Should You Discount Lapsed Customers?
Sometimes the correct win-back offer is 20% off.
Sometimes the correct offer is nothing.
The difference is whether the discount changes behaviour profitably.
The margin question
Suppose:
Product price:
$100
Contribution margin before discount:
60%
Contribution:
$60.
Offer:
20% off.
Revenue becomes:
$80.
Assume product cost stays:
$40.
Contribution becomes:
$40.
The discount cost you:
$20 contribution per order.
If the customer needed the discount, that may be worthwhile.
If they were going to buy anyway, you paid $20 for nothing.
Use the holdout
This is why discount tests should be measured against a no-discount group where practical.
Treatment:
8% purchase rate with 20% off.
Control:
6% purchase rate without offer.
The discount did not “generate 8%.”
The apparent incremental lift is closer to:
2 percentage points.
Now calculate whether those extra orders justify the margin lost on all discounted orders.
That is the real question.
A worked example
1,000 customers.
Discount group
80 orders × $40 contribution
= $3,200.
No-discount group
60 orders × $60 contribution
= $3,600.
The discounted group generated more orders.
It generated less contribution.
That is why conversion rate alone can mislead you.
Delay discounting when natural purchase is likely
NICKIS reportedly used later incentives only after customers had moved beyond the expected second-purchase period, specifically to reduce unnecessary margin loss on purchases likely to happen naturally. (klaviyo.com)
That principle makes sense far beyond one brand.
Use discounts where friction is genuinely financial
Attentive’s 2026 research found promotions were the most frequently cited factor that could bring shoppers back to a brand. So financial incentives absolutely matter. (deloitte.com)
The question is not whether discounts work.
It is whether they are necessary for this customer.
Better hierarchy
- Relevance
- Convenience
- Education
- Newness
- Recognition
- Incentive
Not every campaign must follow this exact order.
But discount should be deliberate.
Frequently asked questions
How much should a win-back discount be?
There is no universal number. It depends on margin, customer value, lapse depth and how sensitive the cohort is to incentives.
Should VIP customers get bigger discounts?
Not automatically. VIPs may respond better to recognition, access or service.
Should the first win-back email have a coupon?
Usually not unless data shows a financial barrier is the primary reason for lapse.
The principle
Do not ask:
What discount converts best?
Ask:
What is the cheapest intervention that creates profitable incremental behaviour?
That is the better question.