How Manufacturers Enter E-Commerce: When a Warehouse Built for B2B Meets the Individual Customer

This episode of Rozmowy Logistyków is a recording of a panel from the Council of Supply Chain Management Professionals conference, where alongside Adam Sobolewski and Piotr Skobało, Tomek Zwiercan, Managing Director of GLS Poland, also took part. They discuss the logistics challenges facing a manufacturing company that starts selling direct-to-consumer through e-commerce.

DISTRIBUTION CENTERS BUILT FOR THE WRONG MODEL

Consumer-brand (CPG) manufacturers that suddenly hit double-digit percentages of revenue from e-commerce discover that their distribution centers across Europe were designed for store deliveries — large shipments, consolidation, a different work rhythm. An e-commerce order averages 1.7-2.7 units, requiring picking and packing at a completely different level of labor intensity than a pallet to a store. Without investment in automation and process optimization, these two flows — B2B and e-commerce — simply don't work well together in the same distribution center.

THE INDIVIDUAL CUSTOMER REACTS INSTANTLY, THE RETAIL CHAIN DOESN'T

A key psychological difference: a delayed delivery to a store is an internal logistics problem; a delayed delivery to an individual customer ends in a Google complaint within hours — even if technically it's the same B2B shipment going to a pickup point. Large players handle this by physically separating the flows: an order delivered to a store, even if it originates from a "new" e-commerce center, travels on separate transport from a pallet from the main distribution center, even though it ends up at the same location.

DATA: GREATER SCALE, GREATER PREDICTIVE VALUE

Machine learning and AI models need a large number of data points — B2B sales (large, infrequent transactions) generate fewer of them than e-commerce (small, frequent orders). The paradox: CPG companies entering e-commerce gain access to much richer datasets, but most of them — even companies with tens of millions of złoty in revenue — still don't build predictive models on that basis. Courier operators handling tens to hundreds of millions of parcels a year have more data points than even the largest single online store, which opens the door to data-sharing partnerships that benefit both sides.

ROUTE OPTIMIZATION: MACHINE LEARNING INSTEAD OF COURIER KNOWLEDGE

Courier companies today use algorithms that learn optimal routes based on how a given courier normally drives — critical in peak season, when up to 25% more drivers than usual are hired. Instead of requiring a new driver to learn a route over a week or two, a scanner or navigation system guides them from day one. This opens the door to predicting delivery time already at the point of order in an online store — technically, it's just a matter of linking optimized-route data with the shopping cart.

SUSTAINABILITY: DECLARATIONS VERSUS WILLINGNESS TO PAY

Research in Western Europe shows a gap: 40-60% of respondents say they want zero-emission delivery, but only a few percent agree to pay more for it. A real example from practice: even an eco-friendly mailer bag 10 groszy more expensive than the standard one doesn't sell as a customer option. This means companies that don't build the cost of green delivery directly into their own P&L can't deliver on ambitious emissions-reduction pledges.

D2C AS A STRATEGY THAT INCREASES THE CARBON FOOTPRINT

More and more manufacturers (example: L'Oréal, targeting over 50% of global sales direct-to-consumer by 2030, mainly in Asia) are moving into D2C sales, bypassing distributors — but that by definition means more small shipments instead of full cartons to a store, which raises the carbon footprint per unit of product, even with a low return rate in categories like cosmetics (around 1%, versus 20-30% in fashion). This trend isn't just the domain of global corporations — a Polish cosmetics and supplements manufacturer with tens of millions of złoty in revenue has already considered opening its own shipping warehouse next to its contract factory to move into direct sales.

CERTAINTY MATTERS MORE THAN SPEED

Consumer research shows that outside of seasonal peaks (toys before the holidays or Children's Day), customers value not the fastest delivery, but the most certain one — knowing the parcel will arrive exactly when they're home or near a locker. Large players (Amazon, Allegro) already offer predictive delivery time windows; companies just entering e-commerce from a traditional B2B model still don't do this, though the flexibility to change the delivery method up to the last moment (e.g. from a locker to door delivery) is another step still missing from most current offers.