Selected work · Product & service design
Ten years of user-centred design across government, public-sector and donor-backed organisations in Africa and Europe. I turn field research into products that busy non-specialists can actually use, and build the reusable, accessible interfaces that carry them to scale.
Designing, and then delivering, a credible way to reach small businesses for the Private Sector Foundation Uganda, the country's apex private-sector body, working alongside government and development partners.
Uganda's business support market is fragmented and uneven. Most providers cluster around start-ups in cities, quality varies with no shared standard, and good services are unaffordable without subsidy. Much of the available content is designed abroad and misses local context. The sector aims to reach 40,000 providers, but only with an offer they trust and can sustain.
I ran the research-to-product process end to end. I interviewed three customer types, providers, entrepreneurs, and funders, then clustered every finding into five recurring themes so the team could see where needs converged rather than reacting to anecdotes.
Single trainer, urban, smartphone but not an advanced user, works in Luganda and some English
End users are experienced specialists who lack the know-how to grow a business
"My goal is knowing how to grow my business and the link to more money. I don't want more training."
I moved the team off solutioning and onto a structured frame, taking each persona through problem, proposition, product, positioning and promotion in turn. Problem and positioning statements used fill-in templates so the language stayed in the user's words and the differentiator stayed explicit.
The frame that anchored the design
Every proposition had to trace back to a named persona and a problem stated in their language. The templates kept the team honest and made trade-offs visible to funders and government partners who needed to trust the reasoning before committing.
The answer was not another feature, it was a person. I designed the Growth Guide: a trained local agent based at a district Growth Centre who carries support to small businesses instead of waiting for them to find it. I designed the role, the journey behind it, and the training, so a non-specialist in a low-connectivity district can run a credible diagnostic and point an owner to the right help.
Recruited from the local advice network, stipend-supported, and trained on the diagnostic and the digital tools.
Runs the diagnostic with each business, directs them to the right support and finance options, coaches on quick wins.
Central-team check-ins and an annual re-diagnostic. Field data flows back to shape the next round of the programme.
I led the implementation, not just the design. Guides were recruited and trained, the diagnostic and journey went into the field, and the central support loop is running now. The service is built to learn: each annual diagnostic refreshes what a business needs, and what Guides see feeds the next redesign. It is live and ongoing, so headline impact figures are still to come.
The learning programme behind Bank of Industry Nigeria's national Youth Entrepreneurship Support (YES) Programme. I designed it in 2016. It was still running in 2024.
Reach thousands of young would-be entrepreneurs across Nigeria with credible business training, deliver it online so it could scale, and do the hard part: keep people engaged all the way to completion. Online learning is notorious for drop-off, and a course nobody finishes teaches nobody.
I built it as a self-paced library of focused courses rather than one long linear march, so learners could take what their situation called for and return as needs changed. They completed five of eighteen on average, choosing by relevance, and finished almost everything they started. A 98% completion rate across thousands of learners and several years is the number I am proudest of, because online, completion is what usually breaks.
I have to be honest about what drove those numbers. The programme was funded by the Bank of Industry, and finishing the courses was the route to a shot at that funding. That made it competitive, and it pulled people all the way through. So the 98% is as much a story about incentive design as about the learning itself.
That is the uncomfortable part for a designer. A funding gateway gets you volume and completion fast, but completion is an output, not an outcome. A high finish rate can mean people learned and changed how they run their business, or it can mean they ticked the boxes that stood between them and the money. From the completion figure alone, you cannot tell which.
It also set up a gap I had to design around. Learners arrived for the funding, while the programme existed to build capability. Holding both honestly, meeting the expectation that brought people in without letting the course collapse into a hoop to jump through, is the real work when you build for motivation at scale.
The lesson I carry from it is to design for the outcome, not the proxy, and to measure accordingly. The most valuable thing this programme could still do is longitudinal: go back to those learners and test whether the learning actually changed their behaviour and their businesses, rather than treating completion as the finish line.
Extrinsic motivation can fill a programme. Only evidence of behaviour change tells you it worked.
Much learning reaches people on a phone, inside a restrictive platform. I build self-contained, accessible components that survive those constraints, themed from one set of design tokens so a rebrand is a single change. Two of them are live below. Try them.
Working components, not screenshots. Click a card; drag or use the arrows to reorder.
Online and blended programmes lose people to passive screens: watch a video, read a slide, click next. When they do add interactivity, the reflex is a multiple-choice question with a right answer and a tick or a cross. For factual recall that works. For leadership development it backfires.
Leadership judgement is contextual. There is rarely one correct answer, and a learner who senses they are being marked starts hunting for the response the system wants instead of thinking for themselves. The quiet pressure of getting it wrong is exactly what suppresses the honest reflection that growth depends on.
So for reflective content I reach for a ranking activity with no answer key. The learner orders the items by their own view, then justifies their top choice. Nothing is scored. Every order is defensible, which is what keeps it psychologically safe, and the work shifts from recall to reasoning. What a learner produces, their order and their justification, becomes the material a facilitator opens up with the group: why did you put adaptability above vision? The disagreement in the room is the point, not an error to correct.
There are no right answers. Order these by what matters most to you as a leader, then justify your top choice below.
In a session, the facilitator uses the spread of answers across the cohort to open discussion. The red outline keeps each item trackable as it moves; arrows make it keyboard and touch friendly.
One palette, defined once as CSS variables, themes every component, so consistency holds across a growing library and a rebrand is a one-line change rather than a hunt across files. Each component is keyboard operable, honours reduced-motion, and works from a facilitator's laptop to a learner's phone. They ship by iframe so they stay intact inside whatever platform hosts them. This is design-system thinking and interaction design under real constraints, the same consistency-at-scale problem a maturing product suite has to solve.
