2024
Point of Sale Platform
BT’s retail presence was expanding, but inside their EE/BT stores, sales advisors struggled with a cumbersome ordering platform. Long load times, confusing flows, and redundant screens slowed them down — meaning fewer customers sold per hour. Our goal: redesign the store sales tool to surface only what’s needed, speed up transactions, and let advisors serve more people per shift.
Problem
In these high-footfall stores, every minute counts. Advisors had to juggle complex plan options, upsells, and hardware bundles. The legacy tablet system was slow, rigid, and built without a deep understanding of how advisors think and sell. The business goal was clear: increase throughput (more sales per shift), reduce cognitive burden on advisors, and align the in-store tool with BT’s product ecosystem.
Discovery
Heuristic evaluation
As the platform was only available on company provided tablets to the Store Advisors, we were not immediately familiar with the journey. So we printed off each step, stuck it onto a whiteboard and began to go through each step to highlight experience blockers, accessibility issues, and design problems.
User research and persona creation
We had little usable data on in-store behavior, so I partnered with our User Research Lead to visit multiple stores and observe both advisors and customers firsthand. We documented patterns, distilled them into a report, and identified a key gap: while we had personas for online users, we had none for in-store advisors — the people actually selling.
Using our findings, I created new advisor and customer personas that captured real behaviors and sales styles. These have since become a go-to reference across the digital team, offering a more grounded, less marketing-driven view of our users.
Create
Prototyping and testing
Working within technical constraints, I used existing design patterns to prototype the new flow in Axure. Our Product Owner — a former Store Manager — helped bring real Advisors and Managers into the usability lab to test it. Their feedback was positive and directly informed several iterations.
Key insights: Advisors are super-users who prefer minimal information compared to the website; performance issues, not just design, slow them down; and the tablet-first approach fits perfectly with how they sell in stores.
Final design
I took these learnings from my research and moved away from the old list format view. The Advisors had responded well in user testing to our horizontal cards. I took the information that we previously had under each list entry and diluted to the essentials that I’d been told were needed to proceed with a sale. This information then provided the base for each card. I liaised with other designers in the business as they were also redesigning our product listings to a card based proposition.
As I had removed some information from the cards, sometimes this still needed to be available for contractual reasons. So there were various ways to display the extra information. For example, our premium offering of ‘Complete Wi-Fi’ was available through a toggle, and extra information like channel listings could be accessed through a pop up.
The order summary page was previously unclear and we’d received feedback that this was another stage where the Advisor would potentially show the customer the page and talk through what they’d agreed. We also introduced a basket page so that orders could be edited before reaching the final summary page.
I worked very closely with our development team to ensure we provided our Store Advisors with the sales tool they needed. I used Zeplin to deliver the final visuals and assets to them as well as keeping regular dialogue open.
Outcome
Post-launch, advisors saw measurable efficiency gains: they could serve 1–2 extra customers per shift, translating to more sales volume. The new card-based UI reduced screen count and cognitive load. The platform aligned better with BT’s digital pattern (also moving toward card-based product listings). Crucially, the tool became more of a selling assistant, not a pain point.
This project reaffirmed that sometimes the best design isn’t adding features — it’s removing friction. In retail contexts, speed and clarity trump “nice to haves.” If I had more runway, I would instrument deeper tracking (time per step, error rates) before and after launch to quantify impact more precisely. Also, deeper A/B testing on toggle interactions might refine hidden data access.