What usability measures
Usability is judged against a task, not against a page. It asks whether a person who arrives with a purpose — get a price, book a slot, find out whether you serve their area — can finish it without help, without a wrong turn, and without giving up. A site can be attractive and still be unusable, and a plain site with clear labels can be excellent.
Several things sit inside it. Whether a first-time visitor can work out what to do at all. How much effort a return visitor spends on the same task. How often people go wrong, and how easily they recover when they do. Whether the labels mean the same thing to the reader as they do to you. Those are separate qualities, and a site can be strong in one and weak in another, which is why a single verdict such as “the site is fine” is rarely useful.
Why usability matters
Every marketing pound or rupee you spend passes through it. Traffic that arrives on a page where the price is hidden, the form asks for things people will not give, or the phone number is an image, converts worse regardless of how good the targeting was. Usability is the multiplier applied to everything upstream of it.
It matters more the harder the conditions are. On a mid-range phone over a patchy mobile connection, which describes a large share of traffic in Nepal, small annoyances become blockers: a menu that needs a precise tap, a field that fights the keyboard, a page that shifts as it loads. The same design that is merely irritating on a laptop loses the sale on a phone. It also overlaps heavily with accessibility — most of what helps a screen-reader user helps everyone.
Where usability goes wrong
The usual cause is expertise. You know what your services are called, so the navigation uses your internal names rather than the words customers search for. You know the process, so a step that looks obvious to you is unexplained. Nobody on the inside can see this, which is why opinion is a poor substitute for watching someone.
Then there is decoration mistaken for design: animations that delay reading, low-contrast grey text, buttons styled to look like labels, a carousel nobody scrolls. And finally, treating usability as a launch task. Sites drift as pages are added, and a navigation that made sense with a handful of pages becomes unusable at a hundred.
How to act on it
Name the two or three tasks that matter commercially and test only those. Watch a few real people try them, without helping, and note where they stop. That evidence beats any checklist, and it is the core of user testing.
Fix the cheap things first: plain labels, visible pricing or a clear reason there is none, big tap targets, readable contrast, forms that ask only for what you will use, and error messages that say what to do rather than what went wrong. Then measure, so improvement is visible rather than argued about — completion of each task, how long it takes, and where people drop out. Re-check whenever the navigation or a key template changes, not once a year.