Plenty of things about a website's design are matters of taste. A handful genuinely change how many people get in touch, and they're rarely the things that get argued about in feedback rounds.
Here are the ones that do the work.
One thing should obviously be the most important
Look at any page on your site and ask what the eye lands on first. If the answer is "several things", nothing is being prioritised, and a visitor who has to choose where to look often chooses to leave.
Hierarchy is made with size, weight, colour, spacing and position — and it's subtractive. You don't make something more important by enlarging it; you make it more important by calming everything around it.
The usual failure isn't ugliness. It's a page where six elements all shout at the same volume because six stakeholders each wanted theirs to stand out.
Your primary action needs to be the loudest thing
Pick one action per page. Give it your accent colour and enough space around it that nothing competes.
Two mistakes are common. The first is having three equally-weighted buttons, which makes all of them weaker. The second is the opposite — burying the action in a subtle ghost button because the strong version looked "salesy" in the design review. Visitors can't click what they can't find.
Secondary actions should be visibly secondary: outlined, or a plain text link.
Remove friction before you add persuasion
Every field in a form costs you completions. Every extra step costs you more. Before writing better copy on a button, take something away.
Practically:
- Cut fields you don't need. You can ask on the call.
- Set the right input types so phones show the right keyboard.
- Turn on autocomplete for name, email, address and phone.
- Labels above fields, not inside them, so they don't vanish while typing.
- Validate in place, and say what's wrong in words.
- Don't ask twice. No "confirm your email".
A five-field form with clear labels will beat a ten-field form with beautiful microcopy every time.
The interface has to respond
A button that's been pressed should look pressed. A form that's submitting should say so. Anything that takes time needs an indicator.
Absence of feedback is why people click "Send" three times and you get three copies of the same enquiry.
The same applies to errors. "Something went wrong" helps nobody. "That email address is missing an @" is a design decision that saves a completion.
Make things look like what they are
Buttons should look like buttons. Links should look like links — usually underlined, or at minimum a different colour with a hover state.
The trend for flat, borderless, minimal interfaces has quietly cost a lot of businesses clicks, because people can't tell what's interactive. If something is clickable, it should be obvious before the pointer arrives, which matters even more on touch where there is no hover.
Contrast, and not just for accessibility
Body text needs a contrast ratio of at least 4.5:1, large text 3:1, and interface elements and focus indicators 3:1.
Light grey on white is the most common failure we see. It looks refined on a designer's calibrated monitor and is genuinely hard to read on a phone outdoors, which is where a lot of your traffic is.
This is an accessibility requirement, but treat it as a conversion one. Text nobody can read is text nobody acts on.
Give things room
Space is what makes a page feel considered and makes text readable. It's also the first thing sacrificed in the last round of feedback, when someone asks to fit more above the fold.
Resist. Density doesn't increase attention, it dilutes it. The single most reliable way to make a page feel more expensive is to give it more room and less content.
Line length matters too. Somewhere between 45 and 75 characters is comfortable. Text stretched across a full desktop screen is tiring to read, and tired readers leave.
Consistency is a conversion feature
If a button looks one way on the homepage and another on the services page, visitors spend a fraction of a second working out whether it's the same thing. Multiply that across a site and it reads as unprofessional without anyone being able to say why.
Build components once and reuse them. It's faster to build, easier to maintain, and it's most of what "polished" means.
What matters less than people think
Above the fold. People scroll. A cramped hero stuffed with everything performs worse than a clear one that invites scrolling.
Button colour, in isolation. The endless red-versus-green tests mostly measured contrast. A button that stands out from its surroundings wins; which colour it is matters far less.
Fashionable effects. Custom cursors, elaborate scroll animations and 3D are neutral at best and costly at worst, because they slow the page down.
The number of clicks. Three obvious steps beat one confusing one.
And the thing that beats all of it
None of this rescues a page that doesn't say what the business does.
The most common reason a site underperforms isn't its interface, it's that a visitor can't tell within a few seconds whether they're in the right place. Fix the message first. Then fix the interface.
Questions we get asked
How do I know which of these is my problem?
Watch someone use your site. Give a person who's never seen it a realistic task — "find out whether this company could help you and get in touch" — then stay silent and watch where they hesitate. Three sessions will tell you more than any amount of theory.
Should we A/B test this?
Only if you have the traffic for it. Below a few thousand visitors a month you'll never reach significance, and you'd be better off fixing the obvious problems and moving on. Most small business sites have plenty of those.
Our site looks good but doesn't convert. Where do I start?
The headline, then the primary action, then the form. In our experience that's the order the problems usually sit in, and the first one is right far less often than people assume.
If your site looks good but isn't producing enquiries, send us the link. We'll tell you what we'd change first.