Every B2B re-platforming project we have been called into late had the same origin story. Somebody chose a consumer commerce platform, assumed the B2B requirements were a plugin away, and discovered eighteen months later that the domain model was wrong in a way no plugin can fix.
The mistake is understandable. On the surface a B2B storefront looks like a B2C one: catalogue, cart, checkout. The difference is that in B2C the buyer is a person, and in B2B the buyer is an organisation. Everything downstream of that changes.
The three things that break first
Identity is the first casualty. A consumer platform models one account, one email, one address book. A distributor's customer is a parent company with fourteen sites, each with purchasers who can order up to a limit, a manager who approves above it, and a finance contact who never logs in but receives every invoice. Retrofitting that produces a permissions table nobody dares to touch.
Pricing is the second. Consumer pricing is a number with optional promotions. B2B pricing is a negotiated price book per customer, layered with volume breaks, contract overrides, promotional overlays and a margin floor that must never be breached. Order of evaluation matters, and it must be deterministic — because a sales rep will be asked to explain a price and 'the system computed it' is not an answer.
Fulfilment is the third. Consumer orders ship once. B2B orders split across warehouses, partially backorder, get released against credit terms, and are frequently amended after placement. If your order model assumes an order has one state, you will end up encoding the real states in a notes field.
What re-platforming actually costs
The line-item cost of a migration is rarely what breaks the budget. Three costs consistently exceed the estimate:
- Data archaeology. Price books that exist only in a spreadsheet a regional manager maintains, and customer hierarchies that were never modelled anywhere.
- Integration surface. Every ERP, WMS, tax engine and carrier connection has to be rebuilt and reconciled twice, because you run both systems in parallel for a period.
- Behaviour change. Sales reps who built workarounds over a decade will keep using them unless the new path is genuinely faster, not merely correct.
The platform decision is cheap. The domain model you commit to is what you pay for over the following five years.
The strangler-fig route
We almost never recommend a big-bang cutover for a business that is trading. Instead, put the new system in front of a narrow, well-defined slice — one region, one customer segment, one product line — with the old system still authoritative behind it. Move the write path last.
It is slower on paper and dramatically faster in practice, because every problem surfaces while the blast radius is small and the rollback is a routing change rather than a weekend.
How to know if you have this problem
Three questions. Can a customer's purchasing manager approve an order above a subordinate's limit without a phone call? Can you explain any price on any order from data in the system? Can an order be partially fulfilled, amended and re-priced without a manual credit note? If the answer to any of those is no, you are not running a B2B commerce platform — you are running a consumer one with people filling the gaps.
Working on this?
We run a paid two-week diagnostic that ends with an architecture record, a risk register and a costed plan — yours to keep either way.
Talk to an engineer