Inventory optimization does not end with the right quantity

The biggest gain is reacting fast to a bad forecast

The first question: which inventory problems do retailers face most often, and where is the biggest room for improvement?

Food retail in particular struggles to balance product availability against markdowns, or even waste. The biggest gain is in treating the whole problem as one: it is not only about forecasting the right quantity, but about finding the cases where the forecast misses fast enough to discount the product or give it a second placement. Otherwise it ends in a clearance below cost, or a write‑off.

The difference between a markdown and a write‑off is mostly a matter of time. Spot the miss in three days and you are still working with margin. Spot it in three weeks and you are working with a loss.

AI changes the buyer's role, not just the order

The second question was about technology — what moves the needle for retailers today, and how the role of automation and AI is shifting.

AI now plays a big part in automating simple, repetitive tasks. Small activities used to be left alone, because automating each one was expensive. Now it is cheap and fast, and there are a lot of them in any company. That changes the role of the people who manage inventory: instead of deciding what to order and when, buyers and product managers move into an analyst role, setting the limits and targets for the AI and ML models. They make the whole process more efficient and they move the company's financial numbers directly.

The shift is in where the decision happens. The model calculates the order. The person decides what availability and what risk the company is willing to carry, and that decision has a price tag.

Get the brief right before you buy the tool

The third question: where should a retailer start so the investment actually pays off?

The first step is to be clear about what inventory optimization actually means for management, what they want to achieve and why — whether this is a one‑off exercise or day‑to‑day operations. If that is unclear or people disagree, it pays to bring in an outside expert to help frame the brief. Only then does it make sense to deploy a tool that runs the optimization on data and smart algorithms.

Buy or build

The key decision is a ready‑made solution versus in‑house development. Both have their pros and cons and you need to be honest with yourself. We know plenty of companies that went into their own development assuming it would be easy, and got burned. Key people stayed tied up in an internal project and the core business paid for it.

The cost of building in‑house is rarely counted in full. Developer salaries are the visible part. What those people did not do in the meantime is the rest.

Where this appeared

Jak najít rovnováhu mezi prázdným regálem a vázanými penězi (How to balance an empty shelf against tied‑up cash), Pavel Gregor, Retail News 9/2026, pp. 30–31 (print edition, in Czech)