2026 Ecommerce Retention Trends: Reasons, Pauses, and Discounts
Our latest ecommerce data shows why customers cancel, when a pause is worth testing, and why a deeper discount can win more acceptances without delivering much more revenue.
Founded in 2020, ProsperStack has helped subscription businesses process more than 1 million cancellation sessions in Q3 2026 and holds a 5 star rating on G2. Our cancellation flows capture why customers consider leaving and how they respond to retention offers.
For this article, we examined data from a small set of ecommerce merchants with comparable records in 2025 and 2026. We matched calendar periods and checked whether survey questions and discount terms stayed the same. We also examined randomized discount tests from 2026 and the recorded payments that followed over 90 days.
The findings describe these merchants, rather than the ecommerce industry as a whole. They offer three useful observations for planning your 2027 retention tests: which cancellation reasons stayed consistent, what a new temporary-break answer revealed, and why more discount acceptances did not always translate into much more revenue.
1. Price and excess product remained stable at the two merchants we compared
Merchant | Matched period in both years | Cancellation reason | Change |
|---|---|---|---|
Merchant 1 | January–July | Excess product | +0.4 pp |
Merchant 1 | January–July | Price | −1.4 pp |
Merchant 2 | January 1–June 14 | Excess product | −0.5 pp |
Merchant 2 | January 1–June 14 | Price | −1.0 pp |
Both comparisons used unchanged survey questions and answer options. Percentages represent shares of answered cancellation-flow sessions, not all subscribers. Changes are in percentage points (pp).
The differences were small, and two merchants cannot establish an ecommerce-wide trend. Price and product backlogs still deserve attention, with the offer chosen around the customer's reported problem.
2. A new temporary-break answer revealed a new segment in 2026
In January 2026, another merchant added a temporary-break answer to its cancellation survey, along with a few other new options. From January through September 2026, the temporary-break answer was the third most common response.
That creates a segment worth investigating. Customers who expect to return may need a different alternative from customers who cannot afford the next order or consistently receive too much product. Test a pause or skip for a temporary interruption where your platform supports it, and plan follow-up around the expected return date.
The new options also changed how the year-over-year results should be read:
Merchant 3, January–September, 2025 vs 2026
Cancellation reason | Change between 2025 and 2026 |
|---|---|
Price | −16.2 pp |
Excess product | −8.7 pp |
The new options, especially the temporary-break answer, may have captured customers who previously chose price because they wanted to stop spending for a while. Customers with unused product still had a specific answer that described their situation. That could help explain why price responses fell more sharply than excess-product responses, although we cannot tell which answer individual customers would have chosen under the old survey.
3. 2026 discount tests showed stronger acceptance than lasting retention
Where brands kept the same discount in both years, about as many customers accepted it in 2026 as in 2025.
In randomized tests, deeper discounts consistently won more acceptances. At one brand, 20% off instead of 10% doubled acceptances and the recorded revenue that followed, mostly because more customers placed another order. But most of those customers canceled within 90 days: for every 100 customers shown the offer, about one additional customer per 100 shown the deeper discount had made a payment and had no cancellation recorded within 90 days. At another brand, 50% off instead of 40% won about half again as many acceptances with almost no additional revenue.
Before expanding a larger offer, track who is still subscribed and paying 60 to 90 days later, not just who accepted, and weigh that against the discount’s cost.
Plan your 2027 retention tests with ProsperStack
Use these findings to set your 2027 priorities: address the reasons customers leave, give temporary breaks a path back, and judge discounts by the customers who stay and keep paying.
ProsperStack’s customer success team can help you review your cancellation data, choose an offer, and configure your first test. Our cancellation flows support targeted offers and experimentation, with available actions depending on your subscription platform.
Talk with our team about your 2027 retention plans.
About the evidence
Research reviewed October 2026. Reason comparisons use answered cancellation-flow sessions in matched 2025 and 2026 periods, grouped into consistent categories. The unchanged-survey examples cover two merchants; the changed-survey example covers a third. Subscriber-level checks showed similar small movements in the unchanged-survey comparisons.
Discount comparisons use session-level acceptance for eight unchanged offers at three merchants, with offer terms and calendar windows matched. The fixed-mix aggregate weights both years using 2025 presentation volumes. One merchant accounts for 98% of volume. Latest coupon settings and recorded edit dates were checked, but historical offer copy is unavailable. Matched-cohort revenue and margin are not supported.
The 2026 randomized discount examples use January–June presentations with 90 days of follow-up. Payments are retention-attributed, not complete order revenue or margin. External research has different populations and methods and does not validate our year-over-year results. These findings describe our sample, not a market-wide change. The randomized tests had no group that saw no offer, so we cannot tell whether customers who canceled after a discounted order would have canceled sooner without it. “Still subscribed and paying” is measured where cancellations are recorded reliably.
