Case study 01Research operations
Building research for a trade audience at CEF
I built CEF's research function from scratch. It showed search was much harder than the business believed, and the changes that followed raised add-to-basket rates and cut repeat searches.
- said finding products was the main problem
- 80%
- or so members in my user panel
- 600
The challenge
CEF's website and app serve a demanding B2B audience with unusual requirements: trade pricing, discounts on large orders, a VAT toggle and multiple sub-accounts per business. CEF had no research function, so the business made decisions on received wisdom.
What I did
- I set up continuous research and evaluation, including a user panel of about 600 members on UserTesting, which I used to recruit for usability tests and surveys.
- I combined surveys, interviews and usability testing with analytics.
- I observed electricians at trade counters to see how they buy.
What I found
- Electricians aren't one audience. They vary by type of work, organisation size, complexity and formality.
- They buy several times a week and need detailed product information fast.
- Users compared the website with their local branch, not with competitors. The site lacked what made branches valuable: product help, alternatives when items are out of stock, and instant quotes or custom pricing.
- In a survey, 80% of customers said the main problem with the site was finding products.
- Search was harder than stakeholders realised. They believed users knew product codes and often searched with them, and that this made searching easier. Surveys showed that few users knew or used product codes. Search was a much bigger issue than had been anticipated.
- One of the most useful questions I asked was "How similar are the jobs you do?" Many users do the same jobs again and again, so they mostly buy the same items.
- That put the search problems in context. Searches for familiar items were easy. Users struggled most with new or unusual items, and couldn't get any help like they would at the branch.
What changed
- Most purchases are repeat buys, so I pushed for easier reordering. The team prioritised reordering improvements and added their requirements to the backlog.
- We investigated other ways to make finding products easier: product cards in search results redesigned to fit smaller screens, so users could see more products quicker; better suggested products; and a refined "similar stock" idea.
- Evidence from surveys, interviews and analytics led the team to show more relevant options and product information in search results. The team also reweighted search towards the harder, less familiar searches. Add-to-basket rates rose and repeat searches fell.
- Research showed a planned job-management feature wasn't needed, so the team didn't build it.
What I learned
CEF's research improved as I gained experience. Over the years I built up specialist knowledge of the products and customers, and that made each subsequent study better. Stakeholders began to request research frequently.
Many features had been built without a full understanding of the use case, like a "quick order" based on product codes, or without the polish and relevance they needed, like the "similar stock" feature, which needed frequent repeated tests. As a result, customers didn't use them. Full discovery research, and usability testing during refinement, were key to making features successful.
What I'd do differently
When I joined, CEF used research only to validate designs, not to learn what users needed before deciding what to build. As my role grew, I realised the work needed a service design approach, so I built that into my digital work. That approach showed how the branches and the online services had complementary roles, but didn't really communicate with each other. Looking back, I'd build more connections across the company. I'd also show what research could do beyond digital, especially for the branches.
Want to talk about this study? Email me or download my CV (PDF).