Validating a future-state shopping experience
Woolworths had a bold vision for the future of online shopping. Before committing to it, we wanted to know whether it actually worked for real people. I led the research and concept-validation work that put that vision in front of customers, pairing Jobs-to-be-Done discovery with moderated prototype testing to understand how people really shop, and whether the future experience genuinely made things easier.
Executive summary
I led a qualitative research and concept-validation stream for a future-state e-commerce programme, pairing Jobs-to-be-Done discovery with moderated prototype testing. The short version: the direction held up. Customers responded well to where the experience was heading, and the research also handed us a clear, focused set of opportunities to make it better, mostly around helping people find things, trust fulfilment, pay easily, and see content that actually felt relevant to them.
Project context
Woolworths had already done the hard thinking and landed on a strong vision for its future online experience. What it needed next was honest customer evidence before betting on it. We still had real questions we couldn't answer from the inside: how do people actually decide on fulfilment? Where do they start trusting the experience, and where do they lose it? How do they trade speed against reassurance, and would the new direction hold up against the messy reality of how people really shop?
My job was to bring that customer reality into the room, so the next round of design and product decisions could rest on evidence rather than our best guesses.
Role & ownership
I sat right where product design, research and strategy meet. My job was to give the research a clear shape, keep it honest and grounded in how customers actually behave, and turn everything we heard into a story the wider team could act on with confidence. Specifically, I owned:
- Shaping the overall research approach across discovery and concept validation, and defining the objectives with the wider team.
- Structuring the research plan and methodology, and guiding the prototype-testing approach, including task and scenario design.
- Synthesising findings across notes, transcripts, recordings and team debriefs, and identifying the strongest patterns, insights and design implications.
- Translating the work into stakeholder-ready outputs that could support product and design prioritisation.
The wider team ran fieldwork across locations, supported note-taking and participants during sessions, created and iterated the prototype, and contributed to daily synthesis and cross-functional decision-making.
Problem
We needed to help the team make confident calls about the future experience before we'd gone too far down the build. The direction felt right, but 'felt right' isn't evidence. We needed to know whether it genuinely held up against real shopping missions, fulfilment expectations, and the moments where trust is won or lost.
For customers, the real question was simple: would this actually make shopping feel easier, clearer and more trustworthy? For the business, it was about taking the risk down, validating the direction early enough to shape what got prioritised, and being honest about where the friction sat outside of design altogether.
Constraints
- Time pressure. The research had to be planned and executed quickly, with limited preparation time before fieldwork began.
- Prototype limitations. We tested a large, end-to-end journey in Figma, which constrained realism, stability and speed across more complex flows.
- Operational complexity. Run in live retail environments, the work picked up noise, connectivity issues, charging limitations and inconsistent recording conditions.
- Recruitment & capacity. Guerrilla intercepts meant less control over who participated, and not everyone could dedicate full-time capacity, affecting coverage across locations.
- Cross-functional dependency. Some of the biggest issues - retrieval, catalogue alignment, fulfilment logic, delivery trust - were not design problems alone, and would need wider product and operational ownership.
Research & evidence
The work ran in two phases - discovery, then validation - delivered as in-person guerrilla intercepts and moderated prototype tests across shopping locations, primarily in Cape Town (V&A Waterfront, Canal Walk, Tygervalley and Palmyra Junction), with additional reach into Gauteng and Hilton-adjacent areas through network-based recruitment. I combined qualitative discovery with moderated concept validation:
- Customer interviews using a Jobs-to-be-Done lens, to understand real shopping missions, motivations and decision-making.
- Moderated usability testing across future-state journeys, to assess clarity, trust, usability and overall fit.
- Selective A/B comparisons to explore how different design approaches were landing.
- Structured synthesis across notes, recordings and debriefs to find patterns with consistency, plus workshops to align on the questions that mattered most.
What made the approach valuable is that it did more than collect feedback on a concept - it connected the proposed future experience back to how customers actually approach shopping, decisions and fulfilment in context.
A few findings strongly shaped the direction:
Speed and ease mattered, but only when customers still felt they could trust the experience, understand their choices and complete the task with reassurance.
Many issues framed as navigation turned out to be retrieval problems: customers knew what they wanted and how they expected to reach it, but still could not find or confirm the right item.
Delivery type, timing and confidence in the service promise strongly influenced whether a journey felt viable, especially in more urgent missions.
The future-state direction was broadly validated. The work pointed not to reinvention, but to important refinements around retrieval, fulfilment clarity, payment ease and content relevance.
“Set fulfilment, then go to food then search.”
“I’m not finding what I’m looking for.”
“Loved the fact that the cards are pre-loaded.”
Key decisions
We stopped treating the core issue as navigation alone. The future-state experience focused on faster, clearer retrieval - stronger search-first behaviour, cleaner category entry, more direct task completion - while making clear that some of the most meaningful fixes sit beyond design, across search, catalogue logic, stock visibility and fulfilment context.
We explored fulfilment as a more visible but flexible layer across the journey, rather than an early gate or a late checkout detail. Testing showed customers preferred a more detailed fulfilment toggle over a stripped-back one, because it gave enough context to trust the choice.
The useful principle wasn’t whether inspiration appears early or late, but relevance. Customers were open to broader inspiration early on, then expected content to become much more specific once they showed intent - shaping homepage, category and in-journey discovery.
Customers didn’t want the shortest journey at any cost - they wanted a fast experience they could still sense-check at the moments that mattered. This pointed to targeted refinements in cart, checkout, payment and confirmation around reassurance, visibility and repeat-payment ease, rather than a major redesign.
Design response
Across the stream I designed and / or validated:
- Future-state end-to-end journeys across Food, Fashion, Beauty and Home
- Fulfilment selection patterns and toggle variants
- Search-led and category-led shopping paths
- Homepage and in-journey inspiration approaches
- Cart, checkout, payment and confirmation flows
- Selected A/B variants to test clarity, preference and trust
The overall direction was validated, but the work identified where refinement would make the greatest difference. Rather than driving a reset, it helped the team make more confident decisions around retrieval, fulfilment clarity, content relevance, payment ease and trust-building moments across the journey.
Impact
What we walked away with was a much stronger foundation for the decisions ahead, at exactly the point it mattered most. Across two phases we spoke to 59 people - 30 in discovery and 29 in validation - in short, focused sessions of 10 to 15 minutes each. It was enough to say with confidence that the direction was sound, to point clearly at what to refine first, and to tell the difference between an interface problem and something bigger sitting in product or operations.
- Validated that the overall concept was directionally strong.
- Uncovered clearer customer jobs, trust needs and fulfilment behaviours.
- Created a more credible basis for prioritisation across design, product and engineering, and a reusable evidence base for future decisions.
Not all of the impact was measurable in hard numbers at this stage, because this was early qualitative research and concept validation rather than a live launch. The value was in reducing uncertainty, strengthening confidence in the direction, and making the next decisions more grounded.
Reflection
More than anything, this work reminded me how valuable it is to test future thinking against real behaviour before the decisions harden. The learning that stuck with me most: some of the biggest pain points weren't really design problems at all. The research surfaced where things like fulfilment logic, product retrieval and platform behaviour were quietly shaping the experience, and naming that honestly made our recommendations stronger and more realistic.
If I did it again, I'd carve out more prep time upfront, and for the most complex end-to-end journeys I'd push for a sturdier build than a prototype. I also came away with real respect for the operational side of research at this scale - consistent note-taking, daily synthesis and enough hands on deck made a genuine difference to how good the output was.
AI workflow note
AI played a supporting role in the research workflow, mainly helping the team process and synthesise a large volume of qualitative data more efficiently - organising notes, reviewing transcripts, identifying repeated themes and structuring early synthesis outputs.
It was never a replacement for researcher judgement. All outputs were reviewed, checked and interpreted by the team, with human judgement central to what made it into the final findings. AI helped us move faster through the volume of material; the meaning-making, prioritisation and final calls stayed with the researchers and designers.
A useful learning: using AI well still requires strong operational rigour. Notes, transcripts and recordings had to be prepared in a consistent, usable format before they could be analysed effectively - so while AI accelerated synthesis, it also reinforced the importance of clean inputs, clear structure and active human review.
Most valuable as an efficiency and synthesis support tool - helping the team work through large amounts of qualitative evidence quickly, while keeping the research grounded, reviewed and human-led.