TattvaDocs
Use casesPerformance marketing

Creator size mix

Mega vs Micro vs Nano — which creator size is delivering for which product? A one-off finding turned into an ongoing view.

Planned

This use case needs definitions that aren't in the Library yet (NRoAS). Until your data analyst adds them, Tattva can't answer it as shown here.

The finding: "Mega creators work better than Micro, and Micro better than Nano. Now we want to check that across more products and a longer look-back."

Many teams tag creators by size by hand and find a pattern like Mega outperforms Micro, which outperforms Nano for the products they studied. Tattva turns that one-off analysis into a question you can re-ask as the data grows.

Who asks

  • The category or creative head, briefing creators
  • The performance-marketing lead, planning next quarter's creator budget
  • Influencer operations, deciding which tier to recruit

Frequency

Monthly for the portfolio review; as needed when planning a new product's creator strategy.

Data you need

  • ad_creatives linked to creators
  • Creator metadata: follower band, engagement rate, average views, growth rate
  • Size tag: Mega / Macro / Micro / Nano (by hand or from a tagging tool)
  • NRoAS per product, attributed back to the creator

Common gap: creator size often wasn't tracked historically. Older creatives may show as untagged; coverage improves as tagging continues.

Size bands

Typical bands:

BandFollowers
Mega> 1M
Macro100K – 1M
Micro10K – 100K
Nano< 10K

Put your own bands in the question if they differ, or ask your data analyst to record them as a convention so every answer uses the same ones. For a niche product, a 50K-follower specialist may behave like a Mega creator.

How to ask it

  • "For each product, compare NRoAS by creator size band over the last 90 days. Has the ranking changed?"
  • "Are Mega creators still outperforming Micro for SKU-A? Show the cohort retention, not just NRoAS."
  • "Which products show the opposite pattern, where Micro or Nano beats Mega?"
  • "For products launched in the last six months, which creator size paid back fastest?"

What you'll get back

A size scorecard per product (figures made up):

ProductMega NRoASMacro NRoASMicro NRoASNano NRoASSuggested mix
SKU-A3.22.41.61.160% Mega, 30% Macro, 10% Micro
SKU-B2.82.52.31.4More balanced; Micro is competitive
SKU-C2.12.62.01.0Macro leads — product or persona effect?

Ask for the sample size per band. When a band has only a handful of creators, treat that cell as unreliable.

Beyond NRoAS

Follower count isn't the whole story. Ask about four measures side by side:

  1. NRoAS: the headline.
  2. Hook rate: how many viewers don't scroll past the first three seconds.
  3. Hold rate: how many watch to the end or nearly.
  4. Cohort retention at month 3 / month 6: do customers acquired through this size of creator come back?

A creator with high hook and hold rates but an audience that doesn't buy produces engagement, not sales. Asking for all four separates the two.

How to make it recurring

Trackers run hourly, daily or weekly; there's no monthly cadence. Two options:

  • Weekly tracker: click Save as tracker on the answer, then Edit the question and pick weekly. Set the Classification rule to Relative delta on your NRoAS metric with critical if Δ > 15, so a large swing is flagged. The rule doesn't detect a change in which band leads; read the run's narrative for that. Add the category team's Slack channel on the Delivery tab.
  • Monthly report: ask the question when the month closes and click Save as report. Share it from the report's Export menu.

Pitfalls

  • Treating size as the whole story. A Mega creator with the wrong audience underperforms a well-matched Micro. Combine size with niche and persona.
  • Out-of-date bands. Creators grow; a Macro from six months ago may be Mega now. Refresh the tags.
  • Confusing size with pattern. If Mega creators happened to use this month's winning format, size gets the credit for the format. Ask for the joint view when a result surprises you.
  • Too few Nano creators. Small bands give unreliable results; check the sample size.
Creator size mix — Tattva Docs