Case · 02 / 04
Government · DWP

Universal Credit — from 14 steps to a conversation

Restructuring a high-anxiety benefit application for people at their most vulnerable — through six months of research with low-income, non-digital, and crisis user groups.

RoleSenior UX Designer
SectorUK Government · DWP
Duration6 months
User groupsLow-income · Non-digital · Crisis
The Problem

Fourteen steps for people who can barely take one.

"I tried three times. I kept getting to the same bit and just giving up. It's too much. You're already stressed and then this just… breaks you."

Universal Credit is the primary financial safety net for millions of UK residents. Applying for it often happens at the worst possible moment — job loss, health crisis, relationship breakdown. The service at the time did nothing to acknowledge this reality.

The application journey contained 14 discrete steps, each requiring users to context-switch between their own documents, government guidance, and the form itself. The experience made no distinction between a straightforward case and an extremely complex one — everyone faced the same wall of questions regardless of circumstance.

Research with target user groups revealed the depth of the problem:

  • 01 Emotional barrier. Users in financial crisis described the application as compounding their anxiety — the length and complexity of the form felt punitive rather than supportive.
  • 02 Digital exclusion. A significant proportion of target users had low or no digital literacy — the interface assumed capabilities they didn't have, with no offline or assisted-digital pathway integrated into the design.
  • 03 Loss aversion paralysis. Fear of making a mistake — and the perceived consequences of that mistake — caused users to abandon mid-journey rather than risk getting it wrong.
The Solution

Designing for the person, not the policy.

1
Empathy-led journey mapping

We began by conducting extensive ethnographic research across three user cohorts — people facing redundancy, single parents re-entering the benefits system, and individuals with long-term health conditions. This gave us a rich picture of the emotional context the service needed to operate within, not just the functional tasks it needed to support.

2
Radical simplification of the flow

Working with policy and legal teams, we identified which questions were truly mandatory at the point of application — and which could be deferred. This allowed us to reduce the primary flow from 14 steps to 6, with branching logic that only surfaced additional questions based on the user's specific situation. Complex cases were handled progressively, not all at once.

3
Tone and language as a design tool

Every piece of content was rewritten with emotional intelligence — acknowledging difficulty without being patronising, providing reassurance at high-anxiety moments, and using language that treats users as capable adults going through a hard time. We tested tone variations specifically with crisis user groups across four research rounds.

4
Safety and recovery mechanisms

We introduced a robust save-and-return capability, inline document guidance that reduced the need to leave the service, and clear, non-threatening error recovery that told users exactly what to do next — never just what they'd done wrong. A co-designed "help me with this" pathway connected users to assisted digital support without breaking the primary flow.

The Results

Impact where it matters most.

↑28%
Task completion

Significantly more users completing the application without abandonment — particularly among the most vulnerable user cohorts.

92%
Participant satisfaction

In usability testing, 92% of participants rated the redesigned experience as "easy" or "very easy" — up from 41% on the existing service.

6
Steps (down from 14)

Primary application journey reduced by more than half, with branching logic handling complexity without exposing it to every user.