First, measure the habit
A feeling that customers wait is not a diagnosis. Build three measures from the store's own orders. The share of orders placed with a discount code or at a reduced item price, by month. The same share for each customer cohort, grouped by the month of first purchase, to see whether customers acquired in a sale ever buy at full price later. And full-price orders in the weeks before each recurring sale, compared with the same weeks of earlier years, to see whether demand is being held back.
The records usually exist already. Google's developer guide to measuring ecommerce documents a coupon parameter on the purchase event and a discount field on each item, and the GA4 e-commerce tracking tutorial sends both. If they are empty, take the same fields from the shop platform's order export. Separate SeasonalityPredictable, repeating changes in demand tied to the calendar, such as months, holidays, weekdays or hours. from the habit: a seasonal dip lowers every kind of order together, whereas a learned habit shows as full-price orders thinning while discounted orders hold. Then read which codes and which routes the discounted orders came through: a sale event, a sitewide code, a cart email, a loyalty reward, or a customer who also appears on a marketplace. The route names the row.
Diagnosis table
Diagnosis
| Symptom | Check | Likely cause | Fix |
|---|---|---|---|
| Full-price orders dip in the weeks before a known sale every year | Sale and clearance dates across past years; full-price orders in the run-up | Sales and clearance on the same dates every year | Contain clearance to ending lines and stop fixing it to the calendar |
| Discounted orders include many products that never needed clearing | Which codes worked sitewide; share of discounted items that were clearance lines | Sitewide codes used to clear a handful of lines | Price clearance items directly and retire the sitewide code |
| Returning customers abandon baskets and order only after the cart email | Incentive branch of the cart flow; repeat customers redeeming its code | A discount in every abandoned-cart email | Write an incentive policy that reads the customer's history |
| Members' orders are mostly paid partly with points or member pricing | Reward redemptions against members' order frequency before joining | Loyalty rewards that work as a standing discount | Shift rewards towards perks and keep money off for real changes in behaviour |
| Shoppers mention a lower price elsewhere; branded marketplace sales rise | Prices of the same products on marketplaces and stockists' sites | The same products cheaper on a marketplace or through stockists | Hold one headline price and express the store's advantage another way |
| Sale messages lift orders less each time; complaints about prices that seemed made up | Evidence behind each was-price; whether countdowns restart or sales get extended | Was-prices and countdowns buyers have learned not to believe | Show only reference prices that were charged and keep every end date |
Fix, verify and prevent each cause
A sale customers can put in their diary
A clearance held on the same dates every year turns into a sale people plan around. The guide to clearing excess stock records, from the fashion industry guide, that discounts on predictable dates and at predictable depths teach customers to wait, and recommends keeping sale periods short and limited to genuinely ending lines.
Fix: stop announcing a store-wide event on the usual date; move ending lines to an outlet section reached from the navigation and offer existing customers early access rather than a deeper cut.
Verify: full-price orders in the weeks that used to precede the sale recover against the same weeks of earlier years.
Prevent: decide clearance from stock, line by line, in the buying review, so the calendar no longer sets the date.
A code that discounts the whole range
A code that works on everything turns a stock problem on a handful of lines into a price cut on the bestsellers too. The clearance guide's first containment rule is no sitewide code, with clearance prices set on the clearance items alone and labelled as ending lines, last sizes or graded stock.
Fix: retire general codes and price the lines being cleared directly.
Verify: the share of discounted items that are genuine clearance lines rises, and full-price sales of the core range hold.
Prevent: require a named product list for any new code, approved by whoever owns margin.
A cart email that pays people to leave
The abandoned cart flow tutorial puts it bluntly: discounts recover carts and train shoppers to abandon, and a regular customer who learns that leaving a basket produces a code will leave every basket. It asks for an incentive policy written before the final email is built, with positions such as no incentive, first-time shoppers only, or free delivery instead of money off, read from the customer's record.
Fix: remove the code for customers with order history and keep reassurance, such as delivery, returns and sizing help, in its place.
Verify: returning customers' orders that follow a cart email without a code hold up, and the flow's incentive cost falls against a holdout.
Prevent: review the policy against order value and margin every quarter, as the tutorial recommends.
Loyalty that is a discount by another name
A programme whose rewards are mostly money off, and mostly redeemed by people who were already buying, gives margin away without changing behaviour. The repeat purchase and loyalty guide calls that a discount by another name, and favours rewards that cost the store less than their perceived value, such as free delivery, early access or a members' sale.
Fix: move the reward mix towards perks and reserve cash-equivalent rewards for behaviour the store wants more of.
Verify: members' order frequency holds while the reward cost per order falls.
Prevent: judge the programme on incremental orders and payback, as the loyalty guide asks, never on enrolment.
A cheaper price one click away
If the same products sell for less on a marketplace or a stockist's site, shoppers who check learn to buy there or to wait for the store to match. The comparison of marketplaces and an own store recommends the same headline price everywhere, with the store's advantage expressed through bundles, loyalty, delivery thresholds and service rather than a lower sticker, and a deliberate subset of the range on the marketplace.
Fix: align headline prices across channels and give the store its own reasons to buy there, such as exclusive sets and new launches.
Verify: customers who appear in both channels start buying from the store at full price again.
Prevent: check channel prices whenever the store's prices change, and agree price terms with stockists in writing.
Price claims nobody believes any more
A was-price that was never really charged, or a countdown that restarts on the next visit, teaches buyers that the store's urgency is theatre, so they wait for the next one. The guide to using urgency, scarcity and defaults honestly records from the site's pricing guide that in the UK the Competition and Markets Authority enforces consumer protection law covering reference prices that were never genuinely charged and urgency claims that are not true; it treats a deadline as a statement of fact that is kept. This is educational, not legal advice, and other markets have their own rules.
Fix: show a reference price only where the item genuinely sold at it, with the evidence on file, and end each sale when it says it ends.
Verify: the sale's lift holds steady across events instead of fading, and price complaints fall.
Prevent: route every price claim through compliance review before a promotion runs.
A prevention routine against the discount habit
The habit builds slowly, from decisions that each looked sensible, so the routine has to watch the total. Each month, read the share of orders and of returning customers buying with a code or a reduced price, by cohort, beside full-price sell-through of the core range. Keep a register of every live discount: who can use it, on which products, until when, and why it exists. A code nobody can explain is usually the start of the next habit.
Before any promotion, check its timing against the last few years so that it does not land where customers already expect it, list the products it covers, and record the evidence behind any reference price. After it, measure what it cost on Contribution MarginRevenue left after every cost that varies with each sale., including the full-price orders that might have waited for it; the guide to measuring a sale or promotion separates incremental orders from pull-forward. To check whether fewer personal codes would pay before cutting them for everyone, the full-price protection test is a design that compares the current offer rhythm with offers at planned events only on customers assigned at random. A basket that keeps its full price but shrinks is a different symptom, covered on the page on a dropping average order value, and the other store symptoms sit in the e-commerce troubleshooting section.
Gloamtwig Candles breaks its customers' waiting habit
Gloamtwig Candles, an invented candle maker, sells through its own store. Over a year, 58% of orders carry a code or a sale price, and for customers in their second year the figure is 81%. Full-price orders in the three weeks before its fixed January and July sales run 34% below the same weeks after them. Reading the coupon parameter by route, 41% of discounted orders come from the two sales, 27% from the third cart email's 15% code, which returning customers redeem as often as new ones, and 19% from loyalty points worth £5 per £50 spent. The maker drops the fixed sale dates and clears ending scents through an outlet page, with past buyers told first. It removes the cart code for anyone with an order history and keeps it, as free delivery, for first-time shoppers. It replaces points with early access to new scents and free delivery for members. Over the following two quarters, the share of discounted orders falls to 33%, second-year customers' full-price share rises from 19% to 46%, and gross margin per order rises by £2.80, while total orders fall 4%.
Common mistakes
Cutting the depth of every sale while keeping its date. Removing the cart code for new shoppers along with returning ones. Fixing the store's prices while a marketplace still undercuts them. Inventing a was-price to make a smaller discount look larger. Measuring a promotion on revenue and ignoring the full-price orders it displaced.
