The challenge
Reece's strategic goal is to shift a significant proportion of sales online., but tradespeople use the language of the job: colloquial names, site slang, terms passed down through apprenticeships. A plumber looking for a 'veggie sprayer' or a 'fowler button' is searching the name they've always used - not the name in a product catalogue. When maX couldn't recognise those terms, customers hit zero results or irrelevant products and had no choice but to ask branch.
Compounding this, Reece has high branch staff turnover. Product knowledge - the ability to translate what a customer asks for - was walking out the door. The product couldn't keep relying on branch expertise to bridge that gap.
Findability was consistently the lowest-scoring category in the bi-annual maX Experience Survey, at 3.78/5. Customers named search as an active impediment to buying online, and 13% of searches were returning zero results.
Lingo was the response: a query reformulation approach that bridges the gap between how trade customers talk and how Reece's catalogue is structured - reducing dependency on branch for product discovery, and encoding product knowledge that had previously lived only in people's heads.
"When I type in the name of something, it comes up with 0 results. In branch they know what it's called, so I just go to them."
- Trade customer
Customer problem
As a trade customer, I want to find materials in maX using the language I'm used to, but the catalogue doesn't recognise the way I search, so I give up and call the branch instead.
Business problem
Search failures were driving customers back to branch, adding pressure on staff and blocking the shift to online sales.
My process
1. Understanding the search failure pattern
I worked with the tech lead, who had conceived the technical approach - a query reformulation model that analysed how customers searched after a failed first attempt (Query Reformulation Analysis). My role was to define the customer-facing experience: when and how to surface suggested terms, and how to make the recommendation impossible to miss.
2. In-branch usability testing
The PM and I ran usability sessions with maX customers in a Reece plumbing branch - intercepting trade customers waiting for their orders to be picked. We tested three prototype iterations. The first two revealed the same problem: customers were missing the suggestion UI entirely. The third solved it with hard-to-miss call-to-action buttons that prompted customers to choose the keyword most associated with their search. That framing landed immediately - customers understood the task and acted on it.
3. Solution flow
The Lingo logic works across two states:
If a search returns results but the terms are deemed problematic, the product listing page displays alongside suggested alternative terms
If a search returns zero results and the terms are deemed problematic, a zero results page surfaces suggested terms instead
Non-problematic searches route to the standard product listing page unchanged - the intervention only fires where it can genuinely help.
Results and impact
Zero results rate dropped from 13% to 8.6% - a meaningful reduction in dead-end searches across maX
Search-to-cart conversion on reformulated terms reached 12.5% - customers who accepted a suggestion went on to add to cart or view a product
Online orders increased by 1.25% over 6 months following rollout
Findability score in the February 2026 maX Experience Survey increased from 3.78 to 3.96 - the first recorded improvement in Reece's lowest-scoring customer experience category
125 of the 1,000 classified problematic terms resulted in customers accepting a suggestion and converting downstream
The solution required no new platform infrastructure, built entirely on existing maX search architecture
Branch staff contributed directly to synonym accuracy, creating a feedback loop between trade knowledge and the product - reducing reliance on individual expertise that turns over with staff
Why this mattered
Search is one of the highest-frequency interactions in maX - performed hundreds of thousands of times daily. Every failed search is a moment where a customer loses confidence in the platform and reaches for the phone instead. At that scale, even incremental improvements compound quickly.
But the bigger opportunity is what Lingo represents for Reece long-term. As more problematic terms are converted, the synonym model becomes more accurate, conversion continues to grow, and the Findability metric continues to improve - without additional branch effort. The product gets smarter over time, encoding trade knowledge that would otherwise leave the business.
That's what makes this more than a search fix. A maX that speaks the trade's language becomes genuinely hard to replace - a platform customers trust to help them work faster, and a competitive advantage that deepens with every search.