Member Lapse Prediction Agent
Flags loyalty members likely to stop shopping and why.
- Industry
- Retail
- Department
- Customer
- Team
- Customer Lifecycle
- Kind
- Investigator
- Approver
- Customer Service Manager
- Code
- CUS-14
It flags loyalty members whose shopping pattern says they are about to lapse, such as a longer gap than their usual cycle, a switch away from a category or repeated failed deliveries, and explains the likely reason for each group. It proposes a win-back route for each reason, from a service recovery call to a category offer. **The problem it solves:** Why these loyalty members are drifting away from their normal shopping pattern, and which reason applies to each group.
How it works
6 steps, and a person signs off before step 6.
- 01Looks it up
Finds the record the request is about in the loyalty and CRM platform: members' purchase history, usual cycle, category mix, failed deliveries, open cases and consent..
- 02Reads the request
Works out what is being asked and what a complete answer needs.
- 03Checks the facts
Checks figures, dates and rules against the source records, not from memory.
- 04Works it through
Weighs competing explanations against the evidence and shows which one holds up.
- 05Asks a person
Hands the work to Customer Service Manager with its reasoning before anything goes out.
- 06Acts
Once approved, it acts in the loyalty and CRM platform: it tags the flagged members with their lapse risk and reason for the win-back journey. and records what was done.
What it works with
Your knowledge, your systems, your people.
Knows
The lapse model: each segment's normal purchase cycle, the lapse signals and their weights, the known reasons members leave (delivery failures, price, range, store closure), and the win-back routes allowed.
Reads from
the loyalty and CRM platform: members' purchase history, usual cycle, category mix, failed deliveries, open cases and consent.
Acts in
the loyalty and CRM platform: it tags the flagged members with their lapse risk and reason for the win-back journey.
Hands back
An object with flagged_count, groups (a list of reason, members, risk_score_average, evidence, confidence and win_back_route) and missing_signals (a list).
Asks a person when
A group's win-back route includes a discount or points offer, or the likely reason is a store closure or a delivery partner.
Settings
Your limits, set by your managers.
| Setting | Default | What it controls |
|---|---|---|
| Gap that signals lapse (times the usual cycle) | 2 | A member whose gap since last purchase is this many times their usual cycle is at risk. |
| Failed deliveries that signal lapse | 2 | This many failed or late deliveries in 90 days is a lapse signal. |
| Flag members with a risk score above | 0.6 | Members above this score, from 0 to 1, are flagged. |
| Most members flagged per run | 5000 | Only the highest-risk members up to this number are flagged. |
Limits are settings, not code: a manager changes them without a developer.
Rules it always follows
- Flag a member when their gap since last purchase reaches 2 times their usual cycle, or when other signals push the risk score above 0.6.
- Count 2 or more failed or late deliveries in 90 days as a strong signal, and route those members to service recovery before any offer.
- Group flagged members by likely reason and weigh at least two reasons per group.
- Flag at most 5000 members, highest risk first.
- Never use protected characteristics as signals, and contact only members with marketing consent.
See the Member Lapse Prediction Agent agent on your own work.
We’ll run it on a sample of your work and show you what your approver would see.