Answer in brief
On Wildberries and Ozon your margin is moved by prices you never set. Listing SEO, an orders-and-stock export, a competitor radar and a Telegram price watchdog split the work of getting it back.
Your price is judged against prices you did not set
A marketplace seller does something no shop owner has to do: publish a price into a room where every competitor can read it, copy it and undercut it before lunch. Nothing about that number is private. On Wildberries and Ozon the figure that decides whether a unit moves this afternoon is not the one you typed in this morning, it is the distance between yours and the cheapest credible alternative sitting on the same screen, in the same grid, half a thumb-scroll away.
That is why marketplace work splits so cleanly into two piles. In the first pile is everything a seller decides alone: the words in the card, the attributes actually filled in, the photographs, the stock that genuinely exists in the warehouse, and the cost price that only you know. In the second pile is everything decided elsewhere: the ranking formula, the advertising auction, the commission schedule, the logistics tariff, and the number a stranger types into an identical listing at nine in the evening.
The four services in this article sit on either side of that line, and it helps to say which is which before talking about money. Listing SEO and the orders-and-stock export work on the first pile. The competitor radar and the price watchdog exist precisely because the second pile changes without consulting you. The catalogue files the first two under Marketplaces and the other two under Parsing and Monitoring, which describes the actual work more honestly than any promise about growth would.
Read the rest of this piece with that division in mind. Every time a paragraph promises something, ask which pile it belongs to. Optimisation of a card is an input you control and a result you influence. A competitor's price cut is an event you cannot prevent and can only meet faster or slower. Money spent on the first is an investment in position; money spent on the second buys reaction time, which is a different product entirely.
The listing decides which searches you are even in
Internal marketplace search is not Google, and treating it like Google is the first expensive mistake. A buyer types two or three words into the platform's own box and receives a grid of cards. Your item either enters that grid or, for that particular query, it does not exist at all. There is no second page of patient discovery, no long tail of referral traffic arriving from elsewhere, no way to be found by a phrase your card never contained.
The catalogue describes the outcome of listing SEO in exactly those terms: cards begin to show for a significantly larger number of internal search queries, so impressions and orders rise without the advertising budget rising alongside them. That last clause is the commercially interesting half. Paid placement also puts you in the grid, but it stops the moment the budget stops. A card that matches more queries keeps matching them while you sleep, and keeps matching them next month.
The work itself is unglamorous and mostly consists of filling things in properly. The title has to carry the phrases buyers actually type rather than the phrases your supplier uses on the invoice. The attribute and characteristic fields have to be complete, because empty fields quietly remove you from filtered results. Synonyms, plurals, common misspellings and the category the item genuinely belongs in all need deciding once and applying consistently across the whole range.
The catalogue prices this at $30 - $80 per card, comparing it to Amazon listing optimisation as the nearest equivalent market, and in roubles at from 1 500 ₽ per card, with market packages of roughly 38 000 ₽ for 1-5 SKU, roughly 58 500 ₽ for 6-20 SKU and roughly 76 500 ₽ for 21-40 SKU. Its difficulty label is not beginner: the catalogue marks this one as needing experience, which is a staffing note rather than a compliment.
What Wildberries and Ozon keep for themselves
The ranking formula is the platform's property and it is not published, not stable, and not negotiable. You supply inputs; the platform supplies weights; the weights change when the platform decides they should. This is not a complaint, it is a design fact of any marketplace, and a seller who plans around it does better than one who keeps hoping to reverse-engineer a permanent trick. Anything sold to you as a guaranteed position is being sold dishonestly.
The same applies to the machinery around the ranking. Commission percentages, logistics tariffs, storage charges, the rules of platform-wide discount campaigns, the mechanics of the advertising auction and the way returns are attributed all belong to the platform. They can move, and when they move they move for everybody at once. Your defence is not influence over these numbers, it is knowing your own cost structure well enough to see immediately what a change did to your margin.
Then there is the largest uncontrollable of all: the price on somebody else's card. A competitor can drop it at any hour, for any reason, including reasons that have nothing to do with you, such as clearing stock before a season ends. You will not be consulted, you will not be warned, and unless something is watching, the first evidence you get will be your own sales curve sagging several days later, when the cause is already cold.
Splitting the world this way changes where the budget goes. Money aimed at controlling the uncontrollable evaporates. Money aimed at two specific things does not: improving the inputs you own, and shortening the gap between an event happening and you knowing about it. The four services here are simply those two aims made concrete, one addressing inputs, the other addressing latency, and both priced in the catalogue rather than negotiated per conversation.
Twenty to forty minutes per SKU is a number about templates
The catalogue gives listing SEO a duration of 20-40 minutes per SKU after the template is set up, and the second half of that sentence carries most of the meaning. The template is the keyword map, the attribute checklist, the naming convention, the tone of the description and the rules for what goes where. Building it takes real thinking. Applying it to the fortieth card in a range takes considerably less, which is exactly why the figure is quoted per SKU.
Here is arithmetic any reader can repeat, offered as an example rather than as a measurement of a real project: a range of forty SKU at 20-40 minutes each comes to 800-1 600 minutes, or roughly 13 to 27 hours of card work once the template exists. That is a fortnight of evenings for an owner doing it personally, or two to four working days for someone doing nothing else. The spread between those two ends is entirely about how similar your items are to each other.
The two pricing formats in the catalogue reward reading together. Per card the price starts from 1 500 ₽; the packages are quoted at roughly 76 500 ₽ for a 21-40 SKU batch. Multiply the per-card floor by forty and you get 60 000 ₽, which sits below the package figure. That is arithmetic, not a hidden discount, and the honest reading is that the floor belongs to simple cards in a familiar category rather than to forty consecutive ones with attributes that all differ.
The practical consequence is that a range should be done as a range. Optimising six cards, waiting a month, then optimising six more means paying the template cost in attention three times over and ending up with three slightly different conventions in the same catalogue. If the budget only stretches to part of the range, choose the part by revenue concentration, not by which cards happen to annoy you most when you look at them.
Three cabinets, one sheet, and margin per SKU
Every marketplace shows a seller their numbers in the shape that suits the marketplace. Orders here, returns there, commissions in a report you have to generate, stock in a fourth place, and none of it in the same units or the same period as the others. A seller on two platforms therefore spends the first hour of the day reconciling instead of deciding, and by the time the picture is assembled it describes a situation that has already changed.
The catalogue's description of the export is deliberately concrete: one table instead of three personal cabinets, holding orders, sales, stock, commissions and returns across all platforms, with margin calculated automatically for each SKU, and the numbers already sitting there in the morning. Every noun in that sentence is a column somebody currently copies by hand. The automatic margin calculation is the load-bearing part, because it is the only figure that answers the question a price war actually asks.
Margin per SKU matters more than turnover per SKU for a reason that catches new sellers repeatedly. Commission, logistics, storage and the return rate can turn a headline price that looks profitable into a unit that costs money to sell. Without those deductions attached to each SKU, a decision to follow a competitor down looks survivable when it is not, and a decision to hold firm looks brave when the item actually had room to move all along.
The catalogue prices this at $120 - $450, or 8 000 - 35 000 ₽, with a duration of 1-2 days for two platforms, and marks it as needing experience. It also makes an unusually frank note about the market: cheaper versions exist on Kwork at 2 500-6 000 ₽, but those are a bare export, while on freelance.ru a genuine WB plus Ozon feed into a working table goes for 10 000 ₽ and upwards. The difference is whether commissions and returns arrive with the orders.
A daily radar answers what changed, not when
The competitor radar is described in the catalogue as a Google Sheet into which a matrix of competitors' products pulls itself every day: name, price, availability, link and the date of capture. Those five fields are chosen well. Name and link make a row actionable, price and availability make it meaningful, and the capture date is what stops an old row being mistaken for a current one three weeks later when somebody reopens the file.
The second half of the description is the part that makes it usable. Alongside the main matrix there is a changes-in-the-last-24-hours tab, so nobody has to read the whole sheet. This is the real product. A full competitor matrix is a wall of numbers that a busy seller opens twice and then stops opening; a short list of what moved since yesterday is something a person can genuinely read with coffee before the working day starts.
The catalogue puts this at $80 - $300, or 5 000 - 25 000 ₽, and it splits that band openly: a simple catalogue sits at 5-8 thousand roubles, while five or more sites with pagination and variations sit at 20-25 thousand. Duration is quoted as 4-8 hours for 3-5 sites with AI assistance, plus another 1-2 hours of anti-bot debugging. Difficulty is beginner and the category is Parsing, which is the plainest available name for the job.
The honest limitation is built into the word daily. A radar that refreshes once a day tells you what a competitor's price is now, but it can tell you about a cut up to twenty-four hours after it happened. For slow categories that is entirely adequate. For anything where a weekend promotion can empty your position before Monday, a daily cadence is a report, not a warning, and the next service exists because of that gap.
The watchdog is a message with four things in it
The price watchdog is described in the catalogue by quoting the message it sends: item X, was 4 900 ₽, is now 4 300 ₽, link. Four elements, nothing else. That restraint is the entire design. An alert has one job, which is to let a person decide in the three seconds between glancing at a phone and putting it back in a pocket, and every extra field in the message makes that decision measurably slower rather than better informed.
Three events trigger it, according to the same description: a competitor dropping a price, an item disappearing from availability, and a new SKU appearing. The second and third deserve more attention than they usually get. A competitor going out of stock is an opening, not a threat, and it is often the cheapest sales opportunity a seller will see all week. A new SKU appearing is the earliest possible sight of a competitor's next move, weeks before it shows up in your own numbers.
The catalogue states the value of speed without dressing it up: the client reacts within minutes rather than when they notice a dip in sales. That contrast is the whole argument for the service, and it is worth restating in your own terms. Reacting in minutes means you choose the response. Reacting after a sales dip means the market chose it for you, and you are now discovering the decision after the fact by reading a chart.
Pricing is $100 - $400, or 6 000 - 30 000 ₽, with a duration of 3-6 hours, and 1-2 hours on top if a parser already exists. Difficulty is beginner and the category is Monitoring. That conditional clause is worth planning around: as an example calculation, a radar at 4-8 hours plus 1-2 hours of anti-bot work, followed by a watchdog reusing the same parser for 1-2 hours, is a smaller total than commissioning the two separately.
Why the sheet usually goes in before the radar
Knowing that a competitor cut a price to 4 300 ₽ is useless if you do not know your own floor. The export is what supplies that floor, because it is the thing that attaches commission, returns and logistics to each SKU and produces a margin figure per item. Without it, an alert prompts a guess. With it, the same alert prompts a decision that takes seconds, because the number you need is already in the column next to the item.
The stated durations make the sequencing cheap to plan rather than dramatic. The export is quoted at 1-2 days for two platforms. The radar is quoted at 4-8 hours for 3-5 sites plus 1-2 hours of anti-bot debugging. The watchdog is 3-6 hours, or 1-2 hours if the parser is already there. None of these are quarters of a year, so the order should be chosen by dependency rather than by which one finishes soonest.
Listing SEO does not belong in that queue at all, and that is a feature. It is per-SKU work at 20-40 minutes a card after the template exists, so it runs alongside everything else without blocking anything. In practice a range can be worked through card by card while the export and the radar are being built, and the two efforts meet naturally when the first month of clean margin data arrives to show which cards deserved the attention.
There is one defensible reason to invert the order. If you already know exactly which competitor is taking your sales, and the only thing missing is the speed to respond, then the watchdog first is rational: it is the cheapest of the four in hours, at 3-6 hours, and it starts producing information immediately. Just be clear that you are buying reaction time before you have bought the knowledge of what a good reaction would be.
What pushes each job to the top of its price band
For the radar the catalogue tells you outright. A simple catalogue costs 5-8 thousand roubles; five or more sites with pagination and variations cost 20-25 thousand. The escalating factors are therefore countable before anybody quotes: how many sites, whether the listings paginate, and whether products carry variations in size or colour that have to be captured as separate rows rather than collapsed into one. Add the 1-2 hours of anti-bot debugging that the catalogue lists separately.
For the export the scope word is platforms. The stated 1-2 days covers two of them, so a third is not free. The fiddly part is almost never the orders, which arrive in a reasonable shape, but the commissions and the returns, which arrive late, in their own format, and sometimes attached to a period that does not line up with the one the orders came in. That reconciliation is what separates a bare export from a working table.
For listing SEO the scale is SKU count, which is exactly how the packages are drawn: 1-5 SKU, 6-20 SKU, 21-40 SKU at roughly 38 000 ₽, 58 500 ₽ and 76 500 ₽ respectively. Underneath that, the real variable is how alike your items are. Forty near-identical variants share nearly all their template work. Forty items across four categories effectively means four templates, and the time per card drifts towards the 40-minute end of the stated band.
The difficulty labels are worth reading as staffing advice rather than as marketing. The radar and the watchdog are both marked beginner, which is a reasonable signal that a capable in-house person could own them after handover. Listing SEO and the WB plus Ozon export are both marked as needing experience. That is not about the tooling being harder; it is about the number of quiet judgement calls that only show up as mistakes weeks later.
The part of the margin that is still yours
Strip out everything the platform owns and a surprising amount is left. You own the words in the card, and they decide which searches you appear in at all. You own the completeness of the attributes, which decides whether filters include you. You own your true cost per SKU, which decides how far down you can follow anybody. And you own the delay between a competitor moving and you knowing, which is the only one of the four that is bought purely in engineering hours.
Priced together, that is $30 - $80 per card for the listings, $120 - $450 for the export, $80 - $300 for the radar and $100 - $400 for the watchdog. In roubles the same four run from 1 500 ₽ per card, 8 000 - 35 000 ₽, 5 000 - 25 000 ₽ and 6 000 - 30 000 ₽. None of these numbers are large next to a month of marketplace turnover, and all four are quoted as bands, so the conversation is about where in the band your case sits.
The failure mode to avoid is buying the reactive half without the controllable half. A watchdog on a badly filled card produces fast reactions to a market you are barely visible in, which is an expensive way to feel busy. The reverse is safer but slower: good cards with no radar simply mean you find out about price moves later than necessary, and lose some margin you could have kept by responding on the day.
So the sensible reading of these four services is not a menu but a division of labour. Two of them improve what you put into the platform, at 20-40 minutes per SKU and 1-2 days for two platforms. Two of them shorten the distance between the market changing and you noticing, at 4-8 hours and 3-6 hours respectively. The platform will keep the rest, and it was never going to hand any of it over.
Questions and answers
How much does SEO for a Wildberries or Ozon card cost?
The catalogue prices listing optimisation at $30 - $80 per card, benchmarked against Amazon listing optimisation as the comparable market. In roubles it starts from 1 500 ₽ per card, with market packages of roughly 38 000 ₽ for 1-5 SKU, roughly 58 500 ₽ for 6-20 SKU and roughly 76 500 ₽ for 21-40 SKU. Production time is 20-40 minutes per SKU once the template is set up.
Can orders and stock from WB and Ozon go into one Google Sheet?
Yes, and that is exactly what the export service does: one table instead of three personal cabinets, carrying orders, sales, stock, commissions and returns across all platforms, with margin calculated automatically for every SKU and the figures already present in the morning. The catalogue prices it at $120 - $450, or 8 000 - 35 000 ₽, and quotes 1-2 days for two platforms.
What is the difference between the competitor radar and the price watchdog?
The radar is a daily snapshot: a Google Sheet with competitors' names, prices, availability, links and capture dates, plus a changes-in-24-hours tab, priced at $80 - $300 and taking 4-8 hours for 3-5 sites plus 1-2 hours of anti-bot debugging. The watchdog is instant: a Telegram message when a price drops, stock disappears or a new SKU appears, at $100 - $400 and 3-6 hours.
Do I need a developer for marketplace price monitoring, or can my team run it?
The catalogue marks both the competitor radar and the price watchdog as beginner difficulty, so an in-house person can reasonably own them after handover, especially since the watchdog needs only 1-2 hours on top when a parser already exists. Listing SEO and the WB plus Ozon export are both marked as needing experience, because the judgement calls there surface as mistakes weeks later.