Influencer audience overlap analysis: how to deduplicate reach across a shortlist

To analyse influencer audience overlap, rate every pair on your shortlist for shared audience using three tests you can run by hand (commenter overlap, geography and language, and shared collaborations), count high-overlap pairs as one audience in your reach forecast, and estimate unique reach with a haircut that you set and label. The aim is simple: stop paying twice for the same people.

Our guide to how many influencers you need covers reach and budget, and the influencer vetting checklist scores creators one at a time. This post is a portfolio check that runs after scoring, on the shortlist as a whole. It matters most in the large micro-influencer plans described in how to run a micro-influencer campaign, and it sits under our complete guide to influencer marketing in India.

Why audiences overlap

Overlap is a property of pairs, not of one creator, so it never shows up in a single creator's insights. That is why vetting alone misses it.

Test 1: commenter overlap

For each creator, list the handles of commenters on their last five posts, at least 100 handles, using the method in influencer comment quality analysis. For each pair, count handles that appear in both lists and divide by the smaller list.

Commenters are a small and loyal slice of an audience, so shared commenters understate overlap. Treat this test as a lower bound: shared commenters prove overlap, and a lack of them does not prove independence. We found no published thresholds for what counts as high, so rank the pairs on your own list: the top third by shared commenters are your high pairs.

Test 2: geography and language

Take the location and age data from the creators' insights, as in how to check influencer audience location. Two creators whose audiences are mostly women of 18 to 24 in Delhi NCR who read Hinglish will share people even if no commenter appears twice.

Test 3: shared collaborations

Open each creator's tagged tab and recent collaboration posts. Note any co-created Reel or collaboration post in the last six months, any giveaway they co-hosted in the last year and any shared event. Ask each creator which other creators on your list they work with. A yes counts as a collaboration flag.

What to ask the creators

The three tests use public traces, and a short written ask fills the gaps. Ask each creator on the shortlist the same questions:

1. Which other creators on this list, if you can see it, do you post with or co-host giveaways with?
2. Have you run paid promotion to gain followers in the last six months, and which cities and languages did you target?
3. Do you share a manager, agency or content group with other creators in this niche?
4. Do you have a second account, or the same content on another platform, and are its viewers the same people?

Send the question about the other creators only after you have a short list, and share names carefully: a creator may not want a rival to know they were shortlisted. Creators with nothing to hide usually answer in a line, and a creator who says "none" while the tagged tab shows three co-posts has told you something about reliability. Our outreach message templates show how to word a data request politely.

Build the overlap matrix

Rate each pair on one rule set, and write the rule down so the next person applies it the same way.

Then work down your shortlist in order of serviceable views, largest first. Count the first creator in full. For each later creator, find the highest rating they have against anyone already counted and apply a haircut for that rating.

Estimate unique reach, and show the range

The haircuts are assumptions, not measurements. Label them that way and show three cases: none, a base case and a pessimistic case. A usable starting set is high 50% and medium 25% for the base case, high 75% and medium 50% for the pessimistic one, and low at zero in both. Change them if your own campaigns give you evidence.

Treat views as a proxy for reach. Views include repeat plays, so use reached accounts where creators share them.

If you know the size of the pool you are targeting, for example from an ad platform's audience-size estimate for the same targeting, a simple formula gives a ceiling. With a pool of P people and creators who reach r1, r2 and so on, unique reach under random overlap is P × (1 − (1 − r1/P) × (1 − r2/P) × …). Two creators who each reach 40,000 people in a pool of 100,000 give 100,000 × (1 − 0.6 × 0.6) = 64,000, not 80,000, even if their audiences were unrelated. Real audiences cluster by city, language and niche, so the true figure is lower.

Plan overlap on purpose

Overlap is waste only when it is accidental. Two creators with a shared audience, posting a few days apart, work as repeat exposure before a sale or launch. That is our own planning logic, not a published result, so test it. For reach goals, the opposite applies: spread creators across clusters and stagger posts so the same person is not served three times in a day.

Test overlap with your own audience as well. A creator whose followers already follow your brand adds little awareness, so sample their commenters and check how many already engage with your account.

Mistakes that overstate reach

• Adding a creator's Instagram and YouTube audiences together. One person can be both.
• Treating followers as reach. Use views or reached accounts.
• Counting a creator's second account as a new audience.
• Ignoring meme pages that repost each other's content.
• Forgetting your own followers.
• Using the same haircut for every campaign. Review it after each campaign, with results in a campaign learnings library.

Where overlap sits in the Gully Score

The Gully Score rates one creator at a time, so overlap carries no points. It acts after scoring, in the creator mix of the plan, where tiers, niches, languages and cities are chosen to work together. Its inputs come from the sheet: A1 (audience cities and countries) and A3 (reach per post) in audience authenticity, and C1 (overlap with your buyer) in buyer fit. Those checks give you the serviceable views used above.

Overlap analysis pays off when the list is long, because small-creator plans share audiences faster than macro ones. The creator mix in our plans spreads tiers, niches, languages and cities, and our micro and nano influencer marketing service is where long lists are built. Send us a brief and the plan returns within 48 hours of the brief call.