Do Discount Customers Become Profitable? A 90-Day Ecommerce Cohort Analysis

An ecommerce cohort analysis groups customers by when and how they were acquired, then tracks what they do next at equal ages. It is the only honest way to answer a question that comes up every December: did the customers we bought with discounts turn into profitable customers, or did we just pull revenue forward? This post lays out a 90-day framework for comparing promotional and full-price cohorts on repeat purchase rate, returns, and contribution per customer, built to run on your BFCM 2026 cohort starting the week after Cyber Monday (November 30).

What does the research say about discount customer lifetime value?

The research leans against discount customers, but not uniformly. A Journal of Marketing Research study using customer-level data from an online grocer and a newspaper found that acquisition discount depth was negatively related to repeat-buying rates and customer asset value, with customers acquired at a 35% discount worth roughly half the long-term value of those acquired at full price.

Holiday data points the same way. One retailer dataset published in November 2025 found only 22.59% of first-time Black Friday buyers bought again within 12 months.

But the counterexamples are real. Peel's walkthrough of a discount-acquired cohort shows a bundle-promotion cohort outgrowing its peers on LTV over four months, with the caveat that seasonality or the product itself may explain part of the lift. The depth of the discount, the product it was on, and the channel it ran through all change the outcome. That is why you measure your own cohorts instead of borrowing a rule.

How to set up a promotion cohort analysis

A promotion cohort analysis needs three decisions made before you pull any data.

Define the cohorts

• Promotional cohort: new customers whose first order used a discount code or sale price above a threshold you choose, for example 15% or more off.
• Full-price cohort: new customers acquired in a comparable window whose first order had no discount, or less than 5%.
• Holiday promotional cohort: new customers whose discounted first order fell between November 26 and December 1. Keep it separate, since holiday behavior differs from an evergreen promo.

Compare at equal ages

Compare every cohort at day 30, day 60, and day 90 after first order, never "to date." A September full-price cohort will always look better than a December promo cohort if you measure both on January 15, simply because it has had more time.

Use contribution, not revenue

Contribution per customer means revenue minus COGS, shipping, payment fees, discounts, and returns, averaged across every customer in the cohort, including the ones who never came back. Revenue comparisons hide the two things that hurt promo cohorts most: thinner margins and higher returns.

The three metrics that decide whether discount customers pay off

1. Repeat purchase rate

Repeat purchase rate is the share of the cohort that placed a second order by each checkpoint. For context, Eightx's 2026 cohort guide puts 90-day cumulative repeat rates at roughly 22% to 32% for apparel and 30% to 50% or more for supplements and beauty. Benchmark your promo cohort against your own full-price cohort first, and the industry ranges second.

Also check whether repeat orders were placed at full price. A promo customer who only returns for the next sale is a different asset from one who converts to full price.

2. Returns

Returns rate on the first order, and on repeat orders, by cohort. The NRF expected 17% of holiday sales and 19.3% of online sales to be returned in 2025. If your promotional cohort returns materially more than your full-price cohort, the first-order revenue was never as large as it looked.

3. Contribution per customer

This is the bottom line. Subtract new customer CAC from cumulative contribution per customer at each checkpoint to see whether each cohort has paid back.

Example: promotional vs full-price cohorts at 30, 60, and 90 days

Illustrative numbers for an apparel brand comparing a 25%-off promotional cohort with a full-price cohort acquired in the same quarter:

In this example the promo cohort was cheaper to acquire, which is why so many brands think promotions work. By day 90, though, lower repeat and higher returns leave it slightly underwater while the full-price cohort has paid back. That gap is small enough that a better post-purchase flow might close it. The value of the analysis is knowing which situation you are in.

What to do with the results

• If promo cohorts trail but are close: keep the promotion, change what happens after. Move promo buyers into a full-price second-purchase flow rather than the next discount email.
• If promo cohorts trail badly: cut discount depth, or restrict the discount to products whose first-time buyers repeat well. Product mix inside the promo often matters more than the percentage.
• If promo cohorts match or beat full price: you found an acquisition offer worth scaling. Document which code, product, and channel produced it.

Then use your audiences deliberately. AdBeacon's RFM segments let you watch where each cohort lands over time, whether promo buyers move into Potential Loyalists and Loyal Customers or slide into About to Sleep. The Discount Shopper lookalike is useful for clearance campaigns, and keeping it out of full-price prospecting protects your mix. For more on predicting who buys next, see how to predict the next purchase.

Attribution matters for cohort analysis too

Cohort analysis tells you whether discount customers paid off. It also needs to tell you which channel acquired them. If platform-reported results are used to assign customers to campaigns, a view-through claim can put a full-price customer into a promo campaign's numbers, or the reverse. AdBeacon assigns orders with click-only, first-party attribution and shows new vs returning customers, nCAC, and POAS by channel next to platform-reported numbers, which is more accurate and more verifiable than combining each platform's own claims. Our post on the customer journey to sale and beyond explains why that journey view matters after the first order.

Run it on your BFCM 2026 cohort

Set the cohort definitions now, before November 27, so you are not choosing them after the results come in. Check the 30-day mark on December 30, 60 days in late January, and 90 days at the end of February. By then you will know whether this year's holiday discount bought customers or just orders. If you want to see new customer performance by channel next to platform-reported numbers on your own data, book a demo with AdBeacon.

Sources

• Lewis, Journal of Marketing Research (2006): Customer acquisition promotions and customer asset value
• eCommerceNews: Online retailers see Black Friday boost but loyalty lags behind
• Peel Insights: To Discount or Not to Discount?
• Eightx: What Is Cohort Analysis in Ecommerce
• NRF: Consumers expected to return nearly $850 billion in merchandise in 2025