A website gets judged as taste. “I like it” or “I do not like it,” and the discussion ends there because nobody has language for what design is actually doing. But most of what makes a design good is not preference; it is a chain of reasons, and reasons can be examined.
Here is the standard I hold work to: every element on the page should be able to answer for itself.
The four questions an element must survive
- What is it for?
- Who needs it, and when do they reach it?
- What would break if it were removed?
- Why this rather than something smaller, quieter, or earlier?
A headline answers “what is this.” A button answers “what next.” A photograph answers “should they believe it.” Section spacing answers “where does one idea stop and the next begin.”
When an element cannot answer, it is usually one of two things: a habit copied from a template, or the designer solving for their own portfolio instead of the visitor’s task. Both are removable, and the page gets better.
Structure carries meaning before type does
The grid is not a container for content; it is an argument about what belongs together. Two items placed side by side will be read as related whether or not they are. A list of five things with equal weight tells the visitor all five matter equally, which is rarely true and rarely what the business means.
This is why a layout can look tidy and still be unclear: the spacing is consistent, the alignment is clean, and the structure is describing the wrong relationships. Fixing that is a structural conversation, not a styling one.
Type is mostly about sequence
Typography discussions get stuck on personality — this serif versus that one — while the real work is hierarchy: what gets read first, second, and third, and whether the sizes and weights make that ordering unmistakable.
A page where the headline, the subhead, and the body sit too close in weight asks the visitor to rank the information themselves. That is extra cognitive work on every screen, and it is one of the quiet reasons sites feel “heavy” without anything obviously wrong with them.
One deliberate rule, applied consistently, outperforms a distinctive typeface used inconsistently.
Colour is information, so spend it carefully
If orange appears on four different things, it stops meaning any of them. Reserve the strong colour for the few moments that need it — the primary action, the current state, the one thing to notice — and let the rest of the palette be calm.
Contrast is not a style preference either. Body text that fails against its background is a functionality failure, and it is measurable rather than debatable. Small-business sites frequently fail here for an ordinary reason: the design was reviewed on a good monitor in a bright room, by a person with the designer’s eyesight.
Motion should explain, not perform
Animation earns its place when it shows a relationship or a state change: this panel came from that button, these two things belong together, this section follows that one. Everything else is a delay between the visitor and their answer.
The practical discipline: keep movement short, keep it tied to interaction, and provide a reduced-motion path for people who find movement unpleasant or dizzying. On this site, animation is switched off automatically when the browser reports that preference — a two-line decision that removes a real accessibility problem.
Imagery has to be honest
The most common design dishonesty on small-business sites is a photograph that describes nobody’s actual work. Generic stock imagery says “this business is like other businesses,” which is precisely the message a site exists to override.
Real photographs of real work, real premises, and real people beat better-produced fakes. Where a proof image does not exist yet, the honest move is to say so — an illustrative planning image labelled as one, a client-permitted screenshot, or nothing at all — rather than filling the space with atmosphere.
A picture needs the same justification as any other element: who needs to see this, and what does it convince them of?
Taste still matters, in its place
All of the above is reasoning, and reasoning alone produces correct, forgettable pages. Judgement enters where the reasons run out: which of two adequate structures feels right for this business, how much personality the brand can carry, where to break the pattern so the page does not read as a template.
That is real design work, and it is also the part nobody can audit for you. But it should be the last decision made, not the first — because taste applied to the wrong structure produces a beautiful page that still fails to explain the business.
The test I apply to my own work
Read the page top to bottom and, for each section, say what it is for out loud. Any section where you hesitate goes on the list of things to justify or remove.
Then ask one stranger the five-second question, and check whether the design helped or got in the way of the answer.
That is the whole method. It keeps decisions honest, keeps scope small, and keeps a redesign from becoming a costume change on the same unclear message.
Curious whether your current page holds up to this? Make the message clearer, or send the address of the site through the contact form and I will tell you what I would check first.