I spent 8 weeks talking to cohort organizers. The dropout problem isn’t what I expected.

Eight weeks ago I had a simple hypothesis:

Cohort organizers lose students silently - and by the time they notice, it’s already too late.

After talking to multiple organizers, a pattern became obvious.

The pattern

Drop-off doesn’t happen at the start.

It happens around day 20–30.

Not when motivation is high -
but in that specific window where excitement has dropped and habits haven’t formed yet.

Everyone described the same thing:

Someone goes quiet.
No message. No signal.
Just… gone.

Most tools only show completion rates after the cohort ends.
There’s no early warning.

The root cause

It’s not content quality.
It’s not the platform.

It’s that dropping out costs nothing.

No record.
No consequence.
No one whose opinion the student values actually notices.

Quitting is free.

A useful insight

One accountability coach running 100+ day programs described it as a “witness problem”:

Public commitment without a credible witness is just journaling with an audience.

He caps cohorts at ~12-15 people.

Not because of time -
but because real witnessing doesn’t scale.

Where I’m at

I’m exploring a way to make commitment and dropout visible before it’s too late.

Early experiments:

a public record of commitments (can’t be edited or deleted)

a rough idea of how organizers could detect “drift” early

Still very early. No revenue.

Two questions:

• If you run cohorts, where does silent dropout actually happen for you?
• Is “making dropout cost something” a real lever, or just a convenient theory?