Working out what people are trying to do on your site, then removing whatever is stopping them. The visual design comes after that, not instead of it.
They get sold as one thing and they answer different questions. Knowing which one your site actually has a problem with saves a lot of money.
What people are trying to do, the steps they go through, and what gets in the way. Decided before anything is drawn.
The visual layer applied to that structure. Type, colour, spacing, and the state of every button and field.
Most sites that convert badly have a UX problem being treated as a UI problem. A redesign makes it prettier, the structure stays the same, and nothing changes in the numbers. That's the expensive version of this mistake and it happens constantly.
A typical path from arriving to getting in touch. Four steps, and each one has a standard way of going wrong.
The step
Seconds, not minutes. They're checking whether this site is about the thing they searched for.
Where it fails
The first screen is a slogan and a stock image. Nothing confirms they're in the right place, so they go back.
The step
Scanning navigation and headings for the service, product, or answer they came for.
Where it fails
Menu items named the way the company thinks internally rather than what customers call things.
The step
Looking for price, proof, and reassurance that you can actually do this for someone like them.
Where it fails
No pricing anywhere and nothing concrete, so the only way to find out is a call they don't want to book yet.
The step
Filling a form, calling, or booking. The moment everything else was leading to.
Where it fails
Eleven fields, three of them required for no reason, and nothing saying what happens after they press send.
The useful part is that these are measurable rather than a matter of taste. Your analytics shows where people leave, and session recordings show what they were doing when they did. We work from that rather than from an opinion about what looks confusing.
Design happens in stages, and the early ones are deliberately plain. It's the fastest way to change your mind cheaply.
Stage 01
Boxes and labels. No colour, no fonts, no images. Deliberately ugly so the conversation stays on what goes where and in what order.
Stage 02
Real spacing, real proportions, real text lengths. Still no brand colour. This is where the structure gets pressure tested against actual content.
Stage 03
Type, colour, imagery, and every state of every element. Only reached once the structure underneath is agreed and not going to move.
Skipping to stage three is tempting and it's why revisions get expensive. Once something is fully designed, changing where a section sits means redoing the visual work too. Disagreeing at the wireframe stage costs an hour, and the same disagreement at stage three costs days.
You can't judge your own site's usability, and neither can we. Both of us already know where everything is.
Most usability problems show up in the first handful of sessions. Testing fifty people finds the same issues and costs ten times as much.
"Find out what this would cost you" rather than "what do you think of the design". Watching someone struggle in silence is the whole method.
Session recordings show real visitors, not people who agreed to be watched. Useful for finding where the problem is before deciding what to test.
This isn't always necessary. On a small site with clear problems, the issues are obvious from the analytics and testing would just confirm what we already know. We'd only suggest it where the answer is genuinely unclear or where a decision is expensive enough to be worth checking.
Files you own and can hand to any developer, not screenshots in a slide deck.
Page by page, showing structure and hierarchy, with notes on what each section is for.
Every page at desktop and phone width, including the states nobody remembers: errors, empty, loading.
Buttons, forms, cards, and type rules defined once. It's what stops a site drifting as pages get added later.
Linked screens you can click through before anything is built, which is what makes testing possible at all.
Depends on the complaint. If people say the site looks dated, that's visual. If they say they can't find things or the enquiries aren't coming, that's structural.
Making a site with a structural problem look better is the most common way money gets wasted on design. We'd rather establish which it is before quoting anything.
Yes, and it's common. You get the wireframes, designs, and prototype, and your own developer builds from them.
Worth having whoever will build it involved early. Designs made without any input from the people implementing them tend to include things that are unnecessarily hard to build.
Because you'd react to the colours instead of the structure, and the structure is the part that determines whether the site works.
It's also cheaper for you. Wireframes are quick to change and finished designs aren't, so every disagreement caught early saves days of rework.
Two rounds per stage, which in practice is enough because each stage is signed off before the next one starts.
What causes revision problems is reopening a decision made three stages ago. If that happens we'll say what it costs rather than absorbing it quietly and rushing something else.
Some conventions exist because people expect them, and breaking those makes a site harder to use rather than more distinctive. Logo top left, menu top right, that sort of thing.
Distinctiveness comes from typography, colour, spacing, and what you actually say. Inventing a new navigation pattern is a costly way to be different.
The basics, as standard. Readable contrast, keyboard navigation, labelled form fields, and text that scales when someone increases their font size.
Formal compliance to a specific standard is a bigger piece of work with an audit attached. If you need that, say so at the start rather than after launch.
Included in every site we build, and available on its own when someone else is doing the build. Pricing sits on the main web design page.
Other design services
Thirty minutes on whether it's structure or visuals, and which one is worth paying to fix.
Or email hello@culmen.digital