UX · UI
User Experience / User Interface
What is UX?
- UX = the whole experience of using it
- UI = visible screens and controls
- Good-looking ≠ easy to use
Once you take UX/UI seriously, the meeting agenda itself changes. The question shifts from 'what color should this button be?' to 'why did users come to this screen, and what must they finish within three minutes?' It is not about drawing screens; it is about designing user paths. That is why half of UX work comes not from design tools but from observing users and analyzing logs to find where they stall. The deliverable is also not a polished mockup but a change you can state in numbers, such as 'cut the clicks to complete an order from 12 to 5.'
In practice, it helps to tell them apart by deliverables. UI leaves behind things you can see, such as wireframes, components, and design systems; UX leaves behind process and records, such as user interviews, journey maps, and usability tests. You also need to draw the line with neighboring concepts. UX covers the touchpoints while someone is using the product, whereas CX (customer experience) is the entire journey from awareness and purchase to support and cancellation. Validation differs too. The long-standing rule of thumb in the industry is that just 5 participants will surface most of the major usability problems, so running several rounds of 5 people beats gathering 100 at once, and it gives answers much faster in the field.
The most common mismatch is treating 'UX improvement' as 'screen redesign.' If you change only fonts and colors and leave the number of steps untouched, the first week after launch brings only 'Where did the old screen go?' questions, and the drop-off rate stays flat. The second is sequencing. If design is layered on after development is 90% done, the structure can no longer be touched and you can only paper over it, and that cost is carried straight into the next redesign. The third is the moment a decision-maker's taste overrides user data. Once a single remark like 'The boss doesn't like that color' overrules the test results, UX research turns into a rubber stamp. Your evidence should be not who said it, but how many people stopped, and where.
Seungtae ParkHands-on AX instructor who teaches AI business development and marketing automation by building real systems
John KimInstructor John Kim, CEO of Nanum Consulting and author of “B2B Marketing Design”The term UX was popularized by Donald Norman in the 1990s and became a core concept of human-centered design.
Consider a hypothetical case. Company A, a household goods manufacturer with 180 employees and annual revenue of about KRW 90 (≈ US$1) billion, launched a dealer-only ordering website three years ago. Two years on, 62% of all orders still came in by phone and fax, and four people on the sales administration team spent half of each day writing orders down. Management set aside a redesign budget, saying, 'Let's update the screens to a modern style.'
- Before kicking off the redesign, Team Lead Kim of the sales administration team visited 12 dealerships in person and had the owners open the ordering screen in front of him. Seven of them asked, on the very first screen, 'Where is the cart?'
- Pulling six months of access logs showed an average of 14 clicks to complete an order, and 58% of drop-offs happened at a single step: 'choose shipping address.' Each dealer had two or three warehouse addresses, yet had to enter them from scratch every time.
- The PM at the design agency suggested, 'Let's start by sorting out colors and fonts.' But Team Lead Kim persuaded them to postpone the visual design for three weeks and redraw the ordering flow first. That is where the saved shipping address and the 'reorder last month' button came from.
- Before development began, at the paper-prototype stage, they brought in five dealer owners for 30-minute tests. Three of them made the same remark, 'The reorder button was below the fold, so I missed it,' so they moved it to the top.
- Two months after launch, clicks to complete an order fell from 14 to 6, the web share of orders rose from 38% to 71%, and the sales administration team’s order-intake workload dropped from 320 hours a month to about 110, freeing two people to move on to finding new dealerships.