Are you actually selling what your customers want, or are you just measuring how long it takes for a freezing person to give up on their standards? It is a question that most retail planners avoid with the same frantic energy I use to avoid a claimant who I know is hiding a pristine mountain bike in a garage they claim was destroyed by a localized hurricane.

We prefer the neatness of the spreadsheet because the spreadsheet does not complain about the cold. It only records the transaction. But in the world of climate control, particularly when the temperature in Chisinau or Balti begins its inevitable October nosedive, the transaction is often a record of defeat, not a record of preference.

I recently found myself in the digital equivalent of a vulnerable position-joining a high-stakes video call with my camera on by mistake while I was hunched over my desk, looking less like a professional investigator and more like a man trying to decipher a map of the afterlife.

That sudden, prickly heat of exposure is exactly what inventory managers should feel when they look at their end-of-season “best-seller” lists. We see a column of numbers and we call it “demand.” In reality, it is often just a ledger of what was left in the warehouse when the panic set in.

The February Illusion

Imagine a planning meeting in the dead of February. The air outside is grey, but inside the room, the mood is clinical and confident. A slide flickers to life, showing the performance of oil-filled radiators from the previous autumn.

Two models, perhaps a basic 1.5 kW unit with seven sections, show a massive spike in volume during the third week of October. The proposal is immediate and “data-driven”: double the order for next year. Let’s own the “value” segment.

Nobody in that room, not the analyst, not the buyer, not the regional manager, asks the only question that matters: How many of those four hundred buyers actually wanted the 2.5 kW model with eleven sections and a built-in timer?

The data does not show the ghost of the customer who walked into a store in Cahul, felt the draft from the front door, and desperately needed a heater that could handle a thirty-square-meter room, only to find the “Premium” shelf empty.

That customer does not go home and wait for a restock. They cannot. Their toddler is shivering. So they buy the seven-section unit, the one meant for a small home office or a nursery, and they drag it into a living room it was never designed to heat. They “settle” under duress. And the retailer, seeing the sale, concludes that the market is shifting toward smaller, cheaper units. This is how the “Drift” begins.

In my line of work, we call this a “selection bias error,” though in insurance, it usually involves someone claiming they only ever drove their car to church on Sundays, conveniently forgetting the drag racing tracks. In retail, the bias is more subtle.

By feeding “panic sales” back into the purchasing algorithm, you aren’t following the market-you are forcing it into a narrower and narrower hallway. Over three or four seasons, the store’s assortment drifts away from the engineered solutions that people actually need, becoming instead a collection of the things people take when they have no other choice.

PHYSICS The Principle of Thermal Inertia

This is particularly dangerous in the category of oil-filled radiators. These are not impulse purchases like a chocolate bar at the checkout; they are technical solutions to a thermodynamic problem.

An oil-filled radiator works on the principle of thermal inertia. The heating element warms the oil, which then warms the metal casing, which then radiates heat into the room long after the electricity has stopped flowing. It is a slow, steady, and comfortable heat. But that comfort depends entirely on the match between the unit’s power and the room’s volume.

When a customer visits BOMBA.md or walks into one of the physical stores, they are usually looking for a specific match: a Kraft or Adler unit that can handle 25 square meters. They are looking for the 2.2 kW or 2.9 kW models with nine or eleven sections.

If those models are out of stock because the retailer underestimated the “quality” demand, the customer buys the 1.5 kW Heinner or Mesko unit instead.

The result? The customer’s electricity bill goes up because the undersized unit has to run at 100% capacity for twenty-four hours a day just to keep the room at a tolerable eighteen degrees. The radiator’s lifespan is shortened. The customer is mildly annoyed with the brand.

Yet, on the retailer’s spreadsheet, that Mesko unit is a “hero product” because it sold out in forty-eight hours. The company sincerely believes it is a genius for stocking it.

We ignore the psychology of the “cold snap” at our peril. When the frost hits the windows in Soroca, the buyer is no longer a rational actor comparing specs; they are a biological organism seeking survival. Sales data collected during a period of scarcity is not a map of preference; it is a map of availability.

If you want to stop the drift, you have to record what was wanted, not just what was taken. This is an unfashionable and difficult task. It requires looking at search queries that returned “zero results.”

It requires listening to the floor staff when they say, “We could have sold fifty more of the 11-section units if we had them.” It requires acknowledging that the top-selling item might actually be your biggest failure, because it represents the volume of customers you failed to satisfy with the correct product.

The Logic of the Ladder

At BOMBA.md, the approach is different because the catalogue is built on the axes of real-world reasoning: kilowatts, room size, and section count.

But even with a robust catalogue, the risk of the “October Rush” remains. When demand spikes across all twenty-plus stores, the supply chain is tested. The temptation is always there: if the 900-lei model is flying off the shelves, why bother with the 1,400-lei model next year?

The answer is that the 1,400-lei model-the one with more heating sections and a higher thermal mass-is the one the customer actually wanted for their primary living space. If you stop stocking it, you aren’t “optimizing for the market,” you are abandoning the customers who have the most significant heating problems to solve.

This phenomenon isn’t limited to heating appliances. It’s how playlists on streaming services become repetitive-the algorithm sees you listened to a “hit” song three times and decides that’s all you like, ignoring the weird, obscure jazz track you searched for specifically but couldn’t find.

It’s how hiring converges on a single profile-you hire the person who was available and “good enough,” and then you rewrite the job description to match their specific set of flaws for the next round.

The defense against this flattening of the world is a return to first principles. In the context of home heating in Moldova, those principles are dictated by the laws of physics, not the laws of the ledger.

To truly understand your market, you have to look for the “missing” data. You have to look at the gap between the 1.5 kW unit that sold in the thousands and the 2.9 kW unit that sold out in a day. That gap-that “stock-out” period-is where the real preference lives. It is the silent signal that tells you your customers are more sophisticated and have higher needs than your sales report suggests.

As I sit here in my office, feeling the slight hum of a radiator that I chose specifically for its ten sections and its ability to keep this drafty corner at a steady twenty-one degrees, I think about that February meeting again.

I think about the person at the head of the table pointing to the “top two” bestsellers. I want to tell them about the “desperation tax” that customers pay when the right product isn’t there. I want to tell them that their “successful” season was actually a series of missed opportunities to build long-term trust.

But I suppose I’ll just stick to investigating fraud. It’s easier to find a man lying about a stolen car than it is to find a retailer willing to admit their best-selling product is a symptom of their own misunderstanding.

The next time the temperature drops and the rush begins, remember that every “out of stock” notification is a piece of data more valuable than a thousand “sold” receipts. It is the record of what your customer actually hoped to find when they turned to you for warmth.

If you ignore that ghost signal, you’ll find yourself with a warehouse full of things people only buy when they’re desperate-and eventually, they’ll stop being desperate for you. They’ll go find someone who actually listened to what they wanted before the frost set in.

Recording the “No”

Recording the “no” is just as important as recording the “yes.” In a world of automated ordering and algorithmic drift, the human element-the ability to recognize when a sale is a compromise rather than a choice-is the only thing that keeps a catalogue from becoming a graveyard of the “almost right.”

Whether it’s a Kraft unit with seven sections or a Camry with twelve, the goal is the same: to match the heat to the cold, not the stock to the panic.