Automated repricing on Wildberries and Ozon: which rules to use and where it loses money
How automated repricing works on Wildberries and Ozon: rules driven by minimum margin, competitor prices and stock levels, why repricing without unit costs runs into losses, API limits, platform promotions, and what to check before switching it on.
Short answer: automated repricing only earns anything if it has a floor. That floor is the minimum price below which an item is not sold, and it is calculated per item from unit cost, platform commission, logistics and advertising. Without it, a rule that says stay below the competitor ends with two repricing engines knocking each other’s prices down into a loss. Rules are usually built on three things: margin, stock and competitor prices. Before launch they are replayed against sales history, and for the first weeks they run on a small part of the range.
The price floor
The minimum price is calculated for every item separately. The platform’s commission depends on the category, logistics on dimensions and warehouse, advertising on what a sale through promotion costs. One blanket markup across the whole range will not do here: on a small item logistics eats more of the margin than on a large one.
If unit costs in your books are approximate, so is the floor. That is why repricing goes live after profit per item has been calculated properly. How to get there is covered in why the marketplace dashboard is not enough for margin analytics.
Three kinds of rule
From margin. The price holds a target margin when commission or logistics change. The platforms revise their tariffs several times a year, and without this rule an item quietly slips into a loss after the next revision.
From stock. When an item is running out and the next delivery is far off, the price goes up to stretch the stock. When an item has been sitting longer than it should and storage fees are accruing, the price comes down to the floor.
From competitors. The price follows comparable offers within set bounds. This is the riskiest rule: if a competitor has gone into a loss themselves, or is clearing stock, there is no reason to follow them.
The rules work together and in a defined order. The floor always sits above all of them.
Platform constraints
API limits. Prices are changed through the API, and it has request limits. Repricing thousands of items several times a day runs into them, and updates start queueing. What matters most is that zeroing out stock does not sit in the same queue behind prices. There is a detailed account in marketplace API limits and errors.
Platform promotions and discounts. The platforms offer promotions with a mandatory discount and show the customer a final price that includes their own discounts. Repricing has to know which promotions an item is in, otherwise it raises the price and the platform drops the item from the promotion.
The platform’s own view of the price. Ozon compares an item’s price with prices elsewhere and may show it less often if it is noticeably more expensive. A margin rule that pushes the price up has to account for that too.
How to test the rules before launch
Take sales and price history for past months and work out what prices the new rules would have set, and how much profit that would have produced at the same volumes. The calculation is rough, because it ignores the fact that buyers react to price. It does immediately expose the rules that drop a price below the floor or jerk it around ten times a day.
Then switch the rules on for ten to twenty per cent of the range for a few weeks and compare profit against similar items left without repricing.
Common questions
Is an off-the-shelf repricing service enough? For a small range on one platform, often yes. Your own module is needed when the rules depend on unit costs from your accounting system, stock in your own warehouse and several platforms at once. What such a back office costs is in what a marketplace aggregator costs.
How often should prices change? More than once every few hours is rarely worth it. Frequent changes burn API limits and make it impossible to tell which change moved the sales.
Who owns the rules? The person who owns profit for the category. Repricing carries out their decisions faster, the decisions stay theirs, and once a week they read the change log.
The rule
Work out the price floor per item first, then switch the rules on. Repricing without a floor cuts prices faster than any human would.
Related reading
Facing a similar problem?
Tell us what you are building. We will walk through the architecture and give you a budget range in one call.