The UX design process is the repeatable path from a fuzzy problem to a product people can actually use. It isn’t a burst of inspiration or a folder of screens — it’s a sequence of decisions, each one de-risking the next. Get the process right and the interface almost designs itself; skip steps and you ship guesses.
At Biotik we run some version of this on every engagement, whether it’s a two-week sprint or a multi-month build. Here is the process, phase by phase, and where each one tends to go wrong.
1. Discovery & research
Every good process starts before the pixels. Discovery is where you understand the customer, the business goal and the constraints — so you design the right thing, not just a thing done well. That means talking to real users, reviewing analytics, mapping the current journey and agreeing the one metric that would move the business.
- User interviews and, where a product exists, session and funnel data
- Competitive and heuristic review of what’s out there today
- A clear problem statement and the metric you’re designing to move
2. Define & synthesise
Research produces a pile of observations; definition turns them into a direction. This is where patterns become insights, insights become a prioritised problem, and the team aligns on what “good” looks like. Outputs are lightweight but decisive: a sharp problem statement, a few personas or jobs-to-be-done, and a map of the flows that matter most.
3. Ideate & design
Now the visible work begins — but it starts with structure, not colour. Information architecture and user flows come first, then wireframes, then the interface itself. Designing the skeleton before the skin keeps debate on the things that decide success (can people find and finish the task?) rather than the things that feel urgent (which shade of blue?).
- Information architecture and end-to-end flows
- Low-fidelity wireframes to settle structure fast
- High-fidelity UI, on-brand and built from reusable components
4. Prototype
A prototype turns a static design into something people can drive. Even a clickable, mid-fidelity flow exposes problems that no amount of reviewing screens will — a confusing step, a dead end, a moment of doubt. Prototyping is cheap insurance: it’s far less costly to fix a flow in Figma than after engineering has built it.
5. Test & validate
Testing is how the process stays honest. Put five to eight people through the prototype, watch where they hesitate, and you’ll learn more than a month of internal opinions. Combine that qualitative signal with heuristic evaluation and, on live products, real analytics. If you already have a product, a structured UX audit is the fastest way to find what to fix before you redesign anything.
6. Iterate
The process is a loop, not a line. Findings feed back into design, the prototype is revised, and you test again — tightening the experience each pass. The discipline is knowing when a flow is good enough to ship and where the next round of learning should focus, rather than polishing forever.
7. Deliver & hand over
The final phase is what separates a design that ships from one that stalls. You hand engineering production-ready screens, documented states and edge cases, and ideally a design system of components and tokens they can build against. Do this well and the team keeps quality high long after the project ends; do it badly and every future feature re-litigates the basics.
Where the UX design process goes wrong
- Skipping research — designing on assumptions, then discovering them wrong in production
- Jumping to UI — debating visuals before the structure and flow are settled
- Testing too late — validating after build, when changes are expensive
- No handover system — shipping screens instead of a system, so quality drifts
- Treating it as linear — a real process loops back as it learns
None of this needs to be heavy. On a small engagement the whole loop can run in a couple of weeks; on a larger build it repeats per feature. What matters is that each phase happens in order — understand, define, design, prototype, test, iterate, deliver — because each one earns the right to the next.
