[Introduction]
Jul 2023
Helping Severn Trent drive daily engagement and leak detection for smart meter customers with My Smart Water Tracker.
Launched a new in-house web tool that replaces a legacy third-party product, improving water usage understanding and making leak alerts actionable - leading to higher daily engagement and empowering customers to save water and money.
[Description]
This project was a exercise in translating complex, low-engagement data into a simple, motivating, and actionable user experience.
My role as the sole designer required a deeply user-centric and iterative approach to balance business goals with diverse user needs.
Framing the Problem & Defining the User
The core challenge was clear: smart meter data was not driving engagement or behaviour change. My initial hypothesis pointed to two key user barriers:
Cognitive Overload: Legacy tools presented data in complex graphs and tables, overwhelming users with low data literacy.
Lack of Tangible Value: Users didn't see a direct connection between their data and actionable benefits, leading to skepticism.
To validate this, I mapped user journeys from the initial meter installation through to regular usage. This revealed critical drop-off points during the first login and when encountering a potential leak alert.
Ideation
Design Thinking & Process
The Mobile-First, Modular Foundation
With analytics confirming 70% of interactions happened on mobile, a mobile-first approach was non-negotiable. This constraint became a strength, forcing a focus on simplicity and core functionality.
My key architectural decision was the Bento-Style, Card-Based Grid. This wasn't an aesthetic choice; it was a structural solution to several problems:
Reduced Cognitive Load: Each card encapsulates a single concept (e.g., "Today's Usage," "Compare Your Use"), preventing information overload.
Inherent Scalability: The modular system allows Severn Trent to add new features (e.g., a future "Water Saving Tips" card) without a complete UI overhaul. It grows with user confidence and business needs.
Visual Hierarchy: The grid naturally guides the user's eye to the most important information first, making the interface scannable and intuitive.
Designing for Clarity and Trust in Data
A major challenge was presenting data to an audience with wildly varying literacy levels.
Graph Iterations: I moved away from complex line graphs with multiple data points. The final daily usage graph uses a simple bar chart, clearly labeled with litres and cost. We iteratively tested labels like "Enough to fill 50 bathtubs" to provide relatable context instead of raw numbers.
The Comparison Tab: We knew household comparison was a potential point of skepticism. To build trust, the design incorporates progressive disclosure. It doesn't thrust users into a competitive landscape. Instead, it introduces the feature gently, explaining how the benchmark is calculated (e.g., "similar-sized households") and framing it as a helpful guide, not a judgment.
The Leak Alert
The leak alert was the highest-stakes interaction. A false alarm or confusing instruction could erode trust entirely. I designed this as a multi-step journey focused on user reassurance.
Tiered Alert System: The design uses a clear visual hierarchy (colour and iconography) to distinguish between a minor "unusual usage" pattern and a critical "continuous flow" alert.
Action-Oriented Flows: The notification doesn't just state a problem; it immediately offers a path to resolution. The copy is calm and directive: "Let's check for a leak."
Empathy for Edge Cases: A crucial decision was integrating user context upfront. By asking during onboarding if a household has high medical water needs, the system can intelligently suppress potential false alerts for those users, preventing unnecessary anxiety.
Onboarding as a Personalised Foundation
The initial onboarding was redesigned from a simple informational tour into a data-gathering tool that personalises the entire experience. The questions we ask ("How many people live here?", "Any medical needs?") are not arbitrary; they are the levers that calibrate the benchmarks and alerts, making the data relevant to each user's unique context from day one.
Conclusion & Iteration
The launch validated our core design decisions, particularly the modular structure and simplified data presentation. The post-launch engagement challenge confirmed that a great tool needs compelling reasons for repeat visits. This directly informed the next phase, where my exploration into gamification and reward mechanisms is a direct response to this behavioural insight, continuing the cycle of user-centred design.
