First Purchase Discounts: Do Those Customers Return?

The welcome popup with a first purchase discount is on practically every store, and it works: the coupon converts a visitor who was about to leave. The question nobody's dashboard answers is what happens afterwards. Did the first purchase discount buy you a customer who returns at full price, or did you just sell one order below margin to a stranger you will never see again? The two outcomes look identical in the month's revenue and completely different a year later, and your own order history can tell them apart.

This guide runs the margin math, walks through what the research actually found, borrows a painful lesson from JCPenney, and shows how to measure discount-acquired customers on your own cohorts instead of guessing.

What a first purchase discount really costs

A discount comes out of profit, not revenue, which multiplies its true cost. At a typical 30% contribution margin, a 100 euro order carries 30 euros of profit; each discount tier eats a disproportionate share of it:

A 20% welcome code at a 30% margin gives away two thirds of the order's profit, so the campaign around it must produce three times the volume just to stand still.

That math is survivable exactly when the discounted first order buys a RETURNING customer, whose later full-price orders repay the introduction. Which is why the entire question collapses into one measurable thing: do they come back?

Do discount customers come back?

The research is more nuanced than the folklore. Optimove's research, as summarised by Peel, found a sweet spot: new customers were much more likely to make repeat purchases when the first purchase discount fell in the 5 to 20% range, with a negligible difference in repurchase rate and customer lifetime value between 10% and 20% off.

The same summary warns that deeper first purchase discounts can lift order volume while lowering average order value and lifetime value. And the stakes of the second purchase are high in either case: in Enalitica's data (nearly 30,000 customers with an order in the 12 months to 28 September 2026), only 14.7% of customers ordered again within 12 months, yet repeat orders made up 23.4% of all orders.

The honest summary is that a moderate welcome discount is usually fine, an aggressive one usually buys traffic, and only your own cohort curves can say which side of the line your store is on. The same math at the year's biggest scale is Black Friday, where the average discount hits 29%.

How discounts train customers: the JCPenney lesson

Discounts do not only select which customers you attract; they teach the ones you keep. JCPenney is the canonical case: a department store whose customers were used to constant promotions scrapped sales and coupons for everyday "fair and square" prices in January 2012. Customers, trained for years to wait for the deal, walked away: sales fell by $4.3 billion that year and the CEO was replaced in April 2013.

The ecommerce translation: a store whose newsletter is a coupon calendar is running a small JCPenney. Existing customers learn to park their cart until Sunday's code arrives, and every "welcome" discount that leaks to returning buyers converts full-price demand into discounted demand. The training effect cuts both ways, which is why the fix is measurement and discipline, not prohibition.

How do you measure discount customers in your store?

Split your customers by how their FIRST order was paid: with a discount or at full price, then compare the two groups' profit per customer over the following months. Every later purchase counts toward the group of the first order, so the curves answer the exact question the popup raises: do discount-acquired buyers catch up, or stay permanently cheaper?

Three details decide whether the comparison is honest:

1. Count coupons, not just line discounts. Cart-level coupon codes are the most common discount and the easiest to miss when they are recorded separately from item prices; a comparison that skips them classifies most discounted orders as full price.
2. Compare PROFIT per customer, not revenue. The discount already cut margin once on the first order; revenue curves hide that.
3. Give the cohorts time. Judge at month 3 and month 6, not week 2, and read the repeat-purchase share next to the value curve.

A small gap between the curves means the welcome discount works as a greeting: pay it with confidence, and treat it as part of your customer acquisition cost. A large, persistent gap means you are buying one-time orders, and the discount budget deserves to move.

Discount rules that buy customers, not deal hunters

• Stay in the 5 to 20% band. In my experience a code under 5% rarely changes a decision, and above 20% the margin math turns brutal. The research above found almost no difference in lifetime value between 10% and 20%, so if 10% converts nearly as well in your store, take the cheaper greeting.
• Charge it to acquisition, not to everyone. First order only, new emails only, excluded from sale items. The most common leak is an Advantage+ campaign without excluded existing customers, which shows the welcome offer to exactly those buyers.
• Do not let the newsletter become a coupon calendar. Existing buyers learn schedules fast; JCPenney's customers needed years to learn and did not unlearn in time.
• Watch the discounted group's basket. If discount-first buyers also buy smaller baskets, the gap is wider than the code percentage suggests.
• Recheck twice a year. Audiences, products and competitors shift; the curves that justified the code last year may not this year.

A welcome code that returning customers can reuse is not a welcome discount. It is a price cut wearing a costume.

How Enalitica shows whether discount customers stay

Enalitica answers the question from your orders: the cohort card groups customers by first order with a discount versus full price and draws profit per customer for both groups across the twelve months after the first purchase, repeat-purchase share included. Coupon orders are counted as discounted even when the coupon is recorded separately from the products, exactly the case most reports misclassify.

Because the curves are built on profit (product costs, shipping and fees subtracted, same engine as POAS), the comparison already contains the margin the discount consumed. One glance settles whether your welcome code is a greeting or a leak, and the same view shows what a new customer may cost for each group.

Book a demo and bring your coupon report; we will read your two curves together.

Frequently Asked Questions

What is the best first purchase discount percentage?

Optimove's research, summarised by Peel, points to 5 to 20%, with a negligible difference between 10% and 20% in repurchase rate and customer lifetime value. Deeper codes tend to lift order volume while lowering order value and lifetime value. Start at 10%, measure your discount-versus-full-price cohorts, and only escalate if the curves stay together.

Do discounts lower customer lifetime value?

Moderate first purchase discounts usually do not: customers acquired with 5 to 20% codes were much more likely to repurchase in Optimove's research. Deep discounts do, by attracting deal hunters, and recurring sitewide promotions do, by training existing customers to wait. The clean test is profit per customer by first-order type, over months.

How do I keep bargain hunters away from my welcome code?

Constrain the code: first order only, one per email, excluded from already-discounted items, modest percentage. Deal hunters chase depth; a 10% greeting is unattractive to them and nearly free compared to the 25% codes that summon them.

Should returning customers ever get discount codes?

Sparingly, and for a reason: reactivation after a long pause or a service failure apology beats a scheduled monthly coupon. Every code a returning customer expects converts an order they would have placed at full price into a discounted one, which is pure margin loss.