Is your welcome discount building customers or bargain hunters?

The pop-up went live on a Monday: "Get 15% off your first order." By Friday, 70% of first orders were using the code, conversion rate on the homepage was up, and the email list was growing faster than it had in a year. The founder sent a screenshot to her co-founder with one word: "Working."

Ninety days later, a different picture. The customers who had used the code were coming back at roughly half the rate of the ones who had paid full price. When they did come back, most of them waited for the next sale. The code had worked brilliantly at the one thing it was measured on. It had also changed the kind of customer the brand was acquiring, and that cost did not show up on any dashboard until a full quarter had passed.

That is the problem with welcome discounts. They are judged on day one and paid for on day ninety.

What the research says about acquisition discounts

This is not just an agency hunch. Marketing researchers have studied it directly. A well-known study by Michael Lewis, published in the Journal of Marketing Research in 2006, looked at customer-level data from a newspaper and an online grocer and asked a simple question: do customers acquired with promotions behave differently?

They do. In both businesses, the deeper the acquisition discount, the lower the repeat-buying rate and the lower the long-term value of the customers it brought in. In one of the datasets, customers acquired with a 35% discount were worth roughly half as much over time as customers acquired without one.

The explanation is intuitive once you see it. A discount does two things at once. It persuades some people who would not otherwise have bought. It also selects for people who are more motivated by price than by the product. The first effect lifts your conversion rate. The second quietly changes the mix of your customer base.

Four ways a welcome code costs more than the discount

When founders think about the cost of a code, they think about the margin given away on the first order. That is the smallest of the four costs.

Leakage deserves special attention because it is often large and almost never measured. A shared code like WELCOME15 gets picked up by coupon aggregators within days. Browser extensions then automatically try it at checkout for anyone who has them installed, including loyal customers placing their fifth order. Every one of those redemptions is pure margin loss: these people were already buying.

The check: does your code build customers or bargain hunters?

You can answer this with the data you already have. Here is the method.

1. Pick a cohort. Take every new customer from a three to six month period that ended at least 90 days ago, so everyone has had time to come back.
2. Split them in two. Group A: first order used a discount code. Group B: first order at full price. If you run several codes, keep the welcome code separate from influencer and sale codes.
3. Measure four things for each group: repeat rate at 60 and 90 days, the share of repeat orders that also used a code, average order value on repeat orders, and 90-day contribution per customer after product cost, shipping, fees, discounts and returns.
4. Check the mix. Make sure both groups bought similar products first. If code users mostly bought a cheaper entry product, part of the gap is the product, not the code. Our guide to customer value by first product shows how to control for this.

Here is what the result looked like for the brand from the opening. It is an illustrative composite, but this shape is very common.

At day 30, the gap looks small, which is why most brands never notice it. By day 90 it is two to one. By day 180 the full-price cohort has nearly twice the repeat rate, and most of the code cohort's repeat orders landed during sale events.

How to read your own result

• Similar repeat rates in both groups. Your code is doing its job. It is converting people who behave like everyone else. Keep it, and focus on reducing leakage.
• Code group repeats less, but at full price when they do. The code is pulling in lighter buyers. Consider a smaller code or a non-price offer.
• Code group repeats less and mostly with more codes. The code is shaping your customer base around discounts. This is the pattern to fix first, and it is closely tied to discount dependency.

When a welcome discount is the right call

None of this means a welcome code is always wrong. There are situations where it earns its cost, and it is worth being honest about them.

• Durable or one-time purchases. If your product is a mattress, a blender or a piece of luggage, there is little repeat behaviour to protect. The code's only job is to win the first and probably only sale, and a code that lifts conversion can pay for itself immediately.
• A brand nobody has heard of yet. With no reviews and no word of mouth, a first-order incentive reduces the risk a stranger is taking. As social proof builds, the code usually becomes less necessary, and that is the moment to test removing it.
• List building is the real goal. If your email and SMS flows are strong, a subscriber can be worth a lot even before they buy. In that case the offer is paying for a contact, not a sale, and should be judged on revenue per subscriber over 90 days.
• High-intent paid traffic with a low conversion rate. If you are paying $3 a click and converting at 1%, a small code that moves conversion to 1.4% can lower your acquisition cost by more than the discount costs. Check the repeat rate before making it permanent.

The common thread: the code should be a deliberate tool with a measured job, not a default that every DTC brand copies from every other DTC brand.

Why codes get more expensive every year

Two trends are making shared welcome codes leakier than they were even a few years ago.

First, coupon tools have moved into the checkout itself. Browser extensions automatically search for and test codes on almost every store, and AI shopping assistants increasingly do the same when they help someone buy. A code that was meant for new subscribers ends up being applied by a tool on behalf of someone who never saw your pop-up.

Second, customers have learned the pattern. Many shoppers now know that most DTC brands offer something for an email address, so they sign up with a new address on every order, or abandon the cart and wait for the "you left something behind" email with a better offer. If your abandoned-cart flow includes a discount, you are teaching exactly that behaviour.

The fix for both is the same: unique, single-use codes tied to one email address and to customers with no prior orders, and no automatic discount in the first abandoned-cart email.

Five alternatives worth testing

Removing the welcome offer entirely is rarely the right first move: email capture usually drops. The better move is to change what the offer is. These are the five alternatives we test most often, with the trade-offs.

The second-order discount is the most underused option on that list. It costs nothing on customers who never come back, and it pushes customers towards the behaviour that actually builds value: the second purchase. It can be sent in the delivery confirmation or a few days after the first order arrives, once the customer has tried the product.

Here is how the five options performed in a 90-day test for the illustrative brand, measured on what really matters: contribution per new customer.

How to run the test properly

Most welcome-offer tests fail because they are judged on the wrong number. A pop-up with "20% off" will almost always win on sign-ups and first-order conversion. That is not the question.

1. Split traffic, not time. Use your email platform's form A/B testing to show different offers to different visitors at the same time, rather than running one offer in March and another in April.
2. Use unique codes for each subscriber, limited to one use and to customers with no previous orders. This removes most leakage and makes the comparison clean.
3. Decide the success metric before you start: contribution per visitor who saw the form, over 90 days. That captures sign-up rate, conversion rate, order value, discount cost and repeat behaviour in one number.
4. Wait for the repeat window. Read the first signals at 30 days, make the decision at 90.

You can read the full story in our case study on the welcome code.

The bigger point: your offer is a filter

Every acquisition decision filters who becomes your customer: the channel, the creative, the product in the ad, and the offer. The welcome discount is the one that most brands set once and never revisit, even though it touches almost every first order.

If you only change one thing after reading this, run the cohort split above. It takes an hour, and it tells you whether your welcome offer is building the business or renting it. If you want it done across all your codes, channels and products at once, that is exactly what a Blindspot Snapshot covers.