future campaign. The snapshot also does not provide reliable flat-fee rates, commission terms, city-level location, audience demographics or brand-collaboration history. Use this model to size the top of the screening funnel, then apply the qualitative checks in our 7-step screening framework.
For the underlying tier counts and both tier-wide and active-only GMV measures, use the benchmark explorer. For budget framing after a shortlist is built, continue with the campaign budget guide.