OpenKol
Measurement, not predictions

Did that KOL post
actually do anything?

Paste a tweet. See the volume it brought in, whether the engagement was real, and how it compares to every call we've measured.

Free — no account for your first checks

Or look at one we already ran
-1%92%184%POST−90m+2h30
$BRODIE
robinhood
+74%
Check 1 of 3
quiet audience · real money
@iruletrenches · 28,783 followers
25 Jul 2026 · 14:10 UTC
Impact
14.8/40
Performance
39.4/60
Audience
36/40
Percentile
••/100
Check detail
Profile
74%
Replies
Verdict
Small, quiet, and effective.
Per 1k
$3,018 · $86.9K in 1h
Evidence
Almost no visible engagement, but $86.9K traded in the first hour. Top decile of everything we've measured.
What the data says

Follower count is close to
meaningless

Across everything we've measured, the audience that actually trades has almost nothing to do with audience size.

165,640 followers
$191
volume brought per 1,000 followers
loud, expensive, inert
28,783 followers
$3,018
volume brought per 1,000 followers
quiet, cheap, effective

A 15x gap in volume brought per follower — completely invisible if you're buying on reach.

What the corpus says
Median impact across every scored call is 8.6 out of 40. Most KOL posts do nothing.
measured across the corpus
A 28.8K-follower account brought $3,018 per thousand followers. A 165.6K account brought $191.
same week, same chain
Audience quality does not predict returns. We tested it and the correlation is 0.05.
Spearman rank correlation
KOL skill does not persist. First half against second half of the same account's history: −0.02.
split-half test
Reply-content scoring beat account-stat scoring at 91% against hand-labelled posts.
blind evaluation
Roughly 31% of a KOL's token posts are not calls at all.
intent classification
Four answers, never blended

One number would be a lie

Audience quality has no measurable relationship with returns — we tested it, and there isn't one. Blending them into a single score would imply a link the data denies, so we keep them apart.

impact
POST−90m+2h30
$86.9K traded in the first hour

Did it move the market?

Volume and distinct traders brought in against the pair's own baseline, with attribution decaying by how soon the move arrived. A peak fourteen hours later is not the post's doing.

forproject teams · KOLs
performance
0%92%184%
+184% peak against +2.6% for peers

Was it a good call to follow?

Peak, hold and drawdown measured in excess of peer tokens on the same chain — so a market-wide melt-up never reads as skill.

fortraders
audience
REPLY CLASSIFICATION74%21%realfillerfarmed
36/40 — real replies, no farming

Is the engagement real?

Reply content over account stats: wallet-address giveaway entries, contentless praise, absence of substance. Validated at 91% against hand-labelled posts.

foreveryone
profile
QUIET · REAL MONEYLOUD · NO MONEY$0$3Kengagement / 1k →
quiet audience, real money — 28.8K followers, $3.0K per 1k

Loud, or actually effective?

Engagement and volume, both per thousand followers, against the measured p75 of each. Separates the quiet audience that trades from the big one that only makes noise.

forproject teams

Charts rotate through real posts from the corpus. Hover for the underlying numbers.

Measurement for paid crypto promotion

Check one post.Score a whole campaign.Stop paying for reach.