Laboratory supplies distributor · New product
B2B PunchOut platform
A catalog connected to the customer's procurement system, so they build and send their own orders without waiting on the distributor.
The cover and the plant picker are recreated with a neutral system. The other screens are the original Figma files, with the brand, customer and suppliers replaced with fictional ones.
- The problem
- Every order from a corporate customer took one or two days to move through request, quote, purchase order and handling in the ERP.
- What I did
- I led the project end to end in one month, from business and systems analysis to interface design, every state and final testing.
- Outcome
- According to ERP records, time from quote to order dropped by about 80%; quotes and orders arrive instantly. The platform is in production.
- Role
- Project Lead, Analyst and Product Designer
- Duration
- 1 month
- Team
- Front-end and back-end development, QA and PM
- Year
- 2026
- Status
- In production
Context
One of the distributor's corporate customers bought through a manual back-and-forth: they sent a request, got a quote, sent back a purchase order, and everything was managed in the distributor's ERP. Each order took one or two days to be ready for processing.
The customer wanted to buy from their own procurement system, on their own schedule.
Who it's for
A corporate buyer who isn't the distributor's usual user. They arrive from their own system, in a hurry, inside a session that expires, facing a catalog full of technical data: SKU, CAS number, pack size, compliance and lead time.
On the other side was the distributor's team, who needed every order to reach the ERP complete and consistent.
The proposal
A PunchOut platform. From their procurement system, the buyer enters the distributor's catalog, where they only see the products and prices in their contract. They build their order and send it back to their system, where internal approval continues. The order reaches the ERP already structured. There's no checkout or payment: the session ends when the cart is transferred.
The flow
Session check, privacy notice, plant selection, catalog, cart and transfer.
Delivery address
Choose where this purchase order will be received.
Decisions and what they cost
Plant comes before the catalog
One customer has several plants, each with its own address and delivery type, partial or full. Choosing it on entry makes it clear from the start where the order is going, and that context stays visible for the whole session.
The cost is one more step before seeing products. It's worth it: an order sent to the wrong address costs far more than a click.
Session time always in view
Sessions expire, and losing a built cart is the worst possible error in this product. The remaining time lives in the top bar and, before it runs out, a notice offers to continue. Ending the session also asks for confirmation.
The cost is that a visible countdown can pressure the buyer. That's why the notice offers to continue instead of just warning.
The cart lives next to the catalog
The buyer isn't browsing, they're building an order. The cart is a fixed panel with subtotal, VAT and total that updates as products are added.
The cost is less room for the catalog. In return, the buyer always sees the total, which is what they check before transferring.
Designed for real data
I tested the design with product names hundreds of characters long, twelve-digit prices and seven-digit quantities. In the catalog the name is truncated and readable in full on hover. In the cart it's shown whole, because that's where the buyer checks what they're about to order.
I also designed every state before handoff: loading, empty results, empty cart, session error, cancellation, expiring session and transfer in progress.
Testing
On the distributor's side we tested with the QA team: smoke, usability and navigation tests, plus heatmaps to see where attention landed. Testing with the buyers was run by the customer.
Navigation needed no changes. The fixes were visual, and the rest was technical, in how the cart transfers between systems.
Outcome
Quotes and purchase orders now reach the ERP instantly and, according to ERP records, time from quote to order dropped by about 80%. What remains is internal order processing. The customer quotes and buys on their own, from their own system.
What I learned
What worked. Defining the flow and every state during analysis showed up in testing: nobody got lost, and the team could focus on the technical side.
What I'd do differently. Validation with the customer's buyers was the weakest part: the customer ran it and I didn't have a record of who took part. Next time I'd measure real usage in production, such as time per order and sessions that expire before transfer.