UX process & deliverables

This happens before any screen is drawn. Draw screens without it and you end up asking "why is this menu here?" and redoing everything. Clients often push to skip IA on agency work — the time saved here comes back threefold later.

The page count in the sitemap is frequently the basis of the quote. Confirm it before signing, and state that pages added later are a separate negotiation.

Korean industry usage. "Only two levels" means make everything reachable within two clicks. The deeper it goes, the more sharply people fail to arrive.

Drawing only the happy path is doing half the job. Failure, cancellation, and leaving and coming back all need branches, or development will be stuck asking.

It is deliberately unattractive. Make it pretty and the client starts talking about colour and fonts, and the structure never gets discussed. Say up front: "this is not the design, it is for checking the structure".

The stage between wireframe and prototype. The "design" shown to a client is usually a mockup. Seeing one often prompts "so it is done then?", so explain in advance.

Telling wireframe / mockup / prototype apart matters enormously in agency contracts. "Includes the prototype" and "includes the design" are completely different amounts of work. State in the contract where it stops.

It is the reference document that makes "not in this document" mean "not in scope". Developers build from it and QA verifies against it. Write it vaguely and a dispute is guaranteed.

"Add descriptions" means the picture alone is not enough — write the behaviour down. A good description is not "arranged attractively" but a verifiable sentence like "button is disabled when input is under 2 characters".

Always clarify what "three designs" means. Three genuinely different concepts, or three colour variants of one concept? The work differs threefold. And state the number of revision rounds (usually two or three) in the contract or you fall into endless revisions.

"Women aged 20 to 40" is not a persona; it is demographics. It has to be specific enough to make design decisions from: "a 32-year-old office worker who wants to finish an order in under five minutes on the morning commute".

It covers beyond the product itself, not just screens — seeing the ad, waiting for delivery, calling support. It is the tool for finding problems that fixing screens cannot solve.

"It is awkward" has to be an observation, not an impression. Not "sign-up is awkward" but "62% drop out at step 2 of 3" — evidence is what makes it persuasive.

Testing just five people surfaces roughly 85% of the problems (Nielsen). You do not need a crowd. What matters is watching without helping. The moment you say "press that one there", the test is worthless.

Nielsen's ten usability heuristics are the de facto checklist. Faster and cheaper than a usability test, but it cannot catch the unexpected things real users do. Best used alongside one.

It only means anything with a large enough sample. On a site with 100 visitors a day, an A/B test is usually meaningless. And change only one thing at a time or you cannot tell what caused what.

The point is finding the stage that loses the most people. Fixing one drop-off point affects revenue far more than making screens prettier.

This is the language clients use to talk about results. Being able to explain a design proposal in these terms changes everything about how it lands. "It will reduce cart abandonment" is far stronger than "it will look nicer".

The key word in agency scope negotiation. When requirements keep growing, "let us push that past the MVP, ship first and see the response" holds the line. An MVP is built small, not built carelessly.

In practice people mostly use it interchangeably with signifier. "Give it affordance" really means "make it look pressable", which is strictly a signifier. Just knowing that is enough.

Coined by Donald Norman to separate it from affordance. When "what is possible" (affordance) and "what looks possible" (signifier) diverge, people get lost. A button that works but looks inert is the example.

Go against those expectations and it feels awkward however well it is built. The reason the cart icon is top-right, the reason the logo goes home — all mental models. This is not the place to be creative.

Many options, difficult wording and inconsistent rules all raise it. "Don't Make Me Think" is the classic principle in this area.

The evidence for cutting menus. That said, cutting for the sake of it is not the answer — the point is classifying well so that fewer things are visible at once.

The scientific basis for button size and placement. It is why important buttons sit at the bottom of a phone screen where the thumb reaches, and why a "delete" button that must not be hit by accident is placed away from the others.

The answer to "why not do something new?" People spend far more time on other sites than on yours. Following familiar patterns is usually right. Differentiate through content and visuals, not structure.

The theoretical basis for "why this much space?" Things close together read as one group (proximity). The gap between a label and its field has to be smaller than the gap to the next item for them to read as a pair.

The reason to put important things top-left and in the first two lines. People scan rather than read. Short subheadings and lists land far better than long sentences.

An area that can lift results substantially without touching the design. Changing "OK" to "Start for free" alone shifts conversion. Error messages must carry ① what went wrong ② what to do about it.

Onboarding is not explaining a lot. Getting someone to actually do something on the first screen beats five tutorial slides. And skip must always be available.

Terrible when overused. Make someone dismiss five in a row on arrival and nobody reads any of them. Point out one or two genuinely hard-to-find features and provide a way to see them again.

Regulated in Korea under the amended e-commerce act. Hidden auto-renewals, structures that make cancelling difficult, and greying out the decline button all qualify. Even when a client asks for it, the right move is to explain the legal risk and propose an alternative.