How to choose a B2B ecommerce platform

Most platform comparisons are written for shops that sell one price to everybody. B2B does not work that way, and the differences that decide the question are the ones a feature table is worst at showing. A platform can tick every box on the grid and still be unable to price an order correctly for your second largest customer.

These are the five questions worth asking, in the order they tend to kill options.

1. Can it price the same product differently per customer?

This is the question that eliminates most consumer platforms, and it is worth asking first for that reason. A B2B contract is usually a rule rather than a list: fifteen per cent across the catalogue, except safety footwear, which stays at list. What you need is a platform that can hold that sentence as a rule.

The alternative is that somebody types the discounted price against every article, for every customer on a contract. With a catalogue of thirty thousand articles that is not a data-entry problem, it is a permanent liability, because the moment the supplier changes a price every typed figure is silently wrong.

2. Can a customer order by article number?

B2B buyers do not browse. They arrive with a list, often pasted out of a spreadsheet or copied off last month’s invoice, and they know the article numbers. A search built for shoppers, which corrects typos and offers alternatives, is actively wrong here: a mistyped digit is a different product, and auto-correcting it into the wrong one puts the mistake in the delivery rather than on the screen.

What you want is exact matching on article numbers and a page that turns a pasted list into a filled basket in one step. If that is missing, the phone stays the fastest way to order, and the site never becomes where customers buy.

3. Where does the catalogue actually come from?

Most distributors do not own their catalogue. It comes from a supplier, it changes daily, and it arrives as files rather than as a tidy import. The real question is not whether the platform can import products. It is whether it can keep importing them, unattended, without a person in the loop.

Ask what happens when a transfer fails halfway. A sync that deletes what it cannot find will empty your shop the first time the supplier’s server has a bad morning. A sync that only ever adds and updates degrades into staleness instead, which is recoverable.

4. Does tax resolve per article or per site?

A general catalogue crosses tax rates. Office supplies and coffee are not taxed alike in most of Europe, and a platform that treats the rate as one site-wide setting will get one of them wrong on every order. This is a small technical detail that turns into an accounting problem, and it is easy to miss during a demo because demos use one product.

5. What do you own if you leave?

Two things are worth separating here. The first is the catalogue and the customer data, which you should be able to export in full at any time. The second is the pricing logic, which is the part nobody thinks to ask about and the part that is most expensive to rebuild. Contract rules encoded in a platform you cannot export are a reason you will not leave, which is not the same as a reason to stay.

Per-order platform fees belong in this question rather than in a pricing comparison. A percentage of every order is a cost that grows exactly as fast as your success does, and it is worth pricing over the life of the platform rather than per month.

A worked example

Vendox supplies office products, hygiene goods and workwear to Swedish businesses, carrying roughly thirty to forty thousand articles from a distributor. Every one of the five questions above decided something about their build.

• Contract pricing as rules. Their terms are expressed as rules over the catalogue rather than as thousands of typed prices, so a supplier price change flows through instead of invalidating a spreadsheet.
• Ordering by article number. Search matches article numbers exactly and never corrects them, and a quick-order page turns a pasted list into a filled basket.
• An hourly supplier pipeline. Five separate supplier files are joined on article number every hour. The sync never deletes, so a broken transfer degrades rather than emptying the shop.
• Tax per article. Swedish VAT resolves per article rather than site-wide, which matters the moment a catalogue holds both stationery and food.
• Owned outright. The platform is open source and self-hosted, so there is no per-order fee and the pricing logic is theirs.

The point of the example is not the stack. It is that the answers came from the business rather than from a comparison table. A distributor whose customers all pay list price, or whose catalogue is a hundred articles they own themselves, would answer these five questions differently and should choose differently.

What to ignore

Template counts, app-store totals and headline plan prices are the three loudest numbers in this decision and the three least useful. None of them tells you whether the platform can price an order for your customer, and the first two describe a shop window rather than a shop.