UAE · ENBook a demo
RetailCustomerInvestigator

Member Lapse Prediction Agent

Flags loyalty members likely to stop shopping and why.

Member Lapse Prediction AgentChurn Prediction Agent · Customer Lifecycle
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.

  1. 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..

  2. 02Reads the request

    Works out what is being asked and what a complete answer needs.

  3. 03Checks the facts

    Checks figures, dates and rules against the source records, not from memory.

  4. 04Works it through

    Weighs competing explanations against the evidence and shows which one holds up.

  5. 05Asks a person

    Hands the work to Customer Service Manager with its reasoning before anything goes out.

  6. 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.

ApproverCustomer Service Manager

Settings

Your limits, set by your managers.

SettingDefaultWhat it controls
Gap that signals lapse (times the usual cycle)2A member whose gap since last purchase is this many times their usual cycle is at risk.
Failed deliveries that signal lapse2This many failed or late deliveries in 90 days is a lapse signal.
Flag members with a risk score above0.6Members above this score, from 0 to 1, are flagged.
Most members flagged per run5000Only 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.

Member Lapse Prediction Agent · Retail | Basinwright Industries