Case Study — Unite Procurement Portal

§ In brief

Strategy: I brought user evidence into the redesign process, helping the product team challenge assumptions, refine the scope and make more informed prioritisation decisions.
Research: I introduced continuous prototype validation across the core journeys — planning and moderating usability tests, synthesising findings and translating them into product decisions.
Design: I redesigned critical purchasing workflows from product discovery to checkout and returns, turning research insights into tested interaction concepts for the new marketplace.

§ 01 — The starting point

A marketplace built over 25 years.

Unite's B2B marketplace connects professional buyers and suppliers across Europe.

Over more than two decades, new features, customer-specific requirements and evolving business needs gradually shaped the platform into a complex product.

This redesign focused on operative purchasers — the people who use the platform most frequently to manage day-to-day procurement.

25 Years online
under Unite & predecessors

2 User groups
the redesign was scoped around

4 Core flows
rebuilt on evidence

§ 02 — The challenge

The opportunity wasn't just visual — it was how we decided.

Redesigning the screens was the easy part. The real challenge was solving three interconnected problems:

Growing friction across daily tasks

Unclear structures, limited orientation and unnecessary complexity made everyday tasks harder — from searching and evaluating products to checkout, returns and managing users.

Decisions based largely on assumptions

Product decisions had historically been shaped by internal requirements and assumptions. Regular usability testing was still new to the organisation.

A redesign too large to tackle at once

Rebuilding the entire marketplace in one step would have been too large and risky. The team needed to identify the most relevant problems and determine where optimisation was sufficient — and where a fundamentally new approach was required.

We needed to distinguish what still supported real buyer needs, what could be simplified, and what had to be fundamentally rethought.

§ 03 — The goal

Build what matters — not simply what already exists.

The goal wasn't to reproduce the legacy marketplace in a new interface. It was to create a clearer purchasing experience around the actual needs of operative purchasers — while preserving the workflows that still delivered value.

Three questions guided the redesign:

What should we preserve?

Which existing workflows still supported how operative purchasers actually worked?

What was missing?

Which user needs were not adequately supported by the legacy platform?

What could we simplify?

Where did unnecessary complexity make everyday purchasing tasks harder?

§ 04 — My approach

Using evidence to redesign critical workflows.

Working closely with Product Management and Development, I gradually integrated continuous validation into the redesign process.

Every core flow was tested with operative purchasers — but the method depended on the complexity and implementation risk of the change: interactive prototypes for larger, development-intensive concepts like checkout and returns; the live system for smaller, faster improvements.

Define the problem

Solve the right problem before designing the solution.

Explore solutions

Make ideas tangible before investing in development time.

Test it with buyers

Replace assumptions
with real user feedback.

Learn & refine

Improve the design
before it reaches
production.

"For the first time in the company's history, we validated ideas before building them — and made product decisions based on actual user needs."

— The shift this project introduced

§ 05 — The buyer behind the workflows

Anke, Operative Purchaser.

Anke manages recurring day-to-day purchases across products and suppliers while following internal procurement requirements.

Find and compare suitable products efficiently.
Evaluate price, availability and delivery conditions.
Follow internal purchasing rules.
Stay oriented across complex, multi-item orders.

Homepage Search Product evaluation Basket Checkout

Fig. 02 — Anke's core journey, in two sequences: Discovery (homepage, search, product) and Transaction (basket, checkout).

§ 06 — Finding the right product

Homepage, search, product page — three parts of product discovery.

Homepage, search and the product page support different parts of the product discovery journey.

While we looked at the journey as a whole, each step had its own usability problems that needed to be addressed in the redesign.

The old flow

1Homepage
2Account
3Search results
4Product page

Legacy Unite start page — top nav with search bar, exclusive suppliers row, and a dense product-range grid.

Fig. 03 — The legacy flow. Step through Homepage → Account → Search → Product page.

1Homepage & account
2Search results
3Product page

The hypothesis & story

Buyers were spending too many clicks on basic account tasks. Getting to routine tools like order history or saved lists meant opening a top dropdown, landing on a generic summary page, and picking the option a second time from a new set of tabs. I figured moving everything into a simple, permanent sidebar would cut out those extra steps and let buyers just get on with their work.

Testing insights

Repetitive clicks: Clicking a dropdown item dumped buyers onto a generic page with a second row of tabs showing the exact same options — forcing two clicks to reach one destination.

Hidden daily tools: Essential procurement workflows, like order history and shopping lists, were buried behind hover menus instead of being visible right away.

Lost context: Buyers lost their place because the navigation layout completely shifted when they moved from shopping for products to checking account settings.

What changed & why

Hidden dropdown → Persistent sidebar

Change: Replaced the top-right account dropdown with a permanent left-hand sidebar.
Why: Keeps core tools always visible so buyers can switch tasks in one click without opening extra menus.

Removed redundant tab pages

Change: Scrapped intermediate profile pages that duplicated dropdown options in a horizontal tab bar.
Why: Flattens the site structure, removing dead-end pages and keeping navigation predictable across the platform.

The hypothesis & story

Buyers were losing time trying to scan an overcrowded search page. With filters, view modes, and sorting controls scattered to the screen edges while results sat in one flat table, nothing stood out and nothing was easy to compare. I figured rebuilding the page around a clear visual hierarchy — bringing key controls front and center — would make finding products feel effortless.

Testing insights

Scattered & split filters: Attribute filters sat in a huge top block while category filters were tucked into the left sidebar. Buyers struggled to find tools because filtering was split across two separate screen areas.

Unfindable layout switch: The gallery view was labeled "Large product images" and buried at the bottom of the left sidebar. Users missed it entirely, assuming no visual grid view existed.

Hidden table sorting: Sorting was hidden inside tiny column header arrows in the results table. Buyers missed the feature because there was no obvious "Sort by" control.

What changed & why

Split filters → Unified top filter chips

Change: Combined the top block and left sidebar filters into a single row of top-aligned filter chips.
Why: Puts all refinement tools in one predictable spot right under the search bar so buyers can filter without scanning two screen edges.

Buried links → Explicit view & sort bar

Change: Moved sorting into a clear dropdown (Sort by: Relevance) and placed simple view icons (Grid / Table) right next to it.
Why: Replaces confusing text links ("Large product images") with clear visual icons, making view switching instant.

Dense table → Scannable product cards

Change: Transformed row-by-row table data into distinct product cards with clear pricing, stock badges, and direct buy controls.
Why: Groups essential decision facts together so buyers can compare products without reading across wide table columns.

The hypothesis & story

The old product page made a simple purchase feel like a multi-step puzzle. Buyers had to cross-reference top seller lists with bottom pricing tables just to figure out what they were buying, while two competing "Add to Basket" buttons added extra confusion. I hypothesized that consolidating everything into a single, unified Buy Box would eliminate the guesswork and speed up ordering.

Testing insights

Two basket buttons created confusion: Having an "Add to basket" button both in the top header area and at the bottom of the vendor table left buyers unsure which specific offer they were actually buying.

Split vendor & price data: The top table showed seller options, but buyers had to scroll down to a separate section to view volume discount tiers and shipping conditions for that specific seller.

Walls of unformatted specs: Technical details, compliance info, and return rules were crammed into a massive, raw table on the left, making it exhausting for buyers to scan for specific specs.

What changed & why

Two basket options → One primary "Buy Box"

Change: Consolidated the top seller's price, stock badge, quantity selector, and main buy button into one clear card on the right.
Why: Eliminates confusion around which offer is selected and gives buyers a single, predictable place to place an order.

Scattered vendors → "More offers" list

Change: Moved secondary seller options into clean, stacked cards directly below the main Buy Box.
Why: Keeps seller comparison close to the buy action without cluttering the screen with competing full-width tables.

Raw table grid → Collapsible accordions

Change: Structured product specs, description text, volume pricing tiers, and return policies into organized accordions (Product data, Product description, Tiered pricing, Return policy).
Why: Cleans up the initial page view while keeping deep technical details easily accessible when buyers need them.

The redesigned flow

1Homepage
2Search results
3Product page

Redesigned Unite start page — sidebar navigation, seller catalogue tiles and clear category cards.

Fig. 04 — The redesigned discovery flow. Step through Homepage → Search → Product page.

Validation Each of these changes was tested with operative purchasers. The sidebar and filter chips were understood immediately — buyers used them without prompting. The product page took longer: several rounds of the buy box still left buyers unsure which offer they were looking at, and testing hasn't reached full satisfaction yet — it's still being iterated on.

§ 07 — The transaction flow

Basket and checkout — one transaction flow.

The old checkout placed everything on a single page — addresses, payment details, cost centres, order numbers and individual line items. Especially for multi-item orders, this made information difficult to review and mistakes easy to miss.

The redesigned flow guides buyers from reviewing their basket to placing the order. Related information is grouped into four checkout steps, with a progress bar showing where buyers are in the process and what comes next.

The old flow

Legacy Unite/Mercateo basket — items, quantities, delivery and warehouse breakdown, with the proprietary basket-optimisation panel hidden.

Fig. 05 — The legacy checkout: basket, then one endless request page. Switch between the two stages above.

1Basket
2Checkout

The hypothesis & story

Buyers were struggling to make sense of a crowded, technical basket page. The old design dumped raw warehouse breakdown tables and complex optimization radio buttons right next to the checkout button, creating heavy visual noise and making order totals hard to verify at a glance. I figured separating item groupings from checkout totals and simplifying the "Smart Basket" optimization tools would make approving an order fast and stress-free.

Testing insights

Overwhelming data tables: Showing a full "Listed by warehouse" accounting table underneath the main item list forced buyers to read through redundant logistical data just to check their cart.

Competing visual priorities: The primary "Checkout" action sat above a dense green "Basket optimisation" box, making it hard for eyes to establish a natural top-to-bottom checkout flow.

Confusing optimization choices: Radio options like "Fastest delivery", "Best price", and "Without optimisation" were presented as a wall of technical stats, making buyers hesitate over which mode was actually active.

What changed & why

Raw tables → Grouped shipment cards

Change: Grouped products into clear, distinct shipment containers (Delivery 1 of 2, Delivery 2 of 2) with clear supplier info and estimated delivery dates.
Why: Shows buyers exactly how their order will ship without cluttering the page with raw warehouse breakdown tables.

Scattered totals → Dedicated checkout summary

Change: Consolidated subtotal, VAT, shipping, and the primary action button ("Transfer basket to ERP") into a clean, sticky sidebar on the right.
Why: Creates a clear visual hierarchy so buyers can instantly review final costs and proceed to checkout without distraction.

Dense radio list → Visual "Smart Basket" widget

Change: Redesigned basket optimization into a clean "Smart Basket" card placed neatly below the order summary, clearly highlighting the recommended option ("Our recommendation").
Why: Simplifies complex optimization logic into straightforward, scannable cards so buyers can trust the platform's recommendation in seconds.

The hypothesis & story

Buyers were overwhelmed by a massive, single-page checkout form that dumped addresses, payment info, cost centers, and warehouse summaries onto one screen. Trying to fill out complex accounting dropdowns inline while reviewing items caused frequent errors and cognitive overload. I figured breaking the checkout into a focused, 4-step wizard (Addresses → Payment → Cost Centres → Order Summary) would guide buyers smoothly through the process without terrifying them with a wall of fields.

Testing insights

One-page form fatigue: Putting addresses, SEPA mandates, promo codes, cost centers, and item lists on one continuous page made buyers feel like they were filling out a tax form, leading to skipped inputs.

Inline cost-center friction: Assigning cost centers and cost types directly inside the product list (with individual "Apply for all" buttons) clogged up the item review area and caused input mistakes.

Unclear checkout progress: Without step indicators, buyers couldn't tell how close they were to finishing the order or whether clicking "Send purchase order request" was final.

What changed & why

Giant single page → Linear 4-step wizard

Change: Separated checkout into 4 distinct steps with a clear, top progress bar (1. Addresses, 2. Payment, 3. Cost centres & types, 4. Order summary).
Why: Keeps the screen clean and lets buyers focus on one decision at a time without feeling overwhelmed.

Embedded forms → Dedicated step screens

Change: Gave addresses, payment details, and cost center assignments their own dedicated, full-width step cards.
Why: Eliminates visual clutter so buyers can set up complex B2B accounting rules cleanly before seeing the final summary.

Cluttered form → Final review & approval step

Change: Created a dedicated Order summary screen (Step 4) that pulls together delivery addresses, approver notes, and item breakdowns right next to the final "Send requisition" button.
Why: Gives buyers complete confidence that all details, approvals, and totals are correct before submitting the purchase order.

The redesigned flow

1Basket
2Addresses
3Payment
4Cost centres & types
5Order summary

New basket — grouped shipment cards (Delivery 1 of 6, etc.), a dedicated total sidebar, and the Smart Basket optimisation widget.

Fig. 06 — The redesigned flow. Step through Basket → Addresses → Payment → Cost centres & types → Order summary.

§ 08 — The outcome

From "we think" to "we know."

The clearest outcome isn't a single metric — it's that the way the product gets decided changed. The redesign was the visible result; the shift underneath it is what lasts.

Decisions

Grounded in evidence.

Product decisions became increasingly grounded in customer evidence rather than internal assumptions.

Process

Validated before built.

Concepts were tested with operative buyers before implementation — so effort went where the value was.

Culture

Feedback, made routine.

Customer feedback became a regular input to design, not an occasional checkpoint.

And the visible result: core procurement workflows — search, the product page, checkout, and returns — were simplified and modernised, each one shaped by testing rather than guesswork. The project is ongoing; harder numbers (task time, adoption, support load) are being tracked as flows reach more pilot users.

§ 09 — What I'd do differently

Learnings.

★ TBD · Placeholder

Honest reflections — what the evidence-based approach got right, where it was slow or hard to sell internally, and what you'd change next time. Especially powerful here: shifting an organisation from a feature-first habit to validating before building.

→ Drop in: 3 short reflections · ideally one on research, one on design, one on stakeholder buy-in

In short.

We modernised a complex, 25-year-old marketplace by replacing assumptions with evidence — and rebuilt its core workflows around how operative buyers actually work, testing what mattered before building it.

Talk to me about this project

From legacy
to intuition.

A marketplace built over 25 years.