3.7X More Reach for Half the Cost: How Algorithmically Selected Creators Benefit Campaigns

Based on an analysis of 3.3M X accounts – all current followers of 60 Prediction Markets projects.

Most teams pick KOLs by follower count, quality, engagement, and impressions. We built an algorithm that picks them primarily based on audience data.

The difference: a 10-KOL lineup built on the audience size and impressions reached 1,998 people in the Prediction Markets niche at ~$10,000. Our algorithmically selected 10-KOL lineup reached 7,533 at ~$4,000. That's 3.7x more niche reach at less than half the cost. Important: in this research, we define niche reach as estimated niche followers × impression rate per account and use this metric to show the realistically reachable audience within the niche.

If you take that same $10,000 and spread it across 21 algorithmically selected KOLs, you get 12,062 niche readers. 6x more for the same money.

This report explains how we got there.

On prices: All KOL prices in this report come from public sources or market estimates. They're approximations and may have changed. Some creators don't sell sponsored posts at all – being included here says nothing about their availability.

On creators: Every creator we analyzed was selected based solely on audience data – niche concentration, impression rate, and overlap. We did not look at commercial relationships or availability. This report is not an evaluation of any creator's quality. It's about audience composition.

On the data: This reflects follower behavior during the study period (March 31st - April 2nd). Audiences can change over time.

We pulled follower lists for the 60 largest Prediction Markets accounts on X. That gave us 3.3M total followers. But many people follow more than one account. After removing duplicates, we ended up with 2.6M unique accounts.

We then filtered for people who follow at least two Prediction Markets accounts. One follow could be accidental or means being a fan of a specific product, not the niche or narratively. Two or more means you're actually interested. This gave us our core audience: 434,114 accounts.

This is the addressable audience for a Prediction Markets campaign.

We selected 10 well-known KOLs active in Prediction Markets. Selection was based on follower count, X-Score & smart followers, and the number of impressions they get on Prediction Markets content specifically.

Then we mapped their audiences against the 434k net audience.

The result: significant overlap. Several pairs share 20–39% of their niche audience. That means if you book both, you're paying for two placements but reaching the same people twice.

This isn't a problem with any individual creator. It's just how it works – people interested in a niche follow multiple accounts in that niche.

To sum up:

When you add KOLs one by one, ordered by how many new people each one brings:

This means that every KOL you add after the fourth one primarily reaches people who have already seen the campaign. That can be useful – but it should be a deliberate choice, not an accident.

Reach or frequency – you need to pick. Without overlap data, you don't even know which one you're buying.

Seeing the same message from multiple trusted sources increases conversion. Research shows buyers need 6–8 touchpoints before acting. A group of KOLs with a shared audience can deliver that, so the same people hear from multiple voices in a short period.

In our dataset, one high-overlap trio (Jaccard ≈ 0.0696) shares 326 accounts across all three. That's a conversion-optimized combination.

The choice between reach and frequency is a campaign decision. It can only be made intentionally when you know the overlap in advance.

We define the following cohorts, for the sake of brevity:

Here's something counterintuitive:

Influencers with <10k followers have 49x more niche concentration than accounts with 1M+ followers. Small creators built their followings around a single topic. Large creators accumulated followers across many topics.

Big KOLs aren't a bad choice, though. Yes, only 0.1% of 1M+ followers accounts are in the Prediction Markets niche – but 0.1% of 1M is still more niche people in absolute terms than 4.9% of 10k. So when you look at total niche people reached per post, larger accounts still come out ahead. Small accounts are more concentrated, but they're also just smaller.

The data support a layered approach:

The problem with smaller KOLs is the work: more outreach, more negotiations, more briefings, more tracking. Most teams default to a few big names because it's operationally easier – not because it's more efficient.

We compared three scenarios using real Prediction Markets KOL data. Prices are approximations.

Ten KOLs chosen by follower count, quality, engagement, and impressions, the industry standard.

Same number of placements, but selected by niche concentration, minimal overlap, and impression rate.

Take the ~$10,000 from Scenario 1 and spread it across 21 algorithmically selected KOLs. This is possible because smaller KOLs charge significantly less per post.

The algorithm picks KOLs in order: first the one with the most niche reach per dollar, then the one that adds the most new audience on top, and so on. Each new influencer is evaluated for what it adds – not for how big they are.

Reach is overestimated. Of 3.3M gross followers, only 434k are genuinely engaged with the niche, not only one protocol.

Overlap is invisible and expensive. Pairs of KOLs share up to 39% of their niche audience. Without measuring this, you're paying double (or paying more KOLs) for the same reach.

Reach maxes out early. Four KOLs cover 77% of the reachable audience. After that, you're buying frequency – which can be valuable, but should be intentional.

Optimized selection is 3.7–6x more efficient. The difference is not in how much you spend. It's in how you define the goal of your campaign.

All of this can be estimated before you spend a dollar.

This analysis focuses on Prediction Markets, but the same methodology applies to any Web3 niche on Crypto Twitter - stay tuned for upcoming reports.

Already ran a campaign? Send us the lineup – we'll map it against the audience graph and show you the overlap, the niche reach per placement, and what an optimized selection would have looked like.

Planning a campaign? Tell us the niche and goal – we'll run the full analysis: overlap mapping, relevance scoring, and algorithmic selection for your objective.

Wallchain Select built the infrastructure behind this – the audience graph, the overlap model, the selection algorithm. We handle selection, outreach, and execution, so your team doesn't have to.

Reach out via Telegram (@surapyk), and we'll take it from there.