retention-diagnosis
Skill0 starsChanged 2 months agoMITView on GitHub
Read retention and churn honestly using cohorts, separating a genuine retention problem from a mix-shift, a measurement artefact, or an acquisition problem wearing retention clothes. Use when churn rises, before building winback, or when asked why retention is falling.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- 45ck/marketing-product-skillstheir own version, first seen Mar 27, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- 45ck/marketing-product-skills★ 3same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- activationSkill
- capacity-economicsSkill
- channel-fitSkill
- data-trustSkill
- experiment-designSkill
- form-frictionSkill
- funnel-diagnosisSkill
- pricing-decisionsSkill
- unit-economicsSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 26
Look for where the curve flattens.
- Distinctive: line 48
A renewal not yet due is not a retention: Only count customers who reached the decision
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Retention diagnosis
Most retention problems are acquisition problems that surface later.
Always cohort. Never look at aggregate churn.
Aggregate churn moves when the mix changes, with no change in behaviour. Grow acquisition quickly and aggregate churn rises simply because new customers churn faster than tenured ones. The business improved; the metric worsened.
Split by signup month, then read down each cohort. That is the only view where a real change is visible.
The question to ask first
Did the customer ever get value? Retention curves usually break in the first period, not gradually. A cohort that drops 60% in month one and then flattens does not have a retention problem, it has an activation problem, and winback campaigns aimed at those people will not work.
Look for where the curve flattens. If it flattens, the business has a stable core and the issue is getting people to it. If it never flattens, the product is leaking indefinitely and no amount of lifecycle marketing fixes that.
Separate the four causes
| Symptom in cohorts | Likely cause |
| Recent cohorts worse from month one, older ones stable | Acquisition quality changed. Check channel mix and targeting first |
| All cohorts worsen at the same calendar point | Something happened: a price change, an outage, a policy change, a competitor |
| Curve declines steadily and never flattens | Product value does not sustain. Not a marketing problem |
| Only one segment worsens | Mix shift. The average is hiding two populations |
The first row is the most common and the most misdiagnosed. A channel that delivers cheap signups who never activate shows up as a retention problem three months later, and the retention team gets asked to fix it.
Measurement traps
- Define churn before measuring it. Cancellation, failed payment, and non-usage are three different events. Involuntary churn from expired cards is often a large share and is a payments problem, not a marketing one
- A renewal not yet due is not a retention. Only count customers who reached the decision
- Paused is not churned, but do not let it hide churn either. Track it separately and check whether paused customers return
- Survivorship. Reading only surviving customers' satisfaction tells you about survivors
Before recommending winback
Winback works on people who got value and left for a reason. It does not work on people who never activated. Segment by whether the customer reached the value moment before spending anything on reactivation.
When to refuse
- Cohorts too young to show a curve. Say how long is needed
- Fewer than ~50 customers per cohort
- Churn definition unstated
Output
The cohort table, where the curve flattens, which of the four causes fits with evidence, and one intervention aimed at the actual cause. Say plainly if the diagnosis is that marketing cannot fix it.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

