How Ecommerce Teams Protect Margin With Smarter Promotions

Honestly we learned this one by accident. We'd run a 20% off promo on handwritten card orders and the numbers looked great, until I pulled the customer list and saw most of the buyers were existing accounts who order every month anyway. We basically paid our best customers to do what they were already going to do.
Now the rule is simple. Before any discount goes out, we ask who it's actually for. If the answer is "people already on the fence who haven't bought in 60+ days" or "brand new accounts testing us out," fine. If it's a blast to the entire list, it almost never earns its margin back.
A few things that changed for us:
First, we split promos by segment. Lapsed customers get the offer, active ones don't.
Second, we stopped leading with price for loyal accounts and started sending them a handwritten thank you note instead, no offer attached. It costs us a couple bucks and retention went up, not down.
Third, we compare promo week sales to the same week without a promo, not to the week before. That one comparison killed about half our discount ideas.
The lesson I'd give anyone: if you can't name the customer who wouldn't have bought without the discount, you don't need the discount.
Rick Elmore, Founder and CEO, Simply Noted

Don't discount what already sells

In our business, we learned fairly early that discounting isn't always the best way to close a deal. Explainer videos are a considered purchase, so someone rarely chooses us simply because the price dropped for a week.

Instead, we've found it more useful to add value around the service, such as a consultation, script support, or helping clients choose the right production approach. It protects our pricing while giving prospects a reason to move forward.

That all said, the lesson I'd reuse is that if the customer already sees the value, don't train them to wait for a discount.

1. Test a holdout group to verify the discount you are giving away isn't just a subsidy to people who would have bought anyway. At TradingFXVPS we keep 10-20% of our traffic in a holdout group and only they see the promo offer. I also like to check who actually redeemed the code(s) and what behavior they were doing before they applied the promo. Was it mostly return visitors who already had a plan in their cart? Or did the code show up on coupon sites within 48 hours and proceed to extract margin from us? If so, then the promo was probably just another form of margin leak. Only a promo that changes a specific behavior (i.e. monthly subs converting to an annual plan, new traders buying a 2nd server location, etc.) is worth running in my book. If I cannot identify the specific behavior, then I won't run a promo for that.

2. Discount capacity not price! Our customers care about latency and up-time, not about saving 3 dollars. So instead of 30% off the yearly price for a VPS I opted to bundle in some upgrades, like more RAM, a higher CPU core count, or an extra month free on top of an already paid for yearly subscription. Extra RAM on a node that I already pay for costs me almost nothing. A price cut on the other hand costs me on every single renewal, because humans anchor to the lower price and suddenly become very inflexible when the price goes back up to where it was before. To make sure our customers get maximum value out of their VPS I made two simple hard rules for myself: never ever discount the renewal price for a VPS and never ever repeat a promo on the same date in the same year.

We stopped measuring promotions by sales volume during the sale and started measuring incremental lift against a matched control group that saw no discount at all. Most of our "successful" promotions were really just giving away margin to customers who would have purchased regardless.
The lesson that protected margin: before our last major sale, we split our list: 80% received the standard 25%-off promotion, 20% received no offer and just a normal product email. Purchase rate in the discounted group was 14%, but the control group still converted at 9%, meaning the discount only drove about 5 percentage points of truly incremental revenue, far less than the margin given up across the full 25% discount.
We shifted to smaller, targeted discounts, around 10%, reserved only for cart abandoners and lapsed customers. Overall promotional spend dropped by roughly 38%, while incremental revenue attributable to discounts rose by 12% compared to our old storewide approach.
The lesson: always test against a no-discount control. Without it, you can't tell a promotion from a subsidy.

I decide a discount is worth testing only when it has a clear job beyond creating a temporary sales spike. It should bring back a cart abandoner, help move a specific item, or give a hesitant customer a reason to complete an order. If I cannot name the behavior it is meant to change, I would not offer it.

The lesson I learned in my jewelry business came from raising prices, not from a discount test. Sales dropped sharply, especially while buyers were already stretched by inflation. That made me careful about treating price as a simple lever. A promotion that merely lowers the price for customers who were ready to buy anyway gives away margin without creating new demand.

My practical rule is to keep the offer narrow, set the margin floor before it goes live, and compare the order pattern with a normal period. If the offer only shifts purchases forward or trains customers to wait, I would stop it. Customer interest matters, but a promotion should earn its cost.

The first question I ask before approving any promotion is whether the discount is buying a new customer or paying an existing one to do what they were already going to do. In my own stores, a sitewide sale used to look like a win while quietly funding repeat buyers who had items in cart at full price, because I never separated those two buckets.

So I quit blanket discounts and instead gated them behind a behavior the customer hadn't already shown. Someone who has browsed three times without buying, or hasn't ordered in months, is a different animal from someone who opens every email and buys anyway. The second group got early access, bundles, and new product drops instead of money off.

That single change protected margin in a way I didn't expect. My loyal buyers didn't need a coupon to stay interested, they needed to feel ahead of everyone else. And once discounting was reserved for cold or hesitant segments, I could read the numbers, because a promotion's lift stopped being contaminated by sales that would have landed regardless.

When I work through this with operators in my group program, the hardest part is accepting that a quieter promotional calendar usually produces better economics. Fewer sales, aimed at people who needed a reason, beat a permanent discount habit that trains everyone to wait.

Promotional strategy is often framed as a pricing decision, but it is really an alignment test. If finance wants margin, marketing wants volume, and operations is managing capacity separately, discounts become political rather than commercial. A worthwhile offer has a shared definition of success before launch, including what outcome would prove it was incremental and what outcome would require it to stop.

We learned that post-promotion review is where margin is truly protected. Reviewing results without blame encourages teams to surface uncomfortable findings, such as cannibalization or poor repeat behavior. That psychological safety matters because people are more likely to correct a weak assumption early. The strongest promotional programs are not built on flawless forecasts. They are built on honest learning and disciplined adjustment.

As the owner of The Monterey Company, a custom branded merchandise company founded in 1989, I look at promotions through one question. Would this order have happened without the discount? If the buyer was already coming to us with an event date and a design, a discount mostly gives away margin.

In my experience, a promotion is worth it when it changes behavior. That might mean getting a buyer to order sooner, choose a larger quantity, or try a product they haven't ordered before. In custom products, quantity price breaks already reward bigger orders, so I would rather lean on that than run blanket discounts that train buyers to wait for the next sale.

The lesson I would share is that many buyers care more about clarity and confidence than price. A clear quote, fast proofing, and a reliable delivery date often do more to win the order than a percentage off.

My test is pretty straightforward: would this particular customer have paid the full price in the next two weeks regardless of the discount? Then, you're effectively paying that customer to do something they were going to do anyway. What made us smarter with our discounts was that we made sure that the discount actually induced a behavior change in the form of a first-time buy, referral, or bundling a low-margin item with a high-margin one, and not just sending out that generic 20 percent off to our whole list when we had low sales again. Just that change alone reduced the number of discounts by more than half without reducing repeat rates, since those customers trained to wait for a sale are always going to wait for a sale.

I believe in keeping our margins rather than falling into the trap of short-term volume growth to satisfy quarterly earnings goals. I prefer non-price over price promotions, and have found good success using non-price promotions like early access to product launches. Our customers appreciated the early access. Compared to price promotions, this earned us better engagement and stronger loyalty. Most importantly, it allowed us to maintain a competitive margin while rewarding customer loyalty.

When promotions start pulling forward sales that would have happened anyway, I treat the discount like a tool you earn the right to use, not a default setting. I decide it is worth offering when it is tied to a clear moment and message, and when I can test discount depth against margin instead of judging it only by top line revenue. If a smaller offer paired with a strong theme performs well, I would rather take that than a bigger discount that teaches customers to wait. One lesson from our own strategy is to protect attention like it is inventory, because if every email is only a deal, the list gets trained to ignore anything that is not a markdown. We segment by intent so warm shoppers who are already close to buying get more timely nudges, while colder subscribers get more spacing and stronger creative. That helps us avoid blasting a discount to everyone when only a small group needs it. The goal is to make the offer feel like an event with a reason to act, while keeping the brand from becoming only a price tag.

A discount is worth offering only when the new orders it creates are worth more, after margin, than the full price orders it gives away. On ecommerce accounts I buy media for, promo conversion rate jumps because people already about to buy take the deal and deal seekers show up beside them. Fence the offer so new and lapsed buyers still get the discount while returning buyers pay full price, then judge it on incremental margin. The promotion with the best conversion rate is often the one eating margin, since you are paying to discount demand you already owned. This fails when the buyer has never paid full price, because then the discount is the only way in.

We found that the timing of an offer often matters more than its size. Early broad discounts encouraged shoppers to delay their purchase instead of deciding with confidence. This reduced urgency and made full price purchases less likely over time. We shifted to shorter campaigns linked to clear customer behavior and tested each response carefully.

Before launching any offer we create a clear exit rule for every campaign decision. We define the results needed in advance before continuing the offer across future campaigns. We review incremental contribution repeat engagement and new customer participation before making the next move. If those signs stay weak we end the offer because lasting value matters most consistently.

I offer a discount only when it is likely to change behavior, not reward someone already ready to buy. We test this with holdout groups: compare a similar audience that sees the offer with one that does not, then measure incremental orders, gross profit, and repeat purchase—not revenue alone. Incrementality testing is designed to measure outcomes caused by marketing exposure rather than simply crediting transactions that happened afterward.

The lesson that protected margin was replacing broad discounts with intent-led alternatives. For hesitant customers, we used back-in-stock alerts, wish-list reminders, bundle value, free shipping thresholds, or early access before considering a percentage-off code. Customers who had already viewed products repeatedly often needed reassurance or timing, not a lower price.

My rule became: every promotion needs a hypothesis, a margin floor, an excluded purchaser audience, and a post-campaign holdout comparison. If the discount only shifts full-price orders into a cheaper window, it is not growth.

From the deal-site side, the discounts that actually change behavior are the ones that stand out against what's normal for that category.

At WhatNotSell we only count a discount when there's a genuine higher original price behind it, and right now we track about 3,800 of those from 49 stores. The median real discount is 31%, but it varies hugely by category: laptops sit around 14 to 17%, while fragrance is close to 70%. A 20% laptop discount is a genuine event. A 20% fragrance discount is below what shoppers already see every day, so it mostly subsidizes people who would have bought anyway.

The lesson for protecting margin: benchmark your discount against your category, not against your own full price. If your offer isn't clearly better than the typical real discount in that category, it probably won't bring in new buyers. It's better to discount fewer items, less often and more deeply than to run a constant shallow discount that just trains customers to wait.

I'm Charles Liu, founder and Marketing Director of Cubic Promote, a Sydney-based business employing around 30 people across Australia, the Philippines, Vietnam, and India.

We try not to discount an order unless the offer is actually going to change something. If the customer was already likely to place the same order at the same time and quantity, giving them 10% off might make the sales report look good, but we've really just given away margin.

We look at whether the promotion can move an order forward, increase the quantity, bring back a customer who hasn't ordered for a while or help shift a particular product. In marketing, that's called incrementality. Put simply, we're asking how much extra business the discount created that we wouldn't have received anyway.

We've learnt to put conditions around promotions rather than making them blanket offers. For example, a minimum order value or quantity threshold gives the customer something worthwhile while also giving us a commercial reason to offer the discount.

That's particularly important in promotional products because many customers already buy in bulk. If someone was going to order 500 units anyway, discounting those same 500 achieves very little. If an offer makes 750 units more attractive and the numbers still stack up, that's a much healthier promotion.

We also look past the revenue spike. Margin, average order value and what the customer does after the promotion matter because a campaign can generate plenty of sales and still be poor business if customers simply learn to wait for the next discount.

We've become much more comfortable holding back when the offer isn't doing a specific job. Discounts work best for us when they're a nudge, not the normal price with a countdown timer attached.

The question behind every promotion is incrementality: would this unit have sold at full price anyway? On Amazon that is harder to see than on your own site, because the products the deal tools want to feature are your best sellers, and those are exactly the products that need discounting least.

Our rule is to discount in order to buy something, not to buy volume. A discount earns its margin when it purchases a specific asset: velocity to break into organic ranking on a keyword you are stuck below, review acquisition on a new ASIN with no social proof, clearing units before they cross the 181 day aged inventory surcharge tier, or defending a Subscribe and Save base that would otherwise churn. Each of those pays out after the promotion ends.

If the only thing the discount buys is this week's unit count on a product already ranking and converting, you are moving demand forward and paying twice: once on the discounted units, and again when the post-promotion dip makes the next period look weak and someone panics into another deal.

The lesson that protected margin most was dropping blanket percentage offers and moving to targeted coupons on specific ASINs with a redemption cap. A coupon puts a clickable badge in search results, so it buys visibility as well as price, and the cap bounds your exposure. A catalog-wide 20 percent off buys neither.

One more habit: judge a promotion on the two weeks after it, not the deal window. If units settle below where they started, you cannibalized rather than expanded, and the deal report will never tell you that on its own.

After years in ecommerce, I stopped judging promotions by the revenue spike. A discount is worth offering only when it changes behavior I actually want: brings in a new customer, raises basket size, moves aged inventory, or produces a second purchase sooner.

The lesson that protected margin was to stop running sitewide discounts for everyone. Existing customers who were already ready to buy would simply wait for the next code. Instead, I used narrower offers—first-order incentives, bundles, quantity thresholds, or discounts tied to specific inventory—and compared them with a non-discounted group.

I look at contribution margin after product cost, shipping, payment fees, advertising, returns, and the discount itself. If the promotion lifts orders but does not improve total contribution margin or customer quality, it is not incremental growth; it is just paying customers to do what they already planned to do.

The best promotion has a specific job and a defined stop condition.

Skip this if your only lever is another percent off. Mine stopped being that when the same baskets arrived with a code on days they would have checked out anyway.
The lesson that protected margin was to spend the quiet weeks on free UK delivery over £25, a published price match, and the AI Curl Analyser or research that rematches pattern and porosity, then keep cut-price rare. A discount still goes out when it defends a real price promise. It does not go out to invent urgency for a jar that should sell on fit. I still feel the itch to hit send when a week looks soft. Hitting send trains shoppers to wait, and the P&L pays for that habit longer than the bump lasts.
In The UK Curl Report 2026, https://zenvy-beauty.com/blogs/news/uk-curl-report-2026, UK women spent £416 before finding a routine that works. Shortening that path earns more than training them to hunt the next code.

A discount is worth it only when it changes a decision that would not have happened at full price.

When promotions start pulling forward sales you would have captured anyway, I stop looking at revenue on promo days and look at incremental orders versus a quiet baseline week, plus margin after the discount. One rule that protected margin for me is tying every code to a friction the customer names (first order, lapsed 90 days, cart abandon) and refusing sitewide cuts that teach buyers to wait. If the same cohort buys without a code after you remove it, the promotion was a habit tax, not a growth lever.

We do not judge a promotion only by revenue created while active. We look at the weeks that follow, in renovation categories where schedules and installation dates shape buying decisions. Comparing performance before, during, and after an offer gives us a clear view of demand. We review cancellations and order quality to see whether the promotion created sales.

The goal is to avoid making discounts feel normal. If shoppers can predict the next offer, they may delay purchases until prices drop. We prefer promotions tied to moments, such as seasonal planning or inventory changes. This creates urgency without teaching customers to wait for a discount.

We decide by looking at dollars, not percentages. On big-ticket items our margin is thin, so a "small" code can wipe out most of the profit. A 5%-off sign-up code sounds harmless, but on a brand we sell at roughly a 10% margin, it cuts the profit on that order about in half. We saw that on two September orders that used our 5% code.

The lesson that protected margin: offer a perk instead of a percentage on large orders. Right now, every product priced $1,000 or more on our site comes with a free 8-piece towel set, about a $35 value. On a $1,200 vanity, that costs us less than a 5% discount ($60), and shoppers still feel they're getting something extra.

We also only run real price cuts when the manufacturer funds them, so the sale lowers our cost as well as the shopper's price. And when a brand's dealer terms don't allow checkout discounts, we leave that brand out of the code rather than risk the relationship.

Kim Smith, President & CEO, Bath Supply Pros (bathsupplypros.com)

I only discount for one moment: the first purchase. Shuven runs one promo, ever, a 10% welcome code for new customers, and that's it. No flash sales, no seasonal markdowns, no stacking codes. The test I use is simple: if the discount isn't the reason someone tries the product for the first time, it's not earning its keep, it's just training existing customers to wait for the next one.

The margin problem is real for me because my formulas run full clinical doses with no proprietary blends, so the ingredient cost is already tight. A deep or recurring discount doesn't come out of some padded markup, it comes straight out of the ingredient budget. That's the lesson: tie your one allowed discount to acquiring a customer you didn't have, not to moving units you'd have sold anyway. After that first order, I lean on subscriptions to keep them, not another coupon.

Ryan Pepple, Founder, Shuven, https://shuven.co

How I decide: before running a discount, I compare the item's normal weekly sales with what we'd expect from the promotion. If the product already sells steadily at full price, like everyday grocery staples, a straight percentage off mostly gives margin away to customers who were buying anyway. A discount is worth it when it changes behaviour: it brings in a new customer, moves them to the app, grows the basket, or clears stock that would otherwise expire.
One lesson that protected margin: put conditions on the deal instead of cutting the price for everyone. In grocery retail and e-commerce in the UAE, I've seen bundles (buy 2, save), minimum-basket offers and app-only deals keep customers just as engaged as a flat discount, while costing far less, because the regular shopper still pays full price for a single item and the offer rewards the extra purchase.
Practical tip: after every promotion, check sales for the two weeks after it ends. If they drop below normal, the promotion mostly pulled forward demand you already had, and next time it should be smaller or more targeted.
Abdul Hakeem, E-commerce Executive, Pasons Group | Founder, Cortex IDE (cortex-ide.app)

In fast fashion you always have winners and losers, so we judge discounts at the portfolio level, not item by item. Our winners carry enough margin that we can clear a true loser even below cost, because dead inventory sitting on the shelf is worth less than the cash and the space it frees for the next arrival. The lesson that protected our margin was discipline on cadence: regular and new styles stay full price until they have been around a while, and the deep markdowns on slow movers only come a few times a year around special holiday sales, never on a regular basis. That keeps customers from learning to wait for a sale on new arrivals, so demand stays strong exactly where our margin lives. A discount is worth it to clear a genuine loser or to earn a first-time customer who comes back, never to manufacture a sale that would have happened at full price anyway.
-Whitney O'Keefe, co-owner, Avery Mae Boutique (averymae.com)