When we pitched a CPG customer last year on a self-serve B2B portal for their distributors, three of the four executives in the room nodded along. The fourth — their general manager for general trade — pushed back hard. "Our distributors won't use it. They've been ordering through WhatsApp and phone calls for fifteen years. They like it."
He was half right. They did like it. They also hated it. The two coexisted, the way most long-standing professional relationships do. The portal launched anyway, with caveats. Nine months later it processes 73% of orders. The general manager is, somewhat sheepishly, the project's biggest internal evangelist now. Here's what happened.
The problem under the problem
The pitch for B2B self-serve usually goes: distributors can order anytime, you save call-centre cost, everybody wins. That pitch is true but it is not the real reason it works.
The real problem with WhatsApp/phone ordering isn't volume. It's information loss. When a distributor sends a WhatsApp message that says "send 240 of the small ones, the usual scheme please" — a human on the brand side has to interpret "the small ones," remember "the usual scheme," check credit, look up the price list, type it into the ERP, and reply with a confirmation that the distributor then has to either trust or audit. Every one of those steps has a meaningful error rate. And mistakes are expensive — a wrong scheme costs margin, a wrong SKU costs returns, a wrong credit decision costs working capital.
The portal isn't about removing the human. It's about removing the interpretation. The distributor sees the same SKUs, the same schemes, the same credit limit, the same shipment ETAs that the brand sees. There is no translation step.
Why the resistance was rational
The general manager's pushback was based on real evidence. Two previous attempts at a B2B portal — at this customer and across the industry — had failed for the same reason: the portal was a tax. It made distributors do more work for the same outcome.
The failure mode looks like this. The brand builds a portal that exactly mirrors their internal ERP order form. The distributor logs in, has to navigate 14 dropdowns, can't see today's scheme because schemes live in a different system, and gets a different price than expected because the discount didn't auto-apply. After three frustrating sessions, they go back to WhatsApp. Now the brand has spent six months on a project that runs at 3% adoption.
The honest way to think about a B2B portal is: it has to be visibly faster than WhatsApp from the distributor's first session. If it isn't, you've lost. WhatsApp is the bar.
The portal isn't about removing the human. It's about removing the interpretation. WhatsApp is the bar.
What we built differently
Three design choices, in roughly the order they mattered.
One: home screen is a re-order screen. When a distributor logs in, the first thing they see is "your last order, one-click re-order, adjusted for current schemes." For ~60% of B2B distributor orders, this is the only action they need. The "search the catalog" path exists, but it isn't the front door. The front door is the front door of a regular distributor's day.
Two: schemes are visible and live. Every SKU shows its applicable scheme right next to the price. If the scheme changes mid-week, the distributor sees it the next time they refresh. No more "I didn't get the scheme that was supposed to apply" arguments at month-end.
Three: credit, ETA and dispatch ETA are first-class data, not hidden. A distributor placing an order sees: their credit limit, their utilized amount, their expected ETA, their last late-payment status. Everything that would have been a back-and-forth WhatsApp is now read-only data in the order screen. The conversation collapses.
What the numbers say
At nine months in, this customer:
- Routes 73% of orders through the portal (target was 50% at 12 months — we got there in 5).
- Reduced order-handling team from 28 to 7 (the rest were re-assigned to relationship work, not laid off).
- Cut order-error rate from 4.8% to 0.6%.
- Increased average order size 14%, because schemes that distributors had been missing started getting applied.
- Reduced disputes at month-end by 81%.
The biggest surprise: distributor NPS toward the brand went up 22 points. The relationship didn't deteriorate when the human-on-phone interaction reduced. It improved, because the conversations that did happen were higher-quality — about strategy, new product launches, regional schemes — instead of "did my order go through."
What broke that we didn't expect
The biggest unforeseen issue was internal. The brand's general-trade team's compensation had been tied to volume they "handled," and the portal made that metric meaningless. We had to help the customer redesign their compensation plan around relationship outcomes (new outlets onboarded, retention, secondary sales) before adoption fully took off. That was a six-week consulting engagement we hadn't planned for.
The lesson: if you build a portal that disintermediates a job function, you have to fix the metrics for that job function too, or it gets quietly sabotaged from within. We now bake that consulting into every B2B-portal go-live.
Curious about your B2B setup? Talk to a Commerce specialist — we'll walk through your current order flow and where a portal would actually pay back.