Monthly sessions
46K
Average service traffic
Improving access to benefits for those in need
Simplifying fit-note submissions, increasing digital uptake by 9.7pp
- Role
- Product Designer
- Date
- 2024
- Client
- GOV.UK
- Status
- Shipped
Jump to a section
- Role
- Product Designer
- Date
- 2024
- Client
- GOV.UK
- Status
- Shipped
Improving access to benefits for those in need
Simplifying fit-note submissions, increasing digital uptake by 9.7pp
Overview
This case study shows my work on a government benefits service, enabling 500,000 - 700,000 citizens each year to submit fit notes and access financial support.
Digital uptake
+10pp
Increase after launch
Customer satisfaction
+5pp
Increase after launch
Scope and ownership
- ROLE
- Lead UX Designer
- DURATION
- 1 year, 2 months
- TEAM
- 10+ cross-functional members
Constraints
- Low completion and satisfaction rates
- Limited developer capacity
- Nuanced policy requirements
Approach
- Map upstream issues to downstream service impacts
- Iterate research and prototypes to reduce upload failures
- Align design decisions across technology, policy, and operations
Problem
Citizens need to submit fit notes as evidence for their benefit claim, but face significant difficulties when doing so. Through research, we found that primary users are likely to submit their fit note several times due to a recurring health condition.
Uploading a fit note was too difficult
- Guidance was confusing and difficult to follow.
- Technical upload issues interrupted an already stressful journey.
- Unsupported file formats prevented valid fit notes from being submitted.
- Frustration and uncertainty caused some users to abandon the process.
Failed uploads created avoidable work
- Support teams received high volumes of calls from users who could not complete the journey.
- Agents manually reprocessed incorrectly uploaded fit notes.
- Rejected evidence delayed benefit processing for people awaiting financial support.
Persona
Andy, who has chronic health conditions, relies on government support to manage his living expenses. He wants to ensure he receives the benefits he needs to maintain his financial stability.
Consider
Begins benefits claim and obtains a fit note from the doctor.
Explore
Reads instructions. Selects either paper or digital fit note. Follows guidance.
Upload
Selects the correct format. Uploads the fit note. Asks for help from friends or family if needed.
Accept
If successful, Andy gets a text notification, and the fit note is sent for processing. Otherwise, he must post it manually.
"What do I need to do here?"
"Seems clear enough"
"Wait, this was meant to be simple!"
"That took far too long"
Process
My primary tools were Figma and the GOV.UK prototyping kit, which uses HTML, CSS and JavaScript. Our team's user-centred design approach was as follows:
DELIVERY SEQUENCE
Activities
User research
Reviewed the feedback inbox and conducted interviews and usability testing with citizens and staff.
Analysis
Assessed backend performance and user flows, identified bottlenecks, and compared similar services.
Prototyping
Designed low- and high-fidelity prototypes, tested them with users, and refined the journey.
Collaboration
Worked with content, research, development, and policy teams to align feasibility and secure sign-off.
Implementation
Partnered with developers during rollout to preserve design intent and respond to delivery constraints.
Feedback Inbox Analysis
- Analysed 4,767 user comments from the feedback inbox exported as Excel rows.
- Grouped feedback by satisfaction score (SS) using Net Promoter Score (NPS) principles.
- Classified SS1 to 2 as service demoters, SS3 as passive, and SS4 to 5 as service promoters.
Satisfaction Score Distribution
| Satisfaction Score (SS) | NPS Category | Rows | Comments |
|---|---|---|---|
| 1 | Demoter | 919 | 544 |
| 2 | Demoter | 877 | 530 |
| 3 | Passive | 644 | 331 |
| 4 | Promoter | 1324 | 765 |
| 5 | Promoter | 3231 | 2597 |
| Total | 6995 | 4767 |
NPS Demoter (SS1 & SS2)
We prioritised addressing the demoter levels, which had 1,074 comments. The table below shows the top 20 categories. Highlighted rows show specific feature requests from users.
| Rank | Category | Occurrences |
|---|---|---|
| 1 | Communication issues | 230 |
| 2 | Difficulty uploading | 178 |
| 3 | Photo acceptance | 139 |
| 4 | General frustration | 106 |
| 5 | Phone lines general | 100 |
| 6 | Fit note not received | 94 |
| 7 | Payments | 89 |
| 8 | File type | 70 |
| 9 | Difficult to use | 65 |
| 10 | Fit note acceptance | 56 |
| Feature request: 11 | Multiple fit note upload | 56 |
| 12 | Other | 42 |
| Feature request: 13 | Digital literacy | 36 |
| Feature request: 14 | Guidance | 30 |
| Feature request: 15 | Photo orientation | 28 |
| 16 | Fit note feedback | 24 |
| Feature request: 17 | Needs extra info | 21 |
| 18 | Other channels | 21 |
| 19 | Upload preview | 20 |
| 20 | Fit note needed signing | 19 |
Top 7: Most Common Feedback Categories (SS1 & SS2)
Visual representation of the most frequent feedback categories from users with low satisfaction scores (1 and 2).
| Category | Occurrences |
|---|---|
| Comms issues | 230 |
| Uploading | 178 |
| Photo acceptance | 139 |
| General frustration | 106 |
| Phone lines | 100 |
| Not received | 94 |
| Payments | 89 |
RESEARCH ROUND 01
Finding where submissions failed
We combined feedback-inbox analysis with interviews involving citizens and agents, then used affinity mapping to identify the strongest recurring patterns.
Feedback inbox
- Communication problems, upload failures, and missing fit notes dominated negative feedback.
- Payment complaints were common among detractors but sat outside our scope, so we escalated them.
- Uploading multiple fit notes in one session was the most requested feature.
User interviews
- Many users had limited digital confidence or accessibility needs.
- Mobile phones were often the only available way to access the service.
- Repeat submissions made unclear guidance and upload failures especially costly.
Behavioural data
- 3
- average upload attempts
- 37%
- drop-off rate
- 63%
- monthly completion rate
Iterations
- I worked with the Content Designer to improve the guidance page copy.
- I designed variants to help users understand how to take clearer, more acceptable photos.
The first iteration reworded the guidance to make it simpler to follow, using plain language and clearer structure so critical information was easier to find. We took it to design critique, where feedback highlighted the need to make the page shorter.
I then explored visual variants that used imagery and step-by-step prompts to guide users towards taking photos that met system requirements, while also making the page shorter. These variants tested much better in user research.
RESEARCH ROUND 02
Testing the clearest guidance
I took the revised designs through critique and heuristic evaluation, then worked with the team across several usability-testing rounds to understand which guidance people could act on.
What users showed us
- Crosshairs around the fit-note image made the intended framing easiest to understand.
- Data matrices had to remain clearly visible so the service could extract fit-note information.
- “Data matrix” and “QR code” confused users, especially those with lower digital confidence.
What we changed
- I worked with the content designer to describe the data matrix as a “black and white square.”
- The phrasing tested well because each fit note contains only one such square.
“My mum wouldn't know what a QR code is.”
Further iterations
- I also designed and prototyped new error messages and 'multiple fit note' upload flows.
- I explored ways to convey contextual error messages and play back the user's uploaded fit note.
- Due to limited developer capacity, these were placed in the backlog.
These iterations streamlined the upload experience, ensuring clearer instructions and more predictable system responses to reduce user confusion and failed submissions.
I also explored designs for handling multiple file uploads, making it easier for users to attach all of them in one go and reducing the need for repeat submissions.
Solution
After several testing rounds and iterations, we finalised the new guidance content and concepts for 'upload' pages to improve users' understanding of what's needed.
Key features include:
- Clearer, more helpful guidance content and error messages
- Better examples of good and bad fit-note images
- Primary and secondary buttons placed closer together on the upload page
- HEIC and HEIF file support, delivered with developers
These refinements addressed user pain points by simplifying the layout, improving the visibility of critical instructions, and ensuring the process could be completed with fewer steps and less scrolling.
Results
My design changes delivered significant results in under six months, demonstrating the effectiveness of my solutions and the value they provided.
Completion rate
+3pp
Increase after launch
Digital uptake
+10pp
Increase after launch
Customer satisfaction
+5pp
Increase after launch
The final solution consolidated key guidance, optimised visual hierarchy, and reduced unnecessary scrolling. This made the process faster to complete, more accessible, and easier for users to follow without confusion.
Conclusion
Through accessible and thoughtful design, we empowered users to navigate the service more confidently, transforming lives and demonstrating the profound impact design can have on public services.
Data-driven design
In live projects, using quantitative data helps you do Reach, Impact, Confidence, and Effort (RICE) scoring. This method can help prioritise design changes effectively, focusing on areas with the greatest benefit to users in the shortest amount of time. Live dashboards and A/B testing can then help validate how our design changes impact the service.
Digital inclusion
Users were told to take a screenshot of their fit note, but one user thought a 'screenshot' meant taking a photo of your screen with a separate camera. This revealed the 'curse of knowledge' bias, where we assume users understand the same things we do. This highlighted the need for simpler language and relatable analogies to support those with lower digital literacy.
Content can't solve it all
Not all issues can be fixed with front-end changes. To reduce failed uploads and improve processing accuracy, we introduced key back-end enhancements. This included a HEIC file converter to support iPhone image uploads, and improvements to the scanning capabilities. These changes led to fewer errors, less manual rework, and a further 6%+ uplift in user satisfaction.