Human Centred Web Design: Turning User Research Into Better Pages

Human centered web design means building pages around what real visitors are trying to do, using evidence from those visitors rather than internal opinion. A small UAE team can do it without a research department: five short interviews, a handful of task tests and an honest look at analytics will show where pages fail, and each finding can be turned into a specific change.

What human centered web design looks like in practice

Most website projects start from the inside out: a sitemap based on the organisation chart and a homepage that says what leadership wants to say. None of it is tested against what a visitor needs on arrival.

A human centred approach flips the starting point. Before deciding what a page says, the team finds out three things:

• Who arrives on the page and from where (search, an ad, a referral, a WhatsApp link shared by a colleague).
• What they are trying to get done in that visit, in their words rather than the company's.
• Where they get stuck, hesitate or leave.

The output is not a persona poster but a short list of problems, each tied to a page and a fix.

Why light research is enough to start

Usability specialists at Nielsen Norman Group have long argued that testing with around five people uncovers most of the significant problems in a design, because each extra participant reveals less that is new. Their advice is to run several small rounds, fixing issues between each, rather than one large study. For a business website, that means a research round can fit inside a normal two week sprint.

Step 1: Run five focused user research interviews

User research for a website does not need a lab. It needs five conversations of around 30 minutes with people who match the visitors you want: recent customers, prospects who enquired but did not buy, or people who fit the profile but have never heard of you.

Who to talk to

• Two recent customers who chose you, so you hear what nearly stopped them.
• Two prospects who enquired and went elsewhere, if they are willing.
• One person who matches your target buyer but has no relationship with you yet.

In the UAE it is worth checking the language mix early. If a meaningful share of your buyers prefer Arabic, include at least one Arabic speaking participant, because navigation labels and form wording that feel obvious in English can read very differently in translation.

Questions that produce usable answers

1. Think about the last time you looked for a service like ours. What started the search?
2. What did you need to find out before you were willing to get in touch?
3. What made you trust or distrust the websites you looked at?
4. What nearly made you give up?

Ask about past behaviour, not hypothetical preferences: "How did you contact the last supplier you chose?" produces facts, while "Would you use live chat?" produces polite guesses.

Step 2: Watch people complete real tasks

Interviews tell you what people think. Usability testing shows what they actually do, and the two often disagree. A task test is simple: give a participant a realistic goal, ask them to think aloud while they use the site, and watch without helping.

Writing tasks that reveal problems

Good tasks describe an outcome, not a path. Compare these two:

Weak tasks test whether people follow instructions; better ones test whether the site matches how they think.

Test on the devices people use

Run at least half the sessions on a phone, ideally the participant's own. Many Gulf visitors arrive from social links and messaging apps on mobile, and problems such as headers covering form fields or tap targets placed too close together only appear on a small screen. Remote sessions with screen sharing work well.

What to record

For each task, note whether the person completed it, where they hesitated and what they said at that moment. Quotes settle internal debates faster than opinion.

Step 3: Review analytics for the patterns behind the stories

Interviews and task tests give depth; analytics gives scale. Check whether what you saw in sessions happens across all visitors.

Useful views to check:

• Landing pages by entry source. Which pages do people arrive on from search, ads and social? Those pages carry the most weight in first impressions.
• Paths to the contact or checkout page. Which routes do converting visitors take, and where do others drop away?
• Site search terms. If you have internal search, the terms people type show the words they use and the content they cannot find in the menu.
• Form abandonment. If your form tool records it, see which field makes people stop.

Session recording and heatmap tools can add detail here, but they collect behavioural data. Under the UAE Personal Data Protection Law (Federal Decree-Law No. 45 of 2021), and under the separate DIFC and ADGM regimes if they apply to you, it is sensible to confirm with counsel how such tools should be configured, disclosed and consented to before installing them.

Step 4: Map findings to user flows

Once you have interview notes, task results and analytics, put them onto a simple map of user flows (the routes visitors take from arrival to action). Keep it to the two or three flows that matter most commercially, for example "first time visitor from search to enquiry" or "returning customer looking for support".

For each flow, list the stages (arrive, understand, compare, trust, act) and place every finding where it happens. Patterns appear quickly, such as problems clustering around trust.

Step 5: Turn each finding into a specific page change

This is the step most research skips, and it is where the value is. Every finding should end as a change someone can build and measure. Use a simple format: what we saw, what we will change, how we will know.

Hypothetical example: a Dubai based facilities management company runs the process above. Here is how its findings might translate.

Prioritise by impact and effort

Sort the list into quick content edits, template changes that need design and development, and structural changes such as navigation. Ship the quick group first while bigger changes are scoped.

UX research methods compared

There are many UX research methods, and a small team should pick the ones that answer its specific question. The table below covers the methods most useful for an existing business website.

Common mistakes that waste research effort

• Testing with colleagues. Staff know the jargon and the menu. They will always find the page.
• Collecting and not acting. If findings do not reach a backlog with owners, the next round will find the same problems.
• Ignoring accessibility. People using screen readers, larger text or keyboard navigation are real visitors. The Web Content Accessibility Guidelines (WCAG) 2.2 give a practical checklist to run alongside your own tests.
• Doing it once. Repeat a short round after each significant release to confirm the fixes worked.

If you are planning a wider redesign, research also sharpens the brief. It shows which parts of the work are design problems and which are build problems, a split explored in paying for design versus build in Dubai. It also gives you evidence for what professional web design actually changes, and helps you choose the right web design tool stack for prototyping and testing. Teams running projects outside Dubai may also find how an Abu Dhabi project differs useful when recruiting participants.

Get research led page changes built properly

Codeeo's UI/UX design team runs this kind of light research and turns the results into wireframes, prototypes and finished pages, and our web development team builds and measures the changes. If you would rather start with a clear list of problems on your current site, a website audit is a practical first step. We have delivered 200+ projects across the UAE and the US, and because our office is in Dubai you can meet the team in person. Talk to us about your website.