Paula Rosa Kitchens
Scoping a D2C MVP for a B2B Kitchen Retailer
Client
Paula Rosa Kitchens
Year
2025
Discipline
Service Design
Platform
Web (E-Com)
Timeline
4 months
My Role
UX & Service Design
THE TEAM
UX & Service Designer (Me)
Strategist
Account Director
Project Manager
Chief Strategy Officer
Client stakeholder
WHAT I OWNED
Co-led customer discovery (6/12 interviews)
Service blueprint build & delivery
Product data architecture (OOUX & relationship mapping)
MVP tickets & wishlisting pushback
THE PROBLEM
Paula Rosa Kitchens had the manufacturing capability and the product range to compete in the D2C kitchen market however what they didn't have was a website, a consumer-facing catalogue, or any operational infrastructure built for individual orders.
Their entire business was structured around volume B2B, 50 identical units to a single contractor on a building site, not the one bespoke kitchen to a homeowner on a cul-de-sac in Bristol.
THE BRIEF
The brief was to scope an MVP matching their target competitor, DIY Kitchens' core e-commerce experience.
Alongside a senior strategist, we had to work out whether that was achievable given where Paula Rosa's operations were starting from, and what it would take to get there.
DISCOVERY: THE BUSINESS
The decision
My first decision was to sequence discovery deliberately; map Paula Rosa's existing operations before speaking to customers. If we understood what the business could actually support, we could design a D2C service that worked with the existing model rather than assuming capabilities that weren't there.
Going to customers first would have generated an idealised experience brief Paula Rosa's operations couldn't fulfil.
Who we spoke to
We spoke with teams all across the Paula Rosa business; manufacturing, marketing, operations, design support. We expected reasonably mature internal systems. What we found was almost entirely manual. Four operational gaps, each with a direct implication for what could realistically launch:
What we found
Manual order processing, built for legacy tools
Product data structured for internal use only
No digital support layer beyond phone agents
Last-mile delivery unsuited to residential drops
Tour around the warehouse
DISCOVERY: THE KITCHEN BUYING JOURNEY
The decision
We ran 12 in-depth interviews with recent kitchen buyers across DIY Kitchens, Howdens, Magnet, IKEA, B&Q and Wren - covering the full journey from deciding to renovate through to in-life ownership. Three findings directly shaped the MVP.
What we found
Choice paralysis was a big barrier to purchase
Trust is often won before anyone visits the website
The planning tool was a source of real anxiety
Quotes from the sessions
HOW I AI'D
Stress tested the research plan against ChatGPT acting as a researcher, validated the discussion guide, plugged gaps, framed hypotheses
Ran interviews with transcription in the background
Loaded every transcript into NotebookLM, including a generated podcast feature, as the team's basecamp for customer understanding
Observation
NotebookLM surfaces findings and verbatim quotes well, but struggles with real insight, the implicit, unsaid thing sitting underneath the findings. That leap still had to be mine.
THE SERVICE BLUEPRINT
Discovery gave us a detailed picture of where Paula Rosa's operations would struggle under D2C demand. The service blueprint was how we turned that into something the business could act on.
It mapped the full end-to-end D2C service; the customer journey with various touchpoints on one axis and the internal processes, roles, and systems required to support them on the other. The goal was to give operations and the IT team a shared, concrete picture of what needed to change and in what order, before any technical scoping began.
WHO OWNED WHAT?
Step's 1-3: Collaborative
The first three steps; creating the framework, doing a first pass of the customer journey, touchpoints, and frontstage/backstage actions were built collaboratively with the strategist in a series of working sessions.
Steps 4-7: I took ownership
After we were aligned, I took ownership of the blueprint itself: building the full deliverable in Miro, resolving the returns and redelivery logic, and presenting the final blueprint back to the business stakeholders.
GETTING TO THE FINAL BLUEPRINT
Step 1: Understanding the existing process
Step 2: Building the foundations
Step 3: Plotting frontstage & backstage interactions
Step 4: Making sense of the complex world of returns and redelivery
Step 5: Getting deep into returns and redelivery
Step 6: Creating the Final Blueprint
Zooming in on the customer emotional journey, bringing in quotes & findings from research
Detailing all of the different systems used in frontstage and backstage environments
We overdelivered by creating a list of CRM touchpoints the customer would receive throughout the lifecycle
HOW I AI'D
ChatGPT was instrumental in building the service blueprint, validating flows between services and acting as a sparring partner on shape and structure
NotebookLM pulled out key quotes for each step of the customer journey
Both helped compose the CRM touchpoints list
SERVICE BLUEPRINT: KEY AREAS
Four areas that shaped the service blueprint.
Manual order processing
The existing delivery model wasn't designed for residential addresses or small parcel returns. We recommended a third-party last-mile partner to handle flexible deliveries and returns, a dependency that had to be resolved before launch.
Product data built for internal use
B2B returns could take 1–2 weeks without issue. A consumer waiting with two loose cabinet doors in a half-finished kitchen is a different problem. We designed a returns flow that triggered a replacement order simultaneously in the e-commerce platform and ERP, automatically generating a redelivery and collection request.
No digital support layer
Internal roles were built around B2B account management, deep product specialists helping developers configure large orders. D2C required generalist customer agents able to handle quick consumer queries over chat or phone. Several roles therefore needed redefining.
Last-mile delivery an issue for B2C
Mapping the full communication timeline revealed Paula Rosa had no email notification or CRM strategy for consumer lifecycle marketing. The blueprint became the foundation for the marketing team's communication plan, a deliverable that hadn't initially been scoped at the start of the project.
PRODUCT DATA ARCHITECTURE
The gap
The competitor audit confirmed that credible kitchen e-commerce runs on precise, well-structured product data, DIY Kitchens and Howdens had the most granular filtering and specifications of any competitor. A key audience and use case are kitchen fitters who need to locate an exact item quickly. Paula Rosa's catalogue wasn't currently built to support that.
To understand the full scope of what needed to change, I used OOUX (object-oriented UX) to map not just what products existed, but how they related to each other. Cabinets have doors; doors have compatible handles; handles have finishes. That relationship mapping directly informed the cross-sell logic and product detail page architecture and flagged that just a standard feature audit would have missed the dependency chain entirely.
From OOUX to entity relationship mapping
Once the object relationships were mapped, I translated them into an entity relationship diagram for the tech team, showing how product components connected at a data level across product pages, the planning tool, and the catalogue. This fed directly into technical sizing and shaped what was feasible to include in MVP scope versus what would need to come later.
Mapping the various attributes of the components that made up the product catalogue
Showing the relationships between different components to inform the tech team.
SCOPING THE MVP
The strategist led MVP scoping against the three principles above. I defined the early lifecycle tool options, wrote the user stories in Jira, sized design effort per feature, and pushed the deprioritisation rationale on wishlisting specifically, see below.
User story map detailing the different pages & functionality needed for MVP
PRINCIPLE 1
Parity First
Match DIY Kitchens on the fundamentals. Their experience was tried and tested and meant Paula Rosa could get to market quickly.
PRINCIPLE 2
Then differentiate
Target the early lifecycle experience, our research had confirmed that it was a clear gap and that competitors were ignoring it.
PRINCIPLE 3
Stay lean
Buy as opposed to build. Lean on third-party tools that could handle complexity, planning, logistics and hand off.
The differentiator that wasn't in the brief
The early lifecycle tool, a lightweight range finder and cost estimator, came directly from research. Customers were abandoning before they reached the planner because the jump from browsing to planning felt too large. The tool would give customers a starting point before entering the third-party planner: pre-saved selections, no blank slate and warmer leads for marketing. We scoped four fidelity options for the client so they could make a cost-informed decision on how far to take it.
THE CALL I PUSHED BACK ON
Wishlisting was cut on cost grounds. I disagreed. Saving items from the catalogue directly into the planner would have reduced the blank-slate anxiety our research flagged, and made the handoff between Paula Rosa's site and the third-party tool feel connected rather than disjointed. It was deprioritised, but I'd call it the highest-value quick win post-launch, the kind of feature that sits at the intersection of user anxiety and conversion, which is where the commercial case is easiest to make.
WHAT HAPPENED?
The stall
The project stalled after handover. Paula Rosa aligned with the strategic vision, but the gap between pre and post-MVP operational investment created a budget constraint that paused progress. That constraint traced back to the build costing, the Jira tickets I wrote were what the dev team used to size and cost the work, and that costing exercise is what ultimately created the gap. Paula Rosa opted to launch the parity build first and introduce the early lifecycle tool in a later phase, a reasonable sequencing call given where they were starting from.
What held up
Across a personnel change and a period of budget pause, an incoming designer was able to deliver directly from the original outputs without needing to rebuild the rationale, proof that the documentation was concise and easy to pick up from.
WHAT I’D TRACK IN THE FIRST 90 DAYS
PRINCIPLE 1
Planner entry conversion
Are there enough entry points into the tool? This was the biggest distinction in raising AOV. Poor planner conversion would be detrimental to the experience.
PRINCIPLE 2
Planner abandon rate
The primary signal of friction. A high rate validates the lifecycle tool argument retrospectively.
PRINCIPLE 3
Average order value
Whether customers are building full kitchens or browsing single items. The lifecycle tool hypothesis lives or dies here.
WHAT I’D DO DIFFERENTLY
Prototype and test the early lifecycle tool before scoping it: The early lifecycle tool was our strongest differentiator and the feature with the most build complexity but we went into MVP scoping without having validated the concept with a single user. A lightweight prototype tested with five or six recent kitchen buyers would have given us the evidence to push harder for it, pick the right fidelity level with confidence, and make the case to the client from observed behaviour rather than research inference.
Push for a less legacy bound brief: The other thing I'd challenge is the brief itself. The project was anchored to replicating DIY Kitchens, which was commercially pragmatic but creatively limiting. My instinct was that the bigger opportunity was a more guided buying experience — inspiration-led entry, a personality questionnaire mapping to specific ranges, feeding into the planner with pre-configured selections. Less traditional e-commerce, more guided purchase journey. Given more time, I'd have built the commercial case for it properly rather than letting the brief set the ceiling.