Key points
- Onboarding customers to a B2B portal starts with your most repetitive buyers, not your largest account. Repetition gives the fastest payoff and the lowest risk, because those buyers know the range well enough to spot an error themselves.
- Before you invite a customer, three things must exist in their account: their prices, their order history and their assortment. An empty account ends in one login and a return to email.
- Adoption dies on small procedural frictions: a password nobody can reset, one shared login for the customer's entire company, no split between the person who orders and the person who approves.
- Your reps need a stake in the platform: visibility of their customers' orders, the ability to place an order on a customer's behalf, and commission calculated per customer rather than per channel. Commission rules are settled before the first invitation, not after.
The integration works, so the platform shows prices from your system, live stock and order statuses, and whatever a customer submits lands in the ERP without retyping. Technically you are ready. This is where most B2B rollouts come apart: the platform exists, it works correctly, and nobody uses it. Customers keep emailing their orders, reps keep retyping them, and the only person who visits the portal once a month is the one who commissioned it. This chapter is about the two groups the outcome depends on: the customers you have to bring onto the platform, and the sales reps without whom no customer will ever get there.
What has to be ready before you invite the first customer
An invitation goes to a specific company rather than to a contact list, so readiness is checked account by account. Three things have to be in place: their prices, their order history and their assortment. Prices means their terms after discounts and volume breaks, not the list price. History means what they bought over recent months, with a way to repeat an order in one move. Assortment means the few dozen items they actually buy, pulled to the front rather than buried in a catalog of several thousand SKUs. A customer who logs in to list prices and an empty home screen is back on email the same day, and you will not get a second chance to convince them.
Then come the things whose absence only surfaces on the first order: delivery addresses (usually more than one), invoicing details, payment terms, contact people on the customer side. The readiness test is simple and worth running before the invitation goes out: log in to their account and place their typical order without knowing the catalog by heart. If it takes more than a few minutes, or requires knowledge the customer does not have, the account is not ready. What such an account covers in terms of features we describe separately under our B2B ordering portal service.
Sequence: the most repetitive buyers first
You pick your first customers on one criterion: how repetitive their orders are. A company placing several orders a month, mostly of the same items, is the ideal first user. It has the most to gain, it knows the range well enough to catch a wrong price or a wrong unit, and it comes back often enough to actually learn the new way of ordering. Do not start with your largest account, where the cost of a mistake is highest and orders tend to be unusual. Do not start with a customer who orders once a quarter, because on their second login everything will be new again. And do not start with the one who shouts loudest for new features, because their wish list will stall the launch.
The invitation and the first login
A good invitation is personal and announced in advance. The rep who handles that customer mentions the platform in a meeting or on the phone, explains something concrete ("you will have your own prices and your order history, and repeating your last order is one click"), and only then does the email with the link go out. A mass mailing from the system titled "Our new B2B portal is live" lands somewhere between newsletters. The difference is not in the copy, it is in who stands behind the invitation and whether the customer knows what to expect.
Test the first login on yourself before a customer does. The password setup link has to stay valid long enough to survive a weekend and a holiday. Password recovery has to work without a call to the rep, because a forgotten password is the single most common reason an account goes quiet after two weeks. The first order is best placed together with the customer, on the phone or in a meeting, starting from a repeat of their last order. That costs fifteen minutes per customer and is the most effective part of the whole rollout, because after one successful order the perceived risk drops from "a new system" to "the same as always, only faster".
The invitation comes from a rep, not from the system. A wholesale customer does not open an account because a supplier emailed them. They open it because the person they have dealt with for years told them it would save their team time, and showed them what their own screen looks like after logging in. Automated invitations make sense later, as a supplement, once part of your base already orders on its own.
The frictions that kill adoption in week one
The most common friction is not a missing feature, it is one shared login for the customer's entire company. On their side the warehouse orders, purchasing approves, the owner wants to see prices and accounting looks for invoices. A single account then circulates around the company and within weeks nobody knows who ordered what. The fix is simple, but it has to exist before the invitation: multiple users within one customer account and a basic split of roles, meaning who can place orders, who can only browse, who sees prices and documents. It also helps if the customer, rather than your service desk, can add another person on their side. Every step that requires phoning the supplier is a point where the process reverts to email.
Encouragement rather than compulsion
Switching off email and phone orders at launch is tempting and almost always premature. A customer who hits friction with no alternative does not start using the platform, they call to complain. Encourage with what does not exist outside the platform: order status without asking a rep, availability and delivery dates, history, documents to download, a one-click repeat of the previous order. A price incentive, meaning a discount for using the portal, looks like the easiest lever but teaches customers that the channel has a price of its own, and it comes back at every negotiation. If you want to reward the channel, do it with process rather than price: an order placed by the customer needs no retyping, so it can carry a later cut-off for same-day dispatch.
What your sales reps are actually afraid of
A rep who does not want the platform rarely says so directly. They say their customers "are a special case", that they "prefer the phone", that "it will not work here". Underneath there are usually two real fears: about commission and about the relationship. Both are worth naming out loud before the rollout starts, because unnamed fears do not disappear, they turn into passive resistance that no software can outrun.
| What the rep says | What sits behind it | How to set up the platform and the rules |
|---|---|---|
| "I will lose commission on orders customers place themselves" | Suspicion that platform orders will fall outside their numbers | Commission calculated per customer, not per channel, agreed and written down before the first invitation |
| "Customers will stop calling me" | Treating the relationship as identical to taking repeat orders | Visibility of their customers' orders and carts, plus an alert when a customer stops buying a regular item; the conversation moves to new products and terms |
| "A customer will order the wrong thing and I will fix it" | Experience with mistakes in email orders | Placing orders on a customer's behalf, and reviewing an order before it is confirmed |
| "Everyone will see what my account base really looks like" | Fear of being judged, not fear of the tool | The same data available to the rep as well; agree what counts as a measure of their work before the first reports appear |
The common thread: a rep should be using the platform as their own working tool before they start recommending it to customers.
The platform helps a rep when it gives them something they did not have before. Three features make the biggest difference: live visibility of their customers' orders and carts, the ability to place an order on a customer's behalf at that customer's prices (useful on calls and in meetings, and incidentally the fastest cure for the fear of "a system that replaces me"), and simple signals such as who has not logged in for a long time, or who stopped buying an item they used to order monthly. Separately, agree what the time recovered from retyping orders and answering "has my delivery shipped" is meant to go into. If nobody decides that, the time dissolves into daily noise and a quarter later nobody can point to a single benefit.
Commission: decide before launch, not after
Commission is the one item that can kill a rollout within a week. The variant of "a lower rate for orders placed without a rep" looks logical from a cost-to-serve angle and is simultaneously the simplest way to turn the entire sales team against the platform. A rep who loses money on every self-service order has an economic reason not to invite customers, and that alone is enough for the project to stall. The rule that works is simpler: commission follows the customer, not the channel. If you want to reward growth, add a separate component for winning a new customer or for activating one on the platform. Make the decision and communicate it in writing before the first invitation, because changing the rules midway costs more trust than the rest of the rollout combined.
This is an organizational change, not an IT project
It is worth saying plainly, because it sounds less comfortable than a delivery schedule: selling wholesale through a platform changes how people work, and the software is only the instrument of that change. If your reps are against it, the project dies regardless of build quality, because they decide whether a customer gets an invitation and whether anyone calls after the first failed login. The things that genuinely help are non-technical: involving reps in choosing the first customers, testing on real orders before anyone outside is invited, stating clearly what disappears from their job and what stays, and having one owner of the rollout inside the company rather than at the software vendor. Once both sides are ready, the remaining question is what order to launch in and how to tell it is working. That is the final chapter.
Questions
How many customers should you invite at the start?
As many as you can look after personally, which usually means somewhere between a handful and a dozen or so companies. The point is that someone can be on the phone for every first order and that problems surface within a group you know. Inviting the whole base at once produces one result: dozens of accounts without a single order and no information about what went wrong.
Can you require customers to order only through the platform?
You can, but not at launch and not across the board. Compulsion introduced before the path is polished turns every friction into a complaint, and it lands on customers who are under no obligation to make a supplier's life easier. In practice the sensible order is the reverse: first get your repetitive customers ordering on their own because it is faster for them, then discuss limiting the email channel, and treat that as a commercial decision rather than a technical one.
How should commission work on orders customers place themselves?
Most safely the same way as on orders taken by phone, meaning based on the customer assigned to the rep rather than on the channel. Any scheme that cuts the rate for platform orders creates a conflict of interest exactly where you need cooperation most. If cost to serve is to show up in pay, use a separate bonus component for acquisition or for activating customers, and announce the rules before go-live.
