Coupons & Promos

Coupon marketing strategy

When to discount, how much, and how to do it without training customers to wait.

8 min read Updated April 29, 2026

A coupon is a discount with a job. The strategy question isn't "should we run a promo?" — it's which customer, at which moment, gets which code, and what stops the rest of your buyers from training themselves to wait for the next one.

Decide what the coupon is for

Every coupon should map to one of four jobs: acquire a first-time buyer, recover an abandoned cart, reactivate a lapsed customer, or move specific inventory. Mixing jobs is how margins quietly bleed. A 20% sitewide code blasted to your full list does all four badly — it pays existing customers to do what they were already going to do, and it teaches the rest to expect a discount on every order.

Pick the job first. Then the audience, the channel, and the discount size fall out of it. A first-purchase code goes to subscribers who haven't ordered. A reactivation code goes to lapsed buyers past a defined window. A clearance code goes only to inventory you actually need to move. None of them go sitewide unless that's the explicit goal.

Pick a discount size that protects margin

The biggest mistake is anchoring on round numbers — 10%, 20%, 25% — without checking what the math does to gross margin. A 25% off code on a product carrying a 35% margin gives away most of your contribution. A free-shipping threshold on a category with strong AOV often outperforms a percentage discount and costs less.

Percentage off — best for new-customer acquisition where the goal is the first order, not its size.
Dollar off above threshold — protects AOV and rules out tiny qualifying orders. Strong for cart recovery.
Free shipping — converts as well as a meaningful discount in many categories, with less margin damage.
BOGO / bundle — moves volume on slow SKUs without tagging the brand as discounted. See BOGO promotion design for the math.

Run the unit economics before the campaign, not after. If a 20% code can't pay back the next-purchase rate of the customers it acquires, the answer isn't to run it harder.

Distribute it where it converts

Distribution is where strategy survives or dies. A great code put in the wrong place becomes a public code by Tuesday — pasted into coupon-aggregator sites, scraped from referral links, and applied to every order including the ones that would have happened at full price.

Two patterns dominate. Public codes (BLACKFRIDAY, SPRING20) are loud, simple, and leak by design — fine when leak is the point. Unique codes are issued one per customer, expire on a rolling basis, and stay attached to the channel that earned them. The cost of unique codes is operational; the cost of leaky public codes is margin. Distributing digital coupon codes covers the gating patterns that keep codes inside the audience you meant to reach.

Build the mechanics that prevent abuse

Most coupon abuse comes from a small set of patterns: one customer creating multiple accounts to stack first-order codes, codes shared on aggregator sites, and stacking with other promotions. The mechanics that block these are well-known and worth implementing once.

One redemption per customer, enforced on email plus payment fingerprint, not just email. One-per-person coupon mechanics walks through the verification logic.
Hard expiration dates — never "no expiration" on a public code. A 7–14 day window concentrates urgency and limits leak damage.
Minimum order value where the discount math requires it. A 25%-off code with no minimum invites $8 trial orders.
Stacking rules written into the coupon tool, not into your customer service team's head. "Cannot be combined with other offers" is a policy; the cart enforcing it is the mechanic.

Set a cadence that doesn't train the wait

The deepest strategic question with coupons is cadence. Run them every month and your customers stop buying at full price. Run them only on holidays and you give up acquisition wins the rest of the year. The middle path: a steady stream of targeted codes that never go to the full list, plus two or three sitewide events a year that earn their margin hit by driving real volume.

Track redemption by audience segment, not just total redemptions. If your repeat-buyer segment redeems a code at 40%+, you're paying loyal customers to do what they'd have done anyway. Tighten the targeting. Cart abandonment coupon ROI covers the timing rules that prevent customers from learning to abandon every cart.

Measure incremental, not gross

Gross redemption rate flatters every coupon program. The number that matters is incremental revenue — orders that happened because of the code, minus orders that would have happened anyway. The cleanest way to estimate it is a holdout: send the offer to 90% of the segment, withhold from 10%, and compare order rates over the same window. The lift is your incremental.

The strategy in five lines: assign each coupon a single job, size it against margin not vanity, distribute it inside a controlled audience, enforce one-per-customer mechanics in the cart, and measure incremental lift against a holdout — not gross redemption.

Frequently asked

How often should we run sitewide coupon promotions?

Two to four sitewide events a year is a workable baseline for most brands — major holidays plus one or two strategic moments. More than that and customers learn the cadence and time their purchases to it. Targeted codes to specific segments can run continuously without the same training effect.

What discount size converts best for first-time buyers?

In most consumer categories, 10–15% off or free shipping above a threshold converts a first-time buyer about as well as deeper discounts, with materially better margin. Deeper discounts only outperform when the price barrier is the actual reason customers aren't buying.

Should coupons stack with other promotions?

Default to no, then make exceptions intentionally. Stacking is how a 20% acquisition code becomes a 40% giveaway when it lands during a sale. Build the no-stacking rule into the cart logic, not into your customer service playbook.

How do we know if a coupon is incremental or just discounting orders we'd have gotten anyway?

Run a holdout. Send the offer to a random 90% of the eligible segment and withhold from 10%. Compare order rates over the same window. The difference is your incremental — gross redemption is not.

Are unique codes worth the operational cost over public ones?

For acquisition and reactivation, almost always yes — they prevent leak to aggregator sites and let you attribute redemptions to channels. For loud, time-bound public events (Black Friday, a launch), a public code is fine because broad reach is the point.