The argument
Every product is user-tested eventually — the choice is before launch or after
We do not have an office in Saudi Arabia
We serve organisations there from our Riyadh office and our engineering hubs in Lahore and Karachi, travelling for the parts of an engagement that need to happen in person.
Skipping research does not avoid the test; it moves it. A product designed from internal assumption meets its first real users in production, where every discovered failure costs an engineering sprint to fix, a support queue to absorb, and a first impression that does not come back. The teams that think they saved a research budget spent it — several times over — on rework, and paid in reputation for the privilege of learning in public.
The correction is small and unglamorous: before the expensive commitments, put the design in front of a handful of people who genuinely resemble the audience and watch them attempt the tasks that matter. Not a focus group offering opinions — observed behaviour. Users are unreliable narrators of what they would do and startlingly reliable demonstrators of what they cannot do. Five sessions do not answer everything; run on the right questions, they answer the ones the launch depends on.
In this market the language of the session is part of the method. A product serving Saudi Arabia that recruits only English-speaking participants has tested a minority of its audience and generalised to the rest. We run sessions in Arabic with Arabic-speaking users as a default, not an add-on, because the hesitations, the vocabulary and the trust signals differ — and those differences are usually the findings that matter most.
The guesses
Four assumptions that quietly design products in this region
None of these is announced as a decision. Each arrives as common sense, survives because nobody watches a user contradict it, and hardens into the product.
- 01
The user in the boardroom
The team designs for someone like themselves — fluent in the domain, patient with the product, invested in its logic. The actual user has thirty seconds, a different vocabulary, and no loyalty. The gap between the two is where products fail.
- 02
The Arabic user as a translated English user
Assuming the Arabic-speaking audience behaves like the English-speaking one, minus the language, ships a mirrored interface and misses different expectations about names, forms, formality and what a trustworthy page looks like.
- 03
The pattern that worked somewhere else
A flow lifted from a famous global product carries an unexamined assumption: that your audience shares its context, its users’ habits and its onboarding investment. Patterns are hypotheses to test here, not credentials.
- 04
Polish mistaken for progress
Jumping to high-fidelity screens makes the design feel decided and expensive to question. Structure — what the product does, in what order, with what words — should be tested while it is still cheap to be wrong about, in artefacts rough enough to change.
The method
From open questions to a tested design in structured weeks, not open-ended ones
Research earns its place by being scoped to decisions. Every step here exists to settle something the team would otherwise argue about or guess at.
Name the decisions at stake
We list what the team is about to commit to — a flow, a structure, a naming scheme — and design the research to inform exactly those. Research without a decision attached produces a deck, not a change. Research questions each tied to a pending commitment.
Recruit the real audience
Participants are recruited to match the people the product serves — including Arabic-speaking users as a matter of course — not whoever is easiest to schedule from the office’s own network. Sessions whose findings generalise to the launch audience.
Observe tasks, not opinions
Each session sets real tasks on the current product or a prototype and watches where behaviour and intention part ways — the retyped field, the wrong menu opened twice, the page abandoned in silence. Observed failures, each with its moment and cause on record.
Turn findings into design, then re-test
Findings are synthesised into prioritised design changes, the design is revised, and the revision goes back in front of users — because a fix is a hypothesis until someone completes the task it was meant to unblock. A design that has already survived its audience once.
Our role
What the research programme delivers and what your product owners decide with it
- 01
Sessions run properly, in both languages
Recruiting, moderation and note discipline are craft, and doing them badly produces confident wrong answers. We run the sessions — in Arabic and English — so the findings can be trusted enough to act on.
- 02
Findings shaped into decisions
The deliverable is not a report of everything observed; it is a short, ranked set of design decisions with the evidence behind each, sized so a product team can act on it the same week.
- 03
Design carried through to tested screens
Where the engagement includes design, we take the findings through information architecture, flows and interface — in both directions, right-to-left included — and validate the result with users before it reaches the backlog.
What the product should be remains your call; research changes the ground you make it on, replacing the strongest opinion in the room with the recorded behaviour of the people you are building for.
Common questions
What buyers ask
We talk to our customers all the time. Is formal research going to tell us anything new?+
Almost certainly, because conversation and observation measure different things. Customers in conversation are polite, retrospective and inaccurate about their own behaviour; the same person in a usability session, given a task and silence, will demonstrate problems they would never have thought to mention. Sales and support conversations also over-sample your most engaged customers, while a product mostly meets people with no relationship to you at all. Teams rich in customer contact are usually rich in opinion data and poor in behavioural data — and it is the behavioural kind that settles design arguments.
Do usability sessions for a Saudi audience have to be conducted in Arabic?+
For any product whose real audience includes Arabic speakers — which in Saudi Arabia is nearly every product — yes, a meaningful share of sessions should run in Arabic, moderated by someone fluent in it. A participant working in their second language behaves differently: slower, more forgiving, less likely to voice confusion, and interacting with an interface that may not be the one they would choose. Sessions in English with bilingual participants systematically understate the problems the Arabic experience has. The practical pattern is a mixed programme, with the Arabic share matching the audience rather than the office.
Why do researchers keep saying five users are enough for a usability study?+
Because of what a usability study is for. It does not estimate how many people have a problem — that takes volume and quantitative method. It discovers that a problem exists, and the frequent, serious problems reveal themselves within a handful of sessions, with each further participant mostly re-demonstrating what the first few found. The efficient pattern is small rounds and iteration: five sessions, fix what they exposed, five more on the revision. Ten users across two versions teach far more than twenty watching the same flaws repeat. Where you genuinely need to count or segment, we say so and reach for quantitative tools instead.
In this cluster
Start the conversation
A scoped research plan for one product decision you are about to make — who to observe, what tasks to set, and what the findings will settle.
Start the conversation
