Replaced a one-size-fits-all online boiler journey with a fast, self-serve path built from original ethnographic research into an emotionally loaded purchase.
Context
Role
Design Lead, sole designer on the product team.
Timeline
2019, roughly 6–8 months end to end, research through shipped redesign.
Team
Sole product designer, working with engineers on user story mapping and with cross-functional stakeholders in a workshop setting, rather than a fixed embedded team.
Key constraint
There was almost no existing understanding of customer motivations, fears, or goals going in, and British Gas's actual value, a personalized in-home expert visit, had to somehow be replicated online, without the person.
Where things stood
The online boiler-buying journey wasn't meeting people where they were. Most customers didn't even know an online quote was possible, and those who found it still felt they needed to talk to a real expert. The online information alone didn't satisfy them. Meanwhile the business wanted more sales, better-quality leads for advisors, and fewer costly call-centre calls, in an increasingly competitive market.
The problem
The gap wasn't information or usability. It was trust and reassurance.
Understood at the start
The problem looked like an awareness or usability issue: customers weren't using the online journey enough, so better promotion or a smoother flow would fix it.
What it turned out to be
Buying a boiler isn't a typical purchase. It's treated as a basic need, wrapped in real anxiety, and customers wanted support and guidance through what felt like an emotional, complicated event. The gap wasn't information or usability. It was trust and reassurance.
Process
01
Starting from ethnographic research instead of assumptions
What was tried
A diary study and in-home interviews with customers who'd just bought a boiler or were mid-search, followed by personas built from the findings.
Why
There was essentially no existing understanding of customer motivations to design from.
What it taught us
Customers shop either proactively (planned) or reactively (emergency), and across both, they wanted guidance through a purchase most people found stressful and unfamiliar.
Considered and rejected
Designing directly off internal assumptions about what customers wanted, not viable given how little was actually known.
[Diagram: proactive and reactive shopping journeys]
02
Choosing "Straight Swaps" over "Virtual Engineer"
What was tried
Two concepts came out of a stakeholder ideation workshop, mapped on an impact-effort matrix: Virtual Engineer (a video call with an expert, replicating the in-home visit) and Straight Swaps (buy a boiler online, skip the in-home visit entirely).
Why
Straight Swaps was prioritized because it directly addressed the segment causing the most call-centre strain (reactive, emergency shoppers), tackling one of the three business goals head-on, while Virtual Engineer would have needed more operational investment (engineer availability for live video calls) to stand up.
What it taught us
User story mapping with engineers defined the minimum information needed to sell a boiler online with no in-home assessment at all.
Considered and rejected
Keeping the in-home assessment as a required step, rejected in favor of speed, specifically for customers in an emergency.
[Diagram: impact-effort matrix from the ideation workshop]
03
Testing inside customers' actual homes
What was tried
Usability testing conducted in customers' homes, since the flow asked people questions about their own boiler and house.
Why
The flow was contextual by nature. Testing it anywhere else risked missing real friction.
What it taught us
Customers often didn't know key details about their existing boiler (model, age, fuel type) off-hand, which shaped how the flow asked for that information, using guided prompts and visual references rather than assuming customers could self-report technical specs.
Considered and rejected
Remote or lab-based testing, which wouldn't have captured home-specific context.
[Screenshot: guided prompts for boiler details]
Research and evidence
[Research and evidence the decisions above rested on]
The trade-off
Chose
A fast, self-serve online purchase path for customers who wanted speed, especially in an emergency.
Gave up
The guidance and reassurance most customers said they actually wanted from this purchase.
Straight Swaps targeted the speed-driven segment, while Virtual Engineer, which wasn't built as part of this project, was intended to serve the guidance-seeking segment instead.
Outcome
Primary metric
Not formally measured or documented.
What shipped
The Straight Swaps online boiler-purchase journey (mobile and web) and a redesigned New Boilers product page, aligned to a newly established design system, with new patterns contributed back into that system.
What happened after
No post-launch iteration is documented for this project.
What I’d do differently
The concept that most directly answered what research said customers wanted (guidance and reassurance) was Virtual Engineer, and it's the one that didn't get built. I'd want to understand now whether Straight Swaps got prioritized because it was easier to ship, rather than because it was what most customers were actually asking for.