Inventory Management: Where Should the Responsibility Sit?

In this episode of Rozmowy Logistyków, Piotr Skobało and Adam Sobolewski continue their series on logistics costs, this time discussing inventory planning. They talk about who in the organization should own inventory, and what data is worth collecting to plan it more optimally.

WHERE SHOULD RESPONSIBILITY FOR INVENTORY SIT

There is no single correct model. In mature global organizations, demand planning sits within supply chain, which acts as an interface between production/purchasing and sales — production focuses on manufacturing cost, logistics on having the goods where and when they're needed. In other models (e.g. fashion, which works in seasons), the team designing the collection decides how much will sell and where, while logistics is purely responsible for delivery. The key question for choosing a model is: where in the chain is value added? When a product has to be "pried away" from a supplier (e.g. the latest electronics in limited supply), inventory management naturally sits with commercial, who negotiate the contract. When a product is easily available but distributed across an extensive network of many warehouses and thousands of points of sale (classic FMCG), detailed product-location planning is a job for logistics.

S&OP: START WITH THE CONVERSATION, NOT THE METHODOLOGY

Sales and Operations Planning is a recurring process (monthly in manufacturing, sometimes weekly in retail) bringing sales, marketing, purchasing/production, and logistics together around one question: what are we doing in this business, and how do we avoid surprising each other. The condition for success is that leadership believes these meetings matter — if the CEO treats S&OP as unimportant, the process won't work, regardless of methodology. A real example from practice: a product with zero availability ended up on the front page of a promotional flyer because the sales team had blocked it from being purchased (as "being phased out"), and no one asked logistics before putting it in the promotion. S&OP eliminates this type of surprise, turning mutual blame ("why didn't you deliver") into shared planning ahead of time.

THE FOUR STAGES OF THE S&OP CYCLE

Collecting and cleaning data, sales/demand planning, supply planning against those plans (accounting for minimum order quantities and transport constraints), and finally balancing — what to do when the sales plan can't be directly translated into a supply plan, because the supplier doesn't have enough quantity or the warehouse doesn't have the capacity.

LEAD TIME IS MORE THAN A CONTRACTUAL NUMBER OF DAYS

Knowing the standard lead time per product or supplier is a starting point — but the real value lies in knowing how often a given supplier actually deviates (in time or quantity) from what was contracted. That's an extra layer of information that digital-twin-type models rely on today — for example, factoring in a 70% probability of delays during the rainy season in Malaysia, automatically converting a standard 3-week lead time into 5.

THE PARETO PRINCIPLE: START WITH 10 SKUS, NOT ALL OF THEM

With a portfolio running into tens of thousands of SKUs, trying to tackle everything at once is paralyzing. Practical advice: start with the 10 SKUs generating the most sales, break them down in detail (even in Excel), then add another 20 — this builds a foundation that makes the rest of the portfolio easier to manage. A real example from the industry: a company with 800,000 SKUs in its assortment keeps inventory on only 150,000 of them — the rest is fulfilled order-to-order, instead of holding stock on everything, because financially it simply doesn't add up.

SERVICE COSTS OFTEN LEFT OUT OF ORDER CALCULATIONS

A common mistake: calculating the optimal order quantity without accounting for the cost of receiving and storing goods — weekly small deliveries generate a different warehouse workload than one pallet or container per quarter, which affects whether holding a given inventory level even makes sense. The final conclusion: to plan inventory well, you need to step out of functional silos — a conversation about assumptions ahead of time is cheaper than explaining mistakes after the fact.