The distinction matters because it decides where the money goes. A business that believes design is decoration buys a nicer looking version of a site that was not working, discovers it still does not work, and concludes that websites do not produce customers. A business that understands design as ordering fixes the thing that was broken, which was usually not the appearance.
This is not an argument that appearance does not matter. It is an argument about sequence.
What design actually does
A visitor arrives at your site carrying questions. Design is the work of answering them in the order they occur, without the person having to search.
That is nearly the whole discipline as it applies to a small business. Everything else, the typefaces and the spacing and the color, is in service of that, and when it stops serving it becomes decoration.
The reason this framing is useful is that it turns an aesthetic argument into a practical one. You cannot easily settle whether a shade of blue is correct. You can settle whether somebody landing on your page can tell within a few seconds what you sell, whether you serve their area, and roughly what it costs. Those are checkable.
The three questions
Almost every visitor to a small business site arrives with the same three, in roughly the same order.
What is this. Not your mission, not your values, not how long you have been trading. What you actually do, stated in the words a customer would use rather than the words your industry uses. A business describing itself as providing integrated solutions has answered nothing, and the visitor is now one question behind.
Is it for me. Location, size, type of customer, type of job. A plumber who works only on commercial properties and a plumber who works only on domestic ones look identical on most websites, and both waste calls because of it. Being explicit about who you are not for is one of the highest value things on a small business site and almost nobody does it.
What happens next. Price or a range, how the process starts, how long things take, and how somebody contacts you. A visitor who has decided they want you and cannot work out what to do next is the most expensive kind of visitor to lose, because they were already convinced.
Sites that answer these three plainly outperform prettier sites that do not. This is consistently true and consistently ignored, because answering them is boring work and choosing a typeface is not.
Order beats appearance
If you take one structural idea from this, take this one.
The order of information on a page does more work than anything visual. A page that puts what you do first, who it is for second, and proof third will function even if it is plain. A page that opens with a large photograph and a slogan, and puts the actual offering below the fold, will underperform however good the photograph is.
The common failure is a homepage that opens with a statement the business is proud of rather than the statement a stranger needs. Those are rarely the same sentence. The business has been living with the offering for a year and finds it obvious. The visitor has been on the page for four seconds.
A simple test: cover everything on your homepage except the first screen. Ask somebody who does not know your business what you sell and whether you could help them. If they cannot answer, no amount of visual work will fix it, because the information is not there.
The mechanics that are not taste
A handful of things in design are not preferences. They are closer to load bearing, and they are the ones a first year business can get right without hiring anybody.
Contrast between text and background. Pale gray text on white is the most common readability failure on small business sites, and it is chosen deliberately because it looks refined on the designer's screen in good light. It is unreadable on a phone outdoors, and a meaningful share of your visitors will be reading it that way.
Size of text. Body text that is comfortable on a large monitor is frequently too small on a phone. The instinct to reduce it for elegance costs you the people whose eyesight is not what it was, which in most local markets is a substantial part of the paying audience.
Space around things. Elements that are close together read as related, and elements that are far apart read as separate. This is not decoration, it is grammar. A heading sitting equidistant between two paragraphs belongs to neither, and the reader has to work out which, which is a small tax paid on every glance.
Size of anything somebody has to tap. A link or button that is comfortable for a mouse pointer is frequently frustrating for a thumb. Anybody who has tried to tap a phone number on a badly built site while standing in the rain understands this immediately.
Consistency. The same kind of thing should look the same way everywhere. Three different button styles on one site do not read as variety, they read as three different people building it, and the visitor cannot tell which things are clickable.
None of these are matters of taste, and all of them are frequently sacrificed to taste.
You are the worst judge of your own site
This is uncomfortable and it is worth sitting with.
You have seen your own site hundreds of times. You know where everything is, what every word means, and what you meant by the heading that confuses everybody else. You cannot un know those things, which means you cannot experience your own site the way a stranger does.
This is why owners fixate on details nobody notices and miss problems everybody hits. The detail you have looked at four hundred times becomes annoying to you long before it becomes visible to anyone else.
The correction is cheap and almost nobody does it. Watch one person use your site without helping them. Not feedback, not opinions, not asking whether they like it. Watch. The places they hesitate are your actual problems, and they will be different from the places you expected, and you will want to explain things to them, and explaining is exactly what your site should have done without you.
Three people is enough to find most of it. The same problems recur, which is how you know they are real.
What is taste and what is not
Taste is which of two competent typefaces you use, whether your palette is warm or cool, whether the photography is bright or moody, and whether the corners are rounded. These decisions matter for how a business feels and they are genuinely worth caring about once the rest is right. They are also the decisions people argue about longest, because everyone has an opinion and none of them can be settled.
Not taste: whether somebody can read the text, find the price, understand what you sell, work out if you serve them, and contact you on a phone with one hand. These can be settled, usually in an afternoon, and they determine whether the taste decisions ever get seen.
The productive rule is to settle everything in the second list before spending an hour on anything in the first.
The redesign trap
A first year business redesigning its site within a year usually should not be.
The impulse arrives for a predictable reason. The site is not producing enquiries, the site is the visible thing, so the site must be the problem. Occasionally that is correct. Frequently the problem is that nobody is visiting, or that the people visiting are the wrong people, or that the enquiries arrive and go unanswered for two days.
A redesign fixes none of those and costs money that would have fixed one of them. Before rebuilding anything, find out how many people are arriving and what they do. If the number is small, you have a visibility problem wearing a design costume. If people arrive and leave immediately, look at the first screen before looking at the whole site.
The honest version of most redesigns is that the owner had grown tired of looking at it. That is a real feeling and it is not a business reason, because the audience has seen it far less often than you have.
What to do instead, in order
Write the three answers plainly, in the words a customer would use, and put them at the top of your homepage where somebody will see them without scrolling.
Fix the mechanical things. Contrast, text size, spacing, tap targets, consistency. This is a short list and none of it requires a designer.
Watch three people use it. Fix what makes them hesitate.
Then, and only then, spend time on how it looks. At that point the taste decisions are worth making properly, because they are decorating something that already works rather than disguising something that does not.
The part that stays true
Fashions in visual design move constantly and the underlying requirement does not. Somebody arrives with questions and limited patience, and the site either answers them in order or loses them.
Every style that has come and gone has been good when it served that and bad when it fought it. That is the only durable test, and it is available to any business owner without hiring anybody or learning anything technical.