Retention & LTV Analyst

Retention And LTV Analyst

RETENTION
LTV
WEEKLY
ECOMMERCE

Works with the AI tools you already use

+19 agents

Weekly retention review for ecommerce teams: find where cohorts are actually breaking, then recommend the VIP, winback and second-purchase work worth doing.

ABOUT WHAT IT DOES USE CASES EXAMPLE INSTALL

Retention gets reported as a single repeat rate, which moves slowly and explains nothing. Underneath it, cohorts acquired from different channels at different discounts behave so differently that the average describes none of them. A repeat rate holding steady can hide a channel whose customers never come back, masked by one that is improving. This skill reads retention by cohort and by acquisition source, finds where the curve actually breaks, and recommends the VIP, winback and second-purchase work worth doing this week.

What It Does

Splits retention by acquisition cohort and channel rather than reporting one rate.
Finds the point in the customer lifecycle where each cohort actually drops.
Projects lifetime value per cohort against what it cost to acquire.
Identifies the customers worth a VIP program and the ones who are not.
Sizes a winback against the margin the returning orders would carry.
Recommends the campaigns and tactics, with the cohort and the expected value.

Use Cases

Find the acquisition channel whose customers never make a second purchase.
Target a winback at the cohort where it still has a chance of working.
Size a VIP program against what the top cohort actually spends.
See which month of the lifecycle the drop really happens in.

Example Session

Where is retention breaking, and what should we do about it this week?

Discounted acquisition is the problem. Those customers rarely come back.

Cohort | Second purchase | 12 month LTV | CAC | Payback | Action
Paid social, 20% first order | 11% | $74 | $68 | Never | Stop the discount
Paid social, full price | 34% | $218 | $91 | 4 months | Scale
Organic and referral | 41% | $286 | $12 | Immediate | VIP program
Email winback, lapsed 6-12mo | 29% | $134 | $9 | Immediate | Run this week

The discounted cohort costs nearly as much to acquire as it ever returns, because a customer taught to buy at 20% off waits for the next 20% off. Full price paid social from the same channel retains three times better. The lapsed six to twelve month segment is the best available action this week: they already know you and cost nine dollars to reach.

Why This Beats Prompting It Yourself

One Repeat Rate vs Retention By Cohort: The Hidden Risk

A generic prompt reads the repeat purchase rate, sees it holding steady, and suggests a loyalty program. The steady average is concealing a channel whose customers never return being offset by one that is improving. Acting on the average means launching a loyalty program aimed at nobody in particular.

When You Still Need The Discount's Effect On Retention

A first-order discount is judged on acquisition cost, where it looks efficient. Its real cost arrives months later as a cohort that retains at 11% against 34% for the same channel at full price. This skill puts those two side by side, which is the only way the discount's true price becomes visible.

Known Limitations

Lifetime value at twelve months is projected for cohorts younger than that, and the output marks which figures are projected rather than observed. Channel attribution for acquisition inherits whatever your tracking supports, so cohorts from poorly attributed channels are grouped and labeled. It recommends the campaigns and does not create or send them.

Install It

Download the ZIP.
It is free and there is no account to create.
Unzip it into your agent's skills folder.
Claude Code reads ~/.claude/skills/, which is hidden by default: the command in the folder block opens it. Other agents scan their own directory, so drop the same folder there instead.
Ask your agent to use it.
Restart the agent if it was already running, then it picks the skill up with no config.

skills folder

Copy

~/.claude/skills/retention-and-ltv-analyst/
  SKILL.md
  references/cohort-model.md
  weekly/
  # one file per week, with the cohort recommendations

# macOS: create the folder and open it in Finder
mkdir -p ~/.claude/skills && open ~/.claude/skills

# Windows: paste in the Explorer address bar
%USERPROFILE%\.claude\skills\

Frequently asked questions

How is this different from a retention dashboard?

A dashboard shows the curves. This finds where each cohort breaks, projects value against acquisition cost, and returns the specific campaigns worth running this week with the segment attached. The output is work rather than charts.

Which agents does the skill run in?

Any agent that supports the open SKILL.md format: Claude Code, Cursor, Codex CLI, GitHub Copilot, Gemini CLI, Manus, Grok and others load it unmodified. The format is portable, the location is not. Each agent scans its own skills directory, so you drop the same folder into whichever one yours uses. There is no config file to edit and no API key to provision.

What does it need connected to work?

Order history with customer identifiers, acquisition source and date, discount usage, and product margin. Acquisition cost per channel is needed for the payback column and everything else works without it.

Does a first-order discount really hurt retention that much?

In most accounts that separate the cohorts, yes, and the gap is usually larger than anyone expects. It is not that discounts are bad; it is that a discount teaches a purchase rule, and customers acquired under that rule keep applying it.

Why run a winback before fixing acquisition?

Because it is this week's highest expected value and the acquisition fix takes a quarter to show. The skill sequences them rather than choosing: run the winback now, change the discount policy for new cohorts, and measure the second one in three months.

What is included with the skill?

The skill itself, the cohort model reference, the weekly recommendation format, and the lifetime value projection rules. It is free and security scanned. We re-publish the ZIP when the model changes, so download it again if your cohorts start shifting.

Other skills

Pricing Analyst

Sets the price floor, reads elasticity across the catalog, and writes the pricing tests and discount changes worth running next.

Discount Ladder

Decides the next markdown step instead of guessing at 20%: increases, holds, or pulls the discount back.

Out of Stock

Flags the SKUs about to run dry, updates the ship-date promise, and queues the actions to pause Paid on them.

Popular in E-commerce