Designing from evidence to impact

Find the right problem

A project may begin with a quarterly KPI, customer feedback or a wider business objective, such as improving conversion, increasing engagement or reducing support demand.

 

I combine quantitative and qualitative evidence to understand what is happening and why:

• Product analytics and funnel data
• Customer feedback and user research
• Usability testing and session recordings
• Search, support and call-centre data
• Business goals and technical constraints

 

I then conduct a heuristic review and map the findings against the user journey. This connects pain points, behavioural evidence and opportunities in one shared view.

 

Outcome: A clearly defined customer problem linked to a measurable business objective.

Turn insight into a prediction

Before designing a solution, I make the team’s assumptions explicit.

We believe that [a defined change]
for [a particular audience]
will result in [a predicted behaviour].
We will know it works when [a chosen measure changes].

This prediction is shared with product, engineering and other stakeholders. Together, we assess its value, feasibility and alignment with the wider product strategy.

We agree what success should look like, how it will be measured and what evidence might cause us to change direction.

 

Outcome: A testable hypothesis, an aligned product team and clear success measures.

Create and test the solution

I explore different ways to address the problem through user flows, wireframes, content, interface concepts and interactive prototypes.

 

My design decisions consider:

• Clear information hierarchy
• Familiar interaction patterns
• Responsive behaviour
• Accessibility and inclusive design
• Brand and tone of voice
• Design-system consistency
• Technical feasibility
• Loading, error, empty and success states

 

Accessibility is considered from the outset. This includes colour contrast, typography, focus order, keyboard use, semantic structure, touch targets and compatibility with assistive technologies.

I prototype primarily in Figma. When greater realism is needed, I use Figma MCP, Claude Code and VS Code to create coded proof-of-concepts and test responsive behaviour, richer interactions or technical assumptions. The fidelity depends on what the team needs to learn, not how polished the prototype needs to look.

 

Outcome: An accessible, brand-appropriate solution that can be tested, evaluated and delivered.

Measure what changed

Launching the design allows us to compare real behaviour with our original prediction.

 

Depending on the objective, I may monitor:

• Conversion or adoption
• Task completion
• Errors and abandonment
• Feature engagement
• Retention and repeat use
• Customer satisfaction
• Support demand
• Qualitative feedback

 

I loosely use Google’s HEART framework: Happiness, Engagement, Adoption, Retention and Task Success, to help choose meaningful experience measures. The final metrics vary according to the product, audience and business objective. If the solution creates the expected change, the team has a clear story showing how product and design delivered value. If it does not, we use the evidence to refine the experience, test another approach or pivot.

 

Outcome: Evidence of impact and a more informed next decision.