Five situations.
Five individual perspectives at the table.

Your customers already buy through marketplaces. That channel grows by almost 30 per cent a year. The only question is whether your goods are there or those of a competitor who listed faster. Bundle the channels and you keep the customer relationship. Wait, and you become an interchangeable supplier to a marketplace.

Connecting every new channel with its own interface creates exactly the landscape that slows you down today. An integration layer turns that around. Channels dock on instead of being wired into the ERP one by one. The first channel costs the same. Every further one costs almost nothing.

Omni-channel fails on three truths about the same article. One sits in the ERP, one in the shop, one in the marketplace feed. So a decision comes before any connection. Which system owns stock, price and product data? That is data work, and it takes up half the project.

Double sales, stock errors and manual channel work appear in no cost centre. They hide in returns, express deliveries and overtime. Calculate the process cost per order and channel. That number usually makes the investment decision on its own.

Customers expect the same price and the same availability everywhere. Every deviation costs trust, and in B2B trust is the real currency. Channel consistency is part of brand care, not only an IT topic.

67 per cent of your buyers want to buy without sales contact. That is no rejection of your sales team. It frees them for new customers, advice and difficult cases. Companies that make the shift grow without adding headcount.

The portal is the smaller half. The bigger one is the ERP connection. Orders, stock, customer prices and payment terms have to flow automatically. Otherwise a nicer facade sits in front of the same manual process. Plan the integration as its own work package.

Every self-service order creates data that phone sales never delivered. Search terms, drop-offs, ordering rhythms. From that come assortment and pricing decisions. Whoever builds the portal plans the reporting from day one.

A manually entered order costs office time and creates input errors. Clarification cases tie up capital as well. Multiply your annual orders by an honest number of minutes. Against that sum the portal usually pays off faster than planned.

Self-service does not mean anonymous. The customer gets their catalogue, their prices and their order history, at any time. That is more attention than an overloaded office team can give. Which is exactly why customers stay.

Your pricing logic has grown over decades and is a real competitive advantage. That is why it belongs in your systems, not in heads and spreadsheets. What only the sales team knows retires with the sales team.

Pricing logic probably lives in three places today: in the ERP, in Excel and in habit. The shop is the chance to consolidate it. One rule set leads, every channel reads from it. That also reduces the special cases that slow down every system change.

Customer-specific prices require clean customer and contract data. If customer numbers drift apart in ERP and CRM, the shop shows wrong prices. Visibly so. The data foundation comes before the feature.

Special terms that nobody tracks systematically are quiet margin killers. Modelled digitally they become visible, measurable and cancellable. Often the project pays for itself through the terms it uncovers.

74 per cent of buyers switch supplier when ordering gets too complicated. For an A customer the channel becomes serious the moment they see their real price in the shop. Before that it stays decoration.

Product data sounds like administration. It decides speed. How fast is a new range online, a new marketplace supplied, a catalogue printed? In wholesale, data maintenance is time to market.

The ERP is built for stock and bookings. Marketing copy, media and channel variants do not belong in it. A PIM alongside relieves the ERP instead of replacing it. The split stays clear: the ERP owns the numbers, the PIM the content.

Only 12 per cent of companies hold data of AI-ready quality. Anyone who wants AI in procurement, pricing or service lays the groundwork today with a product data foundation. Without it every AI case stays blind.

Add up the hours product management and sales spend searching, retyping and asking. Data maintenance without a system is a personnel cost block. It grows without anyone deciding on it.

Rich product data sells: images, attributes, data sheets, availability. Poor product data creates questions and returns. The difference is not a budget topic but a question of process.

The biggest risk is not the switch. It is waiting. Every year on the legacy system is a year in which competitors build ordering journeys your customers prefer. A staged migration stays controllable, standstill does not.

Big-bang migrations fail on simultaneity. Everything new, everything on one date. Stages with clear cuts are the alternative: first foundation and data, then checkout and integrations, then the finishing touches. After every stage the system runs in production.

The migration is the best chance in years to clean up legacy data. Duplicates, dead articles, contradictory customer records. Take them along unchecked and you migrate your problems into a faster system.

Do not compare licence against licence, compare total cost. That includes maintenance, agency hours for workarounds and orders lost to outages. 81 per cent of buyers weight the digital experience higher than two years ago. The legacy system gets more expensive every year, even when the invoice stays the same.

Replatforming is the chance to think the ordering journey from the customer rather than from the system. The benchmark is not the competitor. It is the best shop your buyers use privately.