The illustration rule, stated first. Every pattern below is derived from the criteria in our entry bar, not from an assessment log; this edition has none. None of them describes a real business, and no business is recoverable from any of them. There are no screenshots, no quotations, no paraphrases of anyone's actual wording, and no combination of sector, place, size or story that could point at anyone. Where a pattern needed an illustration, we invented one.
What this is, and who it is for
People ask to be shown what bad looks like, and they are right to ask. A checklist tells you what to have. A pattern tells you what it feels like to arrive at a site that does not have it, which is the part that makes it click.
So here are nine. They are the failures the entry bar is built to catch, they cut across every kind of business, and each of them has a fix that takes under an hour and costs nothing.
One thing to say clearly at the start. None of these is evidence that a business is badly run. Most of them are what happens when a competent person is busy, or when a site was built once for a good reason that has since expired. The likeliest cause of every pattern here is that the operator has not looked at their own site as a stranger for a long time, which is an entirely normal thing to stop doing.
How these patterns were built, and what was left out
When sites are assessed, the reasons they do not clear our entry bar are recorded as patterns, because the patterns are more useful than the individual cases and because we have no intention of ever publishing the individual cases. No such assessment log exists yet, so the nine below are derived from the criteria themselves.
Building a pattern from a criterion means describing the shape a failure of it takes, with nothing specific to any instance, and checking that what remains could describe many businesses rather than one. Left out: the sector, the location, the size, the number of people, the actual wording, the design, the founding story, and anything a person who knows the business would recognise.
We do not publish teardowns of named businesses and we never will. Naming a business in order to criticise its website is a reputational judgement dressed as education, and it is exactly the behaviour the rest of this site exists to argue against. The criteria behind these patterns are at the refusal patterns these produce and the refusal file, which is anonymised on the same principle.
1. The site that describes a philosophy, not a product
The homepage is about values, approach and a way of working. It is sincerely written and often well written. Several paragraphs in, a visitor still cannot say what they would be buying.
This happens to people who care about their work. After enough years, the thing that feels worth saying is the thinking behind it, and the thing itself becomes too obvious to mention.
What it costs. Every visitor who arrives without already knowing what the business does. They cannot tell whether they are in the right place, so nothing further on the page can be read correctly.
2. The site where the newest thing is three years old
A news section whose latest item is old. An events page listing things that have happened. A copyright year that stopped. Nothing on the site is wrong, exactly. It simply all points backwards.
What it costs. The visitor concludes the business has closed. They do not ring to check, because ringing a closed business feels foolish. This one is invisible to the operator, who knows they are open and therefore never reads the page as a stranger would.
3. The site with five ways to contact and no reply from any
An email address, a form, a phone number, two social handles. The email forwards somewhere nobody opens, the form has been broken since a change nobody noticed, and the number rings out.
Usually there is one channel that does work. The visitor has no way to tell which.
What it costs. The most expensive loss on the list, because it hits people who had already decided to proceed. They try, get nothing, try again, and leave believing the business ignores enquiries.
4. The site that is one long image with no text
The whole page is a picture. Prices, hours, address and the description of the business are all inside it as pixels.
What it costs. Anyone who cannot see the image gets an empty page: a screen reader user, someone on a slow connection, someone whose browser failed to load it. Nobody can copy the address into a map or the number into a phone. Nothing on the page can be searched for, by a visitor or by anything else. And when a price changes, the operator needs whoever made the image, so the price does not change.
5. The site where the shop is a link to somewhere else entirely
"Shop" leads to a different site, with different branding, possibly a different name. No warning, no explanation. The visitor cannot tell whether this is the same business, a reseller, or something wrong.
There is nearly always a sound reason. Selling elsewhere is a legitimate arrangement. The site just does not mention it.
What it costs. The most cautious visitors, which is to say the ones being careful about who they send money to, stop at the moment the name changes.
6. The site that requires an account to see anything
Prices, catalogue, sometimes even the description of what is sold, all behind a sign-up. The visitor is asked to commit before being told what they would be committing to.
Common where a business sells to trade customers and has decided that everyone else is not the audience. The site does not say so, so everyone else simply bounces off a wall.
What it costs. Everyone unwilling to create an account to answer a question they consider basic, which is most people. It also removes any way for a stranger to establish that the business is real.
7. The site whose three pages disagree with each other
The homepage gives one set of hours, the contact page another, the footer a third phone number. A service listed on one page is absent from another. Nothing is deliberately wrong. The pages were written at different times and never reconciled.
What it costs. Doubly. The visitor cannot rely on any of it, and the disagreement itself reads as disorganisation, which transfers to everything else. It is worse than a single wrong fact, because a single wrong fact at least looks like an accident.
8. The site that is really a holding page with a blog attached
A short homepage, a contact form, and a blog with a few posts. The blog carries most of the words. There is no page saying what is sold, at what price, or on what terms.
The blog was intended as a way to look active. It became the site.
What it costs. A visitor who wants to buy has nowhere to go. They may read a good article and leave impressed and unserved, which is the least useful outcome available.
9. The site that is beautiful and says nothing
This one exists to break an assumption, and it is the pattern operators most resist.
The design is genuinely good. Considered typography, real photography, careful spacing, everything loads properly and works on a phone. A visitor arrives, moves through it, and leaves without learning what is sold, what it costs, whether the business is taking work, or how to reach anyone. There is a contact form and nothing else.
Sites like this are frequently the product of real money and real skill. The skill went into how it looks, and nobody was responsible for whether it answered anything.
What it costs. Everything the other patterns cost, with an extra difficulty attached: the operator has evidence that the site is good, so when enquiries do not arrive, the problem is assumed to lie elsewhere. This is the pattern most likely to be answered with a redesign, which is the response least likely to help.
The smallest fix for each
Every fix here is under an hour, needs no budget, no developer and no redesign.
| Pattern | The smallest fix |
|---|---|
| 1. Philosophy, not product | Write one sentence saying what a customer gets for their money, in the words they would use, and put it at the top. Leave the philosophy underneath. |
| 2. Newest thing is three years old | Fix the copyright year. Delete or archive the stale news section. Add one line saying when the page was last reviewed, and mean it. |
| 3. Five ways to contact, no reply | Delete every channel you do not watch. Keep one. Add a line saying when you reply. Send a test message through it and confirm it arrives. |
| 4. One long image | Retype the price, hours, address, phone number and description as text beneath the image. Ten minutes, and the image can stay. |
| 5. Shop is somewhere else | Add one sentence before the link: "Our shop is hosted on another site, which is why the name changes." Warn people the destination differs. |
| 6. Account required | Put the description, the price bands and the contact route outside the login. If it is trade-only, say so on the front page and say who it is for. |
| 7. Pages disagree | Write your facts into one document, then correct every page from it. Half an hour, once, and it stays fixed if the document stays authoritative. |
| 8. Holding page with a blog | Add one page saying what is sold, roughly what it costs, and how to buy or book. It can be short. It cannot be absent. |
| 9. Beautiful, says nothing | Same as pattern 1, plus a price signal and a line on availability. Change no design element at all. |
If more than three of these apply, work down the seven-item floor these violate rather than treating them as nine separate jobs, because they overlap. If you cannot tell which apply to you, the checklist to run against your own is written to be run by the person who built the site, which is the hardest audience for it.
In Plain English
Nine ways small business sites lose people, none of them describing a real business, all of them derived from the entry bar's criteria.
The pattern to take seriously is the last one. A site can be genuinely well-designed, expensive and modern, and still fail to say what it sells, what it costs, whether anyone is available, or how to reach a person. Design quality and answering questions are separate properties, and only one of them is what a stranger came for.
None of these failures means someone is bad at their job. Most of them are what a website turns into when a busy person stops reading it. The fix is nearly always retyping a sentence, deleting something dead, or reconciling two pages that drifted apart.
If you recognise your own site here, that is the point rather than an accusation. Every fix in the table takes less than an hour, and none of them requires a redesign.
Run the findability checklist at free templates, or read the rest of this section at shopfront basics.