Markdown and pricing optimisation
Mark down too early and you give away margin, too late and the stock won’t clear. Markdown optimisation finds the moment in between.
- Free 30-minute call
- No commitment
- You own the code
In short: A markdown optimisation plans when, by how much and in which stores to cut prices on seasonal or slow-selling stock. It uses how strongly each item’s sales react to a lower price. That way the stock sells out and you lose as little margin as possible.
The problem
Six weeks before the season ends, the summer range is still on the racks. Head office sends one rule to every store: 30% off the whole category this weekend. The city flagship sells items it would have sold at full price anyway. In the quiet store out of town, the racks are still full when the season closes.
Signs your markdowns cost more than they should
Markdowns follow the calendar, with the same date and the same discount everywhere. You give discounts on items customers would have bought at full price. Other items still don’t sell. They end up in the bargain bin, or you write them off as a loss. Nobody can say which stores needed the discount. And the end-of-season write-off is a surprise every year.
Why blanket markdowns fail
A markdown is a price cut on goods that aren’t selling. It’s how a retailer turns unsold stock into cash. It’s also where many retailers give margin away without noticing. Margin is what’s left of the sale price after the purchase price. A shirt bought for 12 francs and sold for 30 earns a margin of 18 francs. Mark down too early, and you discount items customers would have bought at full price. Mark down too late, and the season ends with stock that won’t clear at any price. One blanket discount across every store and product usually does both at once.
A blanket discount feels fair and is simple to run. But it gives too much discount in stores that would have sold the goods anyway. It gives too little in stores where the goods don’t sell. And it can’t plan prices week by week. So you give away margin in busy stores and keep unsold stock in slow ones.
What changes
Today, every store takes the same discount on the same weekend. With our model, each product in each store gets its own price for each remaining week. A line that sells well in the city keeps its full price. In the quiet store, the same item gets its discount earlier and a bigger one, so it sells out in time. You review the price plan and approve it.
Markdown optimisation in retail: what you get
Each item gets the markdown it needs, and no more. Sell-through is the share of stock you sell by season end. If you bought 100 shirts and sold 80 of them, sell-through is 80%. You see which products are heading for poor sell-through while there’s still time to act. The final clearance is planned, so the write-off stops being a surprise. You check each price plan against three figures. How much margin do the reduced goods earn? How much do you sell at full price? And how much stock is left at season end?
We build the model on your own sales and seasonal data and on your demand forecast. From that forecast, the model works out the expected sell-through per product per store. In one step, the model decides when to cut, by how much and how the price moves over the weeks. It works back from the date the goods must be out of the store. Every price it proposes keeps your rules for the old price, the crossed-out price you show next to the new one.
- A model written around your own markdown and pricing optimisation rules.
- Results measured in the numbers you already track.
- It slots into how you already work, and you can check every answer it gives.
How it works
Your stock per product and store, your past sales at different prices and the weeks left in the season go into one model. It sets a price for each product in each store for every remaining week. You get the price plan that clears the stock in time and keeps the most margin. You review it and approve it before any price changes.
Price elasticity by SKU and store
The starting point is price elasticity: how much more you sell for each step of discount. If 10% off makes you sell 20% more, elasticity is high. If it adds only 2%, elasticity is low. We measure it per stock-keeping unit (SKU), meaning each product in each size or colour, and per store. A fast-moving line in a city flagship may need no markdown at all. The same SKU can go unsold for weeks in a quieter store, because fewer shoppers see it there. Elasticity differs from product to product and from store to store. So good price optimisation software measures it for each product in each store.
Sell-through against the season target
With sell-through, time matters. Say a SKU has reached 60% of its seasonal target with three weeks left. It may not sell out in time. At 90% it will. Discounting it early then only gives margin away. Your demand forecast shows how each SKU will sell in each store if you change nothing. So the model knows which items are at risk. It leaves the rest alone.
Timing and depth, solved together
Timing answers when to cut. Depth answers how much. A small cut that doesn’t get the item selling only costs you margin. A deeper cut may sell the item out a week earlier. You get your money back a week sooner and can buy new goods with it. That can be worth more than the extra discount. The right depth depends on how much time is left. The right time depends on how deep the cut needs to be. So the model sets both at once.
Markdowns by SKU and store
You keep more margin when you decide per SKU and store instead of for the whole category. Think of a lakeside store and an inland one, or a tourist town and a commuter town. The price that sells out one store may leave goods unsold in the other. Sell-through optimisation in retail therefore sets a price per SKU and store. One discount for the whole category would be too deep in some stores and too shallow in others.
Multi-week price paths and end-of-season clearance
At the end of a season, you plan over several weeks. The model plans a price path, a sequence of markdowns over the remaining weeks. Each step is as deep as it needs to be to sell that week’s planned quantity. Whatever is left over goes into the next step. The last step is the final clearance. There an empty shelf matters more than margin. End-of-season clearance optimisation sets that final markdown so that as little as possible is left.
Reference prices on the price path
Markdown tools rarely mention a limit retailers can’t ignore: the rules for showing the old price. These rules say which old price you may show next to the new one, and for how long. This old price is called the reference price or “was” price. Your rules may differ between online and your stores. The model treats them as limits on the price path, exactly as you define them.
A path that only steps down usually keeps its reference price all the way to clearance. A short promotion just before the markdown can lower the “was” price you may show. Then the discount looks smaller on the sign and attracts fewer shoppers. The model counts that in before you plan the promotion. You see the cost of a price move before the sign goes up.
Why markdown and forecast belong together
One shortcut is tempting: take the demand forecast and add a rule with a fixed percentage and date. That fails. How deep you need to cut depends on the sales your forecast expects. And the forecast changes the moment you cut the price. So the model starts from your forecast. Then it decides in one calculation, across all stores, when to cut, by how much and how the price moves over the weeks. The same calculation works for fast-moving consumer goods (FMCG), everyday goods like food and toiletries. There too, you plan promotions and seasonal goods so that you keep your margin.
Common questions.
Something else? Ask us directly.
When your forecast shows the item won’t sell out at its current price in the weeks left. That date differs per item and per store. A fixed date like “six weeks before season end” is easy to run. But it comes too early for items that sell well and too late for weak ones. Compare every week how much you’ve sold with how much you planned to sell. Cut the price once an item falls behind.
You can, but it costs you later. The “was” price you may show often depends on what you charged in recent weeks. So even a short price cut lowers the “was” price you can show next. After a rise, that waiting time starts again: you can show the higher price as the “was” price only later. So plan your prices to go down only, unless a rise clearly pays for itself.
You need one for every group of stores where the same line sells clearly faster or slower. That’s rarely every single store. Say your city stores sell a line faster than your out-of-town stores. Then one price leaves goods unsold in one group or gives margin away in the other. Where you need one price everywhere, for example online, the model plans a single series of price cuts. It then moves stock between stores instead.
Sometimes it can, and if that’s the case, we’ll tell you. Standard software is built for an average operation, though. Yours runs on its own rules: your contracts, your exceptions, the things you know by heart. A custom model writes those rules in, and that’s where the better plan comes from.
You don’t need either to start. We build the model from the data you’ve already got, even if it lives in three spreadsheets and one person’s head. You get a system you run yourself. We train your team until they no longer need us.
You see a first result with the prototype, before you commit to anything bigger. We start with a free 30-minute call about your operation and the model that would fit it. Then we build a working prototype and run it on your numbers.
Similar pages.
Terms on this page
Every abbreviation explained in plain words, in our article “Metrics and methods”.
Tell us what you’re planning.
In a free 30-minute call with a specialist, you find out whether optimisation can help.
- Free, no commitment
- You talk to a specialist directly
- A clear go or no-go