Syntha Lab RU

Lapsed-customer threshold from repurchase gaps

Derives the threshold from your data and not from a round number: takes the gaps between repeat purchases and finds the period within which a set share of repeats falls (80% by default). A customer with no purchase for longer than the threshold is “sleeping”, longer than the threshold times a multiplier (2 by default) is “lost”. If you enter customers by the recency of the last purchase, shows the shares of active, sleeping and lost customers and the number of customers for a win-back campaign.

Customers and loyalty For whom: CRM managers, customer base analysts and marketers who define when a customer counts as lapsed.

Calculate on your data →

At a glance

What question does it answer?
After how many days without a purchase a customer can be considered sleeping and lost by your data, and how many customers are in each status.
What data is needed?
The gaps between repeat purchases of customers in days for one category or season.
How to read the result?
The threshold is a quantile of your gaps: compare it with the accepted 90 and 180 days and see how many customers turned out “sleeping”: that is the target of the win-back campaign. Example on the default values: Threshold “customer is sleeping”: 248 days; Threshold “customer is lost”: 496 days; Median gap between purchases: 135 days; Gaps in the calculation: 12. The figures are illustrative: enter your own to see your result.

Why and how to reason

A round threshold of 90 or 180 days is not tied to how often your customers buy. We derive the threshold from data: a quantile of the gaps between repeat purchases. In the example 12 gaps from 20 to 400 days at a share of 80% give a threshold of 248 days, “lost” after 496; by customer groups active 60.8%, sleeping 370 people, lost 22.9%.

When to use it

Apply when setting up customer base segments and win-back campaigns, separately by season and category.

How to read the result

First the threshold, then the number of sleeping customers: it is the win-back target before they become lost.

Common mistakes

They accept fixed 90 or 180 days and count gaps from a single order.

What data is needed?

  • The gaps between repeat purchases of customers in days for one category or season.
  • The share of repeats that must fall within the threshold.
  • Customers by the recency of the last purchase (optional).
  • An example: 12 gaps from 20 to 400 days, share 80%: threshold 248 days, “lost” after 496 days (illustrative numbers).

Inputs

FieldUnitValue in the exampleMeaning
Share of repeats that must fall within the threshold%80The quantile of the gaps. The higher the share, the longer the threshold and the fewer false “lapsed”. The example value is illustrative.
The multiplier of the threshold for “lost”×2How many thresholds without a purchase until the “lost” status. The example value is illustrative.

How it is calculated

StepFormula
Order of gapssort ascending, position = (n − 1) × share of repeats
Thresholdthe value at the position with linear interpolation between neighbouring gaps
Statusesactive up to the threshold, sleeping from the threshold to threshold × multiplier, lost beyond

Worked numbers

Default values from the app; in the interactive version you enter your own.

Threshold “customer is sleeping”
248 daysThe quantile of gaps between purchases.
Threshold “customer is lost”
496 days
Median gap between purchases
135 days
Gaps in the calculation
12
Active
60.8 %1,380 customers
Sleeping (win-back target)
37016.3 % of customers
Lost
22.9 %520 customers
Gaps between purchases in ascending order and the threshold, days
Bar chart: 1 20, 2 35, 3 50, 4 70, 5 95, 6 120010020030040011Gap: 2020Threshold “sleeping”: 24824822Gap: 3535Threshold “sleeping”: 24824833Gap: 5050Threshold “sleeping”: 24824844Gap: 7070Threshold “sleeping”: 24824855Gap: 9595Threshold “sleeping”: 24824866Gap: 120120Threshold “sleeping”: 24824877Gap: 150150Threshold “sleeping”: 24824888Gap: 180180Threshold “sleeping”: 24824899Gap: 200200Threshold “sleeping”: 2482481010Gap: 260260Threshold “sleeping”: 2482481111Gap: 320320Threshold “sleeping”: 2482481212Gap: 400400Threshold “sleeping”: 248248
  • Gap
  • Threshold “sleeping”
Data as a table
MeasureGapThreshold “sleeping”
120248
235248
350248
470248
595248
6120248
7150248
8180248
9200248
10260248
11320248
12400248

Customers by status

GroupDays sinceCustomersStatus
Bought in the last 2 months45480Active
2–3 months ago100360Active
4–6 months ago165300Active
7–8 months ago225240Active
9–10 months ago295200Sleeping
11–12 months ago350170Sleeping
More than a year ago540520Lost

Apparel is seasonal: count the threshold by season and category and not as one number for all; do not take gaps from a single order.

What to do with the result: put the threshold into the customer base segmentation and launch the win-back for sleeping customers before they become lost; count the payback of the campaign separately.

A typical mistake: accepting fixed 90 or 180 days and counting gaps from a single order.

Nuances and limits

  • Accepting fixed 90 or 180 days.
  • Counting gaps from a single order.
  • Taking one threshold for all categories and seasons.

Terms used

Calculate on your data →