The drug market grew 13 percent without selling a single extra pack

Two baskets, both from the transaction data of a chain I worked with. One day, one pharmacy, twenty minutes between the two purchases.

The first: an expensive weight-loss drug, a facial serum, vitamins. The total was about ten times the store's average basket. The shopper never once looked at a price tag.

The second: three generics, in the largest packs on the shelf. Everything at the club price, access to which the person had paid for in advance with a subscription. Before leaving home he had checked prices in two apps.

Both stood at the same display and saw the same tags. Statistics will add their baskets together and report the average. There was no shopper with the average basket in that pharmacy.

I advise pharmacy chains, take part in developing pricing services for retail, and am affiliated with parso.ru and flexpricer.ru.

What the till shows

According to DSM Group, the commercial drug market grew by 13.3 percent in 2025 and reached 1,850 billion rubles (₽). Over the same year, 4.4 billion packs went through pharmacies. The gain was 0.02 percent.

Zero.

In December a pack cost 443 rubles on average. The share of items priced above one thousand rubles reached 37.7 percent of the market in money terms. There are about 83,000 pharmacies in the country, and they now open three times more slowly: two to four percent a year, against roughly ten before.

Before I build conclusions on this, let me concede part of my own argument. Nobody indexes the one-thousand-ruble threshold, and prices rise thirteen percent a year, so some items crossed the line on their own, with no help from the shopper. And the habit of buying a bigger pack to lock in the price also pulls the average pack cost up, even though that is the behavior of someone economizing. The same indicator is driven up by both the upper branch and the lower one.

What is left after the deductions. The divergence is not visible in the annual aggregates. It is visible at the till: the spread of baskets within a single day at a single store has become wider than the spread between formats. A pharmacy by the metro serves two economies at once and cannot decide which of them to set its price tag for.

The letter K, briefly

The term was coined by financial analyst Peter Atwater in the summer of 2020 in the Financial Times: the recovery after the lockdowns looked neither like a V nor a U, because one part of the economy went up while another went down, and it happened simultaneously. A K-shaped economy has meant exactly that ever since: diverging trajectories instead of a shared cycle.

The mechanics were dissected best by the Federal Reserve Bank of New York in May 2026. The main channel of divergence is not wages but wealth: the real net worth of the top one percent of Americans has grown by more than a quarter over the past few years, while the middle tier's gain did not reach even ten percent. The second channel is the poor's own inflation. Their basket holds more food, fuel and rent, that is, the things rising fastest, so one and the same official figure means different things to two families. And wages, oddly enough, worked against stratification: in some periods the lower quartiles grew the fastest of all.

How well proven is this. Moody's Analytics reckons that the top twenty percent of Americans account for almost sixty percent of consumer spending, and the share of the top ten percent reached 49.2 percent, the highest since 1989. The government prototype from the Bureau of Economic Analysis, for the same figure, gives 25.7 percent. The Bureau of Labor Statistics comes out at twenty-three. The Federal Reserve Bank of Minneapolis brought all the gauges together and concluded that the data do not add up to a clear K-shaped picture.

I say this not for the sake of academic precision. In a month someone will bring you a presentation with a reference to Moody's. Nobody will bring you the reference to Minneapolis.

The Russian version

We simply have no distributional statistics of the right quality, so arguing about percentages is pointless. But the behavioral readings converge. According to Romir, in 2026 demand is falling for expensive alcohol, smartphones, jewelry and household appliances, and rising for entertainment, delicacies, beauty-salon services, car repairs and travel. This is not the behavior of someone who has grown poor. This is how someone behaves who is shifting the budget from one category to another.

The most vivid confirmation was found inside the lower branch itself. Over 2025 Chizhik grew by 67 percent, to 417.5 billion rubles, while Svetofor lost three percent, fell to 394.4 billion and closed 315 stores. Both chains sell cheap. It does not follow that one manages price better: Chizhik has the logistics and buying power of X5 behind it. What follows is more modest. A low price alone is no longer enough to win.

The price is now fully visible

There is one more circumstance without which any conversation about pharmacy price tags is stuck in the previous decade. The shopper from the second basket had checked prices in apps before he left home, and this has long been the norm, not advanced behavior. Over 2025, online pharmacies handled 297 million orders worth 475 billion rubles, up twenty percent in units and twenty-three in money.

For a pharmacy this means something unpleasant. Before, you could count on the price being remembered for a few familiar items, with the rest taken without a second look. Now any item can be checked in two seconds, and it is checked if the size of the purchase is worth it.

What the leader did

By the AlphaRM ranking, first place for 2025 goes to the Aprel (April) chain: 259 billion rubles, growth of 34 percent, 9,976 stores in 77 regions, and grown without buying up other chains. DSM Group has its own methodology and its own leader, Rigla, so a chain's place in the table depends on whose ranking you open. The place here is not the point at all. The model is what is interesting.

Aprel does not sell a markup on a pack. It sells a subscription. The customer pays for access to club prices for a term of a month to a year and gets a shelf that is fifteen to thirty percent cheaper.

The design is elegant, and I do not like it. The chain charges the most economizing shopper money to stop comparing prices. It works not because the shelf is cheaper. Plenty of chains have a cheap shelf. It works because, once you have paid, you no longer feel like comparing, and what you paid has to be earned back.

As for single pharmacies, the year 2026 finishes them off separately. The base VAT (value-added tax) rate rose to 22 percent, and the VAT-exemption threshold under the simplified tax regime (USN) dropped from sixty million rubles of revenue to twenty million. A single-store pharmacy with a turnover of twenty-five million is already caught by VAT. Now it loses to the chain twice over: on the input price from the distributor and on the cost of its own bookkeeping.

An uncomfortable admission

Before moving on to what to do with all this, let me argue against myself.

From the fact that large chains grow faster than the market, it does not follow that they grow thanks to price management. If anything, the reverse. To gauge how volume responds to price on even a single item, you need tens of thousands of baskets a week: a chain of ten pharmacies does not have them, a chain of a thousand does. You can hold an anchor item at zero margin only if you have the best input price, and that is given for volume. In other words, large chains did not grow because they manage price. They gained the ability to manage price because they grew.

For a single pharmacy this means that part of the leaders' advantage is physically out of its reach, and no analytics will change that. What follows is about what is nonetheless within reach.

What to do about it and where the limits are

What this does NOT mean

This is not about changing prices more often. And it is emphatically not about changing them over the course of a day depending on who is standing at the shelf. A price computed personally for you from your own data is called surveillance pricing, and in a number of countries it has already earned its own regulation. Mixing these things up is dangerous, and there will be a story below about how much such a mix-up cost one chain.

In retail, dynamic pricing means this. Prices are recalculated regularly according to rules. The rules are built on how demand behaves, what competitors do, and what the goods cost to buy. Every change is checked against a control group without fail. The word "dynamic" refers to the process, not to the frequency.

Why a discount does not replace a system

The arithmetic is simple and unpleasant. At a gross margin of twenty percent, a five percent discount takes a quarter of the gross profit, and to get back to the old money you have to sell a third more. A discount works like a bet in a casino: sometimes it comes in.

In a pharmacy the bet is also blind. A discount is usually given on what is selling anyway, which means it is paid for by the shopper who would have bought without it.

How many key value items a chain needs

A key value item, or KVI, is an item by whose price the shopper judges the price level of the whole pharmacy, even though he does not know the prices of the rest of the range.

Here a correction to the common version is needed. "Five to seven prices" describes the memory of a single shopper. For a chain this is not the length of a list. A diabetic needs seven items of their own, a mother with an infant needs hers, a person with hypertension needs theirs, and the sets do not overlap. Across all segments the anchors number several hundred.

They are counted from the baskets: purchase frequency, position in the basket, price elasticity of demand (how far volume responds to a change in price), the ability to pull the rest of the basket along. A noticeable share of the items chains keep in their anchor lists by tradition do not pass on these criteria. Nobody judges the pharmacy by them, and no margin is taken on them.

Where you can manage the markup and where you cannot

Behind the anchor comes a predictable tail into the basket, the accompanying items, which is where the money is made. And here the field of application has to be narrowed, because in pharma it is drawn wider than is permissible.

You can manage the markup where the price is free: parapharmacy (the pharmacy's non-drug range), medical devices, cosmetics, hygiene. A diabetic buys test strips every month, and almost always with lancets and wipes; he knows the price of the strips by heart, and nobody knows the price of the wipes. Diapers pull along wet wipes and cream. Dermocosmetics are trickier, but even there a cleansing gel usually brings a moisturizer, and by summer an SPF product as well.

And here is what you must not do, and this is not my personal preference:

• Prescription drugs. There is no prescription in the basket. A model built on such statistics describes over-the-counter dispensing, not therapy.
• Items on the VED list (vital and essential drugs; a price-regulated category). Their price is capped from above by the manufacturer's registered price and by the maximum markups that each region sets for itself.
• Antibiotics and course anti-inflammatories. Excluded entirely.
• Pediatric and adult forms of the same active substance. In one basket they create a risk of overdose. Never mind the basket then.

There is noticeably less of the range left than the vendors promise. But what is left is what you will not have to make excuses for.

Four constraints absent from grocery retail

The ceiling. On the VED list the algorithm has no right to earn a single kopeck above the limit. This is not sewn into the optimization function as a penalty; it is a wall, and it is put up before the calculation.

The batches. The same drug lies in the warehouse at different purchase prices, because it arrived in different deliveries and at different price lists. The selling price must be computed from the specific batch, otherwise half the calculated effect turns out to be a bookkeeping illusion: you will think you raised the margin, when you have only raised the average temperature across the warehouse.

Shelf life. A tail that has grown more expensive sells more slowly. In groceries that is frozen working capital, in a pharmacy it is also a write-off, because a drug has a date after which it is worth nothing.

The counter. The Good Pharmacy Practice order obliges the employee to tell the shopper about a cheaper analog with the same active substance. A counter pharmacist who has an average-basket target hanging over her will still name the cheap analog if she sees the person counting small change in his wallet. And she will be right to. A well-computed scheme dies within a week if nobody has discussed it with the people standing at the till.

How to tell effect from season

Checking matters more than counting.

The hypothesis is rolled out to ten to twenty percent of the stores, the control group lives on the old prices, and the effect is measured over two to four weeks by volume, margin and average basket. Without a control group you cannot tell the work of price from a cold-and-flu season, and in six months the argument about whether it worked or not will be unresolvable in principle.

It all looked smooth on paper. The "plus five to eight points of margin" range has wandered through vendor presentations for a decade, and I have never once seen it confirmed by a control group. Not once.

There is exactly one public Russian case where there was a control group. A regional chain of 150-plus pharmacies, fifteen pilot stores against fifteen control ones, two months. Gross income rose by 15.62 percent, revenue by 11.7 percent, the number of sales by 9.23 percent. The case was published on Retail.ru in December 2021 by the company myRetailStrategy as advertising for its own system.

The last circumstance matters more than the figures themselves. Divide one by the other: at an initial margin of about twenty percent, gross-income growth of 15.6 against revenue of 11.7 means a gain in profitability of roughly 0.7 percentage points. Not five and not eight. And this is the best of the public results, selected by a vendor for advertising.

I would frame a realistic benchmark like this. A chain that has never had category management will win one and a half to two points of front margin (the markup on the shelf) in the first year. A chain with working category management will win fractions of a point. And you have to count not the front margin but the whole profit, because next to it stands the back margin, money from the manufacturer under marketing/listing contracts, and in pharmacy retail it often weighs more than the markup. Front-margin growth that has eaten the back margin is no result at all.

Others' experience and others' rakes

I keep the foreign examples separate from the Russian material on purpose: less carries over from there than it seems, and the most useful things are not the cases but the mistakes.

Who rolled it out

Farmacorp, Bolivia, 70 pharmacies. A family chain with a three-generation history rebuilt its pricing and added ten percent of revenue in the first full year. The value of the case lies in one caveat from their chief information officer: over the same year they opened one new pharmacy. One. In an industry where any growth is customarily explained by expansion, that is a rare kind of clarity.

A caveat from my side: there was no control group there, the case was published by the supplier of the system, and precisely for that reason I cite it as an illustration, not as proof.

Boots, United Kingdom. The Price Advantage program gives reduced prices on about a hundred and fifty items a month, but only to loyalty-card holders. About 58 percent of the chain's shoppers have the card, and 70 percent of the regulars. The idea is the same as Aprel's: the price becomes a property of the relationship with the shopper. Boots gives the club price for free, while Aprel charges money for it.

Chemist Warehouse, Australia. The country's largest discounter. It had more than 540 pharmacies; after merging with the wholesaler Sigma it became 880-plus. The key figure is a different one: about seventy percent of the company's revenue comes from the non-prescription part of the range, against twenty-seven percent at a typical Australian pharmacy. The company simply built its business where the price is free, and uses the prescription shelf as a traffic generator.

Where they got burned

Boots and Superdrug, January 2026. An investigation showed that of the 690 short-term promotions Boots ran from March to August 2025, 119 were presented in a way that misled the shopper. Every sixth one. Calls went out for the antitrust regulator to check both chains. The lesson for everyone who builds price around loyalty: a club price lives exactly until the day someone asks what price it is counted down from.

Wendy's, February 2024. The chain's CEO uttered the phrase "dynamic pricing" on a quarterly investor call, meaning digital menu boards and discounts in slow hours. The press translated it as surge pricing, a price by peak demand. Within a day a hashtag calling for a boycott went viral on social media, a competitor rolled out a "No urge to surge" campaign with a free burger. Two weeks later Wendy's was explaining that it had not been about to raise prices at peak hours. It did not abandon the plans; it paid the reputational bill in full.

The word did it in.

RealPage, November 2025. The U.S. Department of Justice proposed a settlement in the case over an algorithm that suggested to landlords what rate to set. No fines and no admission of guilt, but with conditions worth reading as a ready-made regulation: models may be trained only on non-public data at least a year old and unconnected to active contracts, and market information may be given to subscribers no more granularly than at the state level. The algorithm itself is not banned. What is banned is the exchange of fresh non-public data between competitors through a shared vendor.

The restriction applies to consultants like me too. Whoever runs the pricing schemes of two chains in one region is exactly such a shared vendor.

What carries over

Not much, and this is worth saying plainly. The Australian maneuver is already being reproduced here, only it is called growth in the share of parapharmacy and private labels: that is precisely the move into the zone of free prices. The British construction arrived earlier and in a harsher form, which means the British side effects will arrive too. The question of what price the club discount is counted down from has not yet been asked aloud here. I bet it will be asked before the end of 2027, and not by the shopper but by the regulator.

From the technological experience, the main and most boring thing carries over. The control group. It is what separates a case from a promise.

Three conclusions

First. The argument about whether the economy is K-shaped or something else is secondary for retail. What is primary is the fact visible at the till without any macroeconomics: one and the same shopper behaves like two different segments within a single basket.

Second. You cannot answer this with a discount, and still less with an across-the-board markup increase. You have to answer by splitting the range by roles: where we hold the price strictly at market because we are judged by it, and where we earn because there nobody checks the price. In a pharmacy the zone for such a maneuver is narrow, and that is a frame, not an annoying inconvenience.

Third. Anything that cannot be checked against a control group should be treated as opinion. Including mine.

And about those two baskets I began with. There were twenty minutes between them. Between the prices that would have suited each of these people there is a difference of about twentyfold. The average basket of that pharmacy describes neither of them. They still plan by it there.