Every business I have worked in had a list of people who had bought once and stopped, or subscribed and cancelled. In almost every case that list was larger than the active customer base, and in almost every case nobody owned it.
Acquisition has a team, a budget and a weekly meeting. Win back has a quarterly campaign, if it has anything. That imbalance is not a considered decision, it is an accident of how teams get built.
Why lapsed customers are different
They know your product, your price and your delivery. You do not have to buy their attention or explain who you are. The only question is whether the reason they left still applies.
That last point is the one that makes win back work or fail. Sending a discount to someone who left because the delivery was late is not a win back, it is a reminder of the problem with money attached.
Sending a discount to someone who left over a late delivery is a reminder of the problem with money attached.
Segment by reason, not by recency
- Left over price. A price mechanic will work, and it is the only group where discounting is the right answer.
- Left over an experience failure. Acknowledge it before offering anything, and only after you have actually fixed it.
- Left because the need ended. Leave them alone until the need plausibly returns, then be specific about why now.
- Left involuntarily, through a failed payment nobody chased. Fix the payment, not the relationship.
- Never really started. Bought once, never engaged. This is an onboarding problem wearing a churn costume.
Most win back programmes treat all five identically, at a recency cut off, with the same offer. Then they conclude that win back does not work.
The involuntary slice
Worth pulling out separately because it is the fastest thing on this list to fix. Cards expire, payments fail, retries happen at unhelpful times, and dunning messages read like collections notices. A meaningful share of what gets counted as churn is really an unpaid invoice with no human follow up.
Nobody is emotionally attached to fixing dunning, which is exactly why it stays broken.
Where to start
Pick the smallest well defined group, usually involuntary churn or the last ninety days of cancellations with a stated reason. Run something deliberately narrow. Measure it against a holdout. Then decide whether to build the programme.
A single successful narrow test will get you the resource. A broad campaign that performs averagely will get win back quietly dropped for another two years.