Watch what most B2C platforms actually create when a consumer registers: a login row in an identity system, and — if marketing had budget that quarter — a lead. The person now exists twice, in two systems, as two different kinds of thing, and neither of them can be invoiced. Somewhere between their first order and their first support call, an operations person quietly re-keys them into the business system as a real customer and staples the records together with an email address. Every future mystery — the order that will not attach, the statement that misses an invoice, the duplicate account with one letter different — is born in that gap. The fix is a decision, not a feature: on a storefront, the person registering is not a prospect to be qualified — they are a customer who has not ordered yet, and the signup should provision them as one.
One signup, three records, one link
When self-signup is enabled on an xMatix portal, registration creates the whole shape in one motion: a customer account typed as a consumer, a contact under it carrying the person's name and details, and an external portal login linked to that contact from birth. The invitation machinery then takes over — the same set-password flow an invited dealer user gets — so the person verifies their email and sets a password before the login is usable. Nothing is deferred to a future migration: the moment they finish signing up, they can put something in a cart, check out, and the order lands against a customer record finance can invoice, with no human in the loop between "register" and "revenue".
Both halves are switches, not code: a portal can allow self-signup at all, and can decide whether a signup provisions a customer or merely creates a login. The default is the full provisioning — because a login that cannot transact is a support ticket with a password.
Always a fresh customer — on purpose
The obvious "improvement" is the one we refused: matching the signup email against existing customers and attaching the new login to a match. It sounds like deduplication; it is actually an account-takeover gift. A signup form is an anonymous surface — anyone can type any email address into it — and a system that links a self-declared email to an existing customer account has let a stranger claim that customer's order history, addresses and credit terms for the price of knowing where they shop. So self-signup always mints a fresh account and contact, and never adopts an existing one. Where a real person genuinely exists twice, merging is an administrator's decision, made with evidence, from inside the tenant — and inviting a known customer's contact explicitly, which links safely because the tenant chose the linkage, remains the right path for established relationships. Deduplication is a data-quality chore; adoption from an anonymous form is a vulnerability. The distinction is worth an architecture.
A consumer is not a small dealer
Provisioning is half the story; the experience is the other half. A portal built for the trade greets its users with credit exposure, outstanding statements, targets and scheme performance — exactly what a household consumer should never see, both because it is meaningless to them and because rendering a credit apparatus around a cash customer invites nonsense data into real books. The consumer's account type drives the portal the person actually gets: unit-first ordering rather than case-and-scheme ordering, consumer banners, an account area of orders and addresses rather than ledgers and targets. One deliberate asymmetry: an account the system cannot classify is treated as trade, not consumer — hiding a partner's credit position is a worse failure than showing a consumer an empty statement card. And when a signed-in consumer orders, it is the same cart, the same freeze-first checkout and the same order pipeline as everything else the platform sells — a consumer order is an ordinary order whose customer happens to have registered themselves.
The anonymous visitor gets a real front door
Before any of this, someone has to arrive. The storefront's public landing page carries the catalogue's real product families at list prices, the offer carousel, and a coverage check that answers "do you deliver to my pincode" from the same data the order router uses — honest answers for a visitor who has typed nothing but a pincode. What it does not do is invent urgency for strangers: scheme eligibility is an account-scoped fact, so a visitor sees no personalised promises until there is a person to promise to. Registration is the door between those two worlds, and because of everything above, walking through it means becoming a customer — not joining a queue to become one later. For the channel-strategy half of this story — what going direct does to the distributors who still move the volume — see The distributor is not your competitor; for the machinery, the consumer accounts documentation.