You do not have a margin problem. You have a cost to serve you never act on

When margin slips, the conversation usually goes one of two ways. Either we need to put price up, or we need to take cost out. Both are real levers, both get pulled, and both feel like doing something. But they skip the question that actually explains the margin, which is who you are serving, through which channel, and what it truly costs you to do it.

Most businesses know their margin by product. Far fewer know it by customer or by channel. And the ones that do almost always look at it once a year in a finance review, nod, and then carry on managing the average. That is the real gap. It is not that nobody has the number. It is that the number never reaches the people making commercial decisions, so it never changes one.

Average hides the answer

A blended margin is an average, and an average is a story about nobody. A healthy looking thirty percent can be a small account at eight and a large one at forty five. It can be one channel quietly subsidising another that bleeds on every order. You manage the blended number, it looks stable, and you feel in control, while the real money leaks out one customer and one channel at a time, none of them big enough on their own to set off an alarm.

The price up, cost down reflex works on the average too. So you raise price across the board and trim cost across the board, and you move the average a point, and you still have not touched the customer who was never going to be profitable at any sensible price.

Cost to serve is a commercial number, not a finance one

This is where it gets misfiled. Cost to serve gets treated as a finance or supply chain exercise, something to be calculated and filed. It is not. It is a commercial weapon, and it lives in commercial decisions.

Drop size. Delivery frequency. Returns. Payment terms. The long tail of tiny accounts that each take a full sales call and an order and a delivery to move almost nothing. Every one of those is a commercial choice, not a ledger entry. Once you can see cost to serve by customer and by channel, it tells you exactly what to do. Which account to reprice. Which one to put on new terms or a minimum order. Which line to delist. And which handful of customers you should, politely and deliberately, let go.

What it changes when you look

I once ran a business that put distribution, retail outlets and a central kitchen under one roof. On the product margin line they did not look wildly different. Underneath, the cost to serve was nothing alike. A convenience channel taking small daily drops costs nothing like a distributor lifting a pallet a week, and a central kitchen feeding both is a different animal again. Treating them as one business, managed to one blended margin, would have hidden which one was actually paying for the others. Once you separate them you stop pricing and terming them the same way, because they were never the same business.

The point

You probably do not have a margin problem. You have a cost to serve you have measured once and never acted on, and a blended average that lets you avoid looking too closely. Put the real number on it, customer by customer and channel by channel, and put it in front of the people who make commercial calls. The answer is almost never where the price up, cost down reflex was pointing.