Why the old growth model stopped working
Let us be honest: the old growth model has broken. Not long ago, stores were running at 25, 30, 40 percent profitability and more. Today real seller profitability varies, of course, but it sits somewhere between 5 and 15 percent.
A few sellers move more volume, a few keep better margins, but those are exceptions. For many, profitability is lower, zero or negative.
At the same time, competition and marketplace commissions keep rising. Over the last three or four years commissions have gone up while our profitability has gone down. Competition is not necessarily exploding, but it is certainly not falling.
Lower margins used to be offset by growing turnover. Today the only resource left for protecting profit is a radical cut in costs. Turnover for the vast majority of sellers is no longer growing either: it is flat or falling.
The real difference is automation
The difference is not in people or in approach. Compare what we do with what sellers do on Amazon in the US. The difference comes down to one word: automation.
Every store today runs on people: managers, analysts, operations staff, designers and other specialists. I recently audited sellers with turnover of around 3 to 4 million per month, and that store was run by six people. At modest turnover we carry very large teams, and those teams often eat the small percentage of net profit that was left.
On Amazon, a large share of processes is already automated. Sellers do very little by hand and let automation do a lot for them. We, by contrast, try to do everything ourselves.
But there is another way to grow. The point is not to say we are doing something wrong. The point is to show that you can structure the work differently and raise net profitability not by cutting people, but by automating the processes that are inefficient or eat time and money and that can technically be automated.
1. Repricing and price monitoring
The first process that is often automated abroad is repricing and price monitoring. Across all the audits I run, sellers who use any kind of repricer are on average 1 to 2 percent of all sellers. The majority still do it by hand.
A manager or the seller sits down every day and monitors prices: what competitors did, how they changed their price, whether their marketplace-funded discount went up or down. Then the seller changes their own price too: once a day, more often, or every two days. Prices move quite regularly.
What does that lead to? When we constantly change something, we lose boosts.
The marketplace sees constant changes in product cards, sees that the card is unstable, that something is always changing, and stops giving price boosts. Conversions fall, positions drop, sales volumes decline, ads perform worse, and the store stagnates.
How it works abroad
Abroad there are various tools. One example is Seller Snap, a US tool that tracks prices every minute and, using game theory, looks for a stable point of maximum margin. It does not dump prices; it finds the specific price for your product that stays within the right range.
These tools are built for Amazon, Walmart and eBay, not for the Russian marketplaces. What matters is the idea: automation is already possible at this stage, you just have to think about it.
The tool promises advanced AI for repricing on Amazon and Walmart with built-in seller analytics: save time, avoid price wars, maximize profit. How familiar that price-war and dumping problem is.
It uses AI strategies and game theory. The tool forecasts competitor moves and responds strategically, like chess: it calculates the most reasonable next move. That is impressive, given that in many stores prices are changed by people who do not think that way. They do not consider the next step. They think: a competitor dropped the price, so I have to drop mine too.
Automation here starts from the principle that profit comes first. It raises the price when competition allows, protecting your margin.
2. Supply forecasting and out of stock
The second process is supply forecasting. What is the problem here? The marketplace has a local sales metric, and there is the concept of out of stock.
People often ask what that means. You sell on the marketplace, your goods are in warehouses, but at some point they start running out. You fail to bring a new batch in time, the product runs out, and you get an out of stock: there is demand, but physically no stock in the warehouses.
What does that lead to? You get lower search rankings. And if the goods are also going to non-local clusters, you get penalties for non-locality that can be 3, 5, 8 or even 10 percent.
At the same time you may have plenty of goods already sitting in clusters. You overload the warehouses, and that pushes you into paid storage.
Three levels of supply management
Some work manually: open the seller dashboard, look at stock and place supplies by hand. That is the basic level.
The second level is Excel: export data and metrics and calculate which supplies you need. That is how advanced sellers work.
The third level is still rare on the Russian and CIS markets: predictive analytics. Supplies are analyzed and planned ahead.
Predictive analytics in practice
A system looks at your sales velocity, seasonality and historical data over past months, then plans on its own: this product needs this many units, to this warehouse, on this date. It calculates everything automatically from historical data.
One example is SoStocked. Its pitch: stop losing profit to stockouts, overstock and fees. Amazon sellers lose hundreds of thousands of dollars a year to inventory mistakes and hidden costs.
That is familiar to us too. Sellers often lose money by overloading warehouses, under-supplying or something similar.
SoStocked optimizes inventory to avoid both shortages and overstock and points to upside in profit. Its inventory optimization is built specifically for Amazon sellers. It forecasts and plans supplies 12 months ahead, factoring in seasonality, order management and stock levels. These are strategic actions that are already automated on the US market.
3. Rule-based promotion management
The third process that is automated in the US is rule-based promotion management. One example is Pacvue. The core idea is the if-then principle: if a condition is met, an action follows.
Examples of conditions and actions
- If stock for a product drops below 30 units, the system automatically pulls it out of a promotion or out of ads. That protects you from out of stock.
- If a competitor is dumping, the manager gets an alert. The system pings that action is needed.
- If margin falls below target, the algorithm sees it and acts. For example, it pauses participation in promotions or ad campaigns.
Why automation wins
A machine has no spare time to recalculate and update numbers. An automation system calculates everything in real time and, if something goes wrong, immediately recommends what to do.
We often act on intuition and by hand. Foreign markets already have tools that automate this work.
Pacvue, for instance, runs ad campaigns with real retail conditions, improves conversion and prioritizes spend. It aligns retail ad campaigns to cut wasted spend across the catalog and optimizes performance based on real-time conditions.
4. Fighting paid storage
The fourth process automated abroad is the fight against paid storage. You know what paid storage is: goods move from free to paid storage, you get costs, and your margin shrinks.
Paid storage often eats tens of thousands of dollars a month, simply because you cannot sell the goods off, they sit in warehouses and consume your net profit. Stores often go to zero or into the red.
Abroad there are tools that automate, speed up and simplify all of this. One example is Sellerboard.
It tracks all these metrics. If it sees that only a few weeks are left before paid storage kicks in, it recommends actions: lower the price, launch a promotion, ship out a batch.
In short, it warns you that very little time is left before paid storage and that you need to act.
Sellerboard costs $15 a month and offers a two-month free trial. It helps maintain stock, avoid overstock, minimize fees, accurately estimate promotion costs and plan smarter promotion investment. It also simplifies purchase order management with suppliers and FBA shipments (Fulfillment by Amazon). In essence it is an automation tool that covers many processes on the marketplace.
5. Product research and niche selection
The fifth process is finding new products, which is also automated abroad. I believe the right product choice at the start is one of the most fundamental and important steps.
Yet products are often chosen on a whim: "my relative has a factory, I will sell his product" or "I will find a source for some product and sell it." That is the most expensive mistake I have seen in marketplace business.
You can set up a product card, ads are not hard, promotions are not hard, generating sales is not hard. But if you picked the wrong product, all those perfect settings and perfectly polished details will not lead to good sales. The market may simply be small, and all your perfect cards will bring in a small turnover.
Small market versus big demand
Sometimes cards are far from perfect: metrics look weak, yet they sell in markets where demand volume is half a million a month. Clearly the seller will get a fraction of that turnover, and net profit will be modest, even though a month of work went into the card.
Meanwhile there are categories where you can earn more. Abroad, many sellers look for products based on a list of a large number of niches.
Automating product research
Product research becomes a science of its own. There is a dedicated service, Helium 10, a product search engine for Amazon and Walmart.
It minimizes research time. The feature refreshes a list of several million listings, leaving only the products that match your price, profitability and operational constraints.
In short, you enter your criteria: what products you want to find, what would interest you. The system then finds a niche with low competition and optimal metric values.
The reverse-engineering method: look at the full catalog, take the metrics and use the result to shape your strategy. That is how products are analyzed on foreign markets.
On our market, all of this product selection is very often done by hand. From audits and consultations, I see sellers doing it all manually.
How to automate your store
If you do not want to wait for Western services (they may never arrive, and local equivalents may never appear), you can automate your store and your business processes right now.
If a process takes your team tens of hours a week or hundreds of hours a month, understand this: if the process is repeatable, it can be automated and will most likely pay for itself within a few months.
Summary: five growth points for margin
Let us sum up what can be done more efficiently and give five growth points for profitability.
Pricing: game theory
Today pricing is often blind dumping, following what competitors do and adjusting to it. Abroad, pricing is built on game theory: not "I will lower prices somehow, raise them somehow and earn somehow," but prices built on mathematical factors and formulas.
Supplies: predictive analytics
Supplies are often built in Excel or, at best, on intuition. Abroad, predictive analytics and predictive math are used: people plan based on formulas. Here it is often intuition; there it is math and calculations.
Promotions: automatic rules
Promotions are often controlled manually. Abroad, automatic rules are set: if a condition is met, the promotion is switched off, increased or decreased.
Stock: early warnings
We often pay penalties, overpay for storage and fail to track these moments. Abroad, the standard is an early reminder, an early warning that goods may fall into paid storage. This is a common pain for many sellers.
Purchasing: big data
We often buy goods on intuition, assuming we will sell them within a certain period, and often guess wrong. Abroad, the standard is to think based on big data, math, formulas and calculations with concrete metrics: how much of a given product we will most likely sell.
Recommendation
Think about which processes and which people in your business may be redundant. The idea for this breakdown came when I saw a company turning over several million a month with minimal margin, eaten up by staff salaries.
Over the coming years everything points to lower profitability and greater complexity. The winner will not be the one with the bigger team, but the one whose business is better digitized and automated.
Even a single implemented automation algorithm can pay for itself within a few months. Automation is not free, but over several months it pays back, and over a year it turns into profit, especially compared with hiring staff to do busywork.
We develop such algorithms and automations for marketplace work and for product businesses in general, based on what is available on the market, on an understanding of the Ozon, Wildberries and Yandex Market APIs and on the processes inside a product business.
You can build your own infrastructure yourself, or contact us and we will help.
What to do next
The winner will not be the one with the bigger team, but the one whose business is better digitized and automated. Even one implemented automation algorithm can pay for itself within a few months. Think about which processes in your business can be automated and start with the most resource-hungry one.
