Good for whom?
Eight hundred and eighty people a month ask Google what makes a good website, and the results are mostly lists of design principles: white space, typography, consistency. Those matter, but they are judged by designers. A business website is judged by two other parties. The customer, who wants to know quickly whether you do the thing, where, and for how much. And Google, which wants to know whether the page answers the search and whether the site can be trusted.
Everything below is a thing one of those two parties measures. Where we have numbers from live sites we built, they are in the table, with the tool and the date.
The nine things
- Fast. First content visible in under a second on a phone, fully loaded in under two. Google measures this as Core Web Vitals and reports it in Search Console; visitors measure it by leaving.
- Stable. Nothing jumps as the page loads. A button that moves under the thumb is a lost enquiry. Google measures it as Cumulative Layout Shift; the good score is close to zero.
- Clear in one screen. Before any scrolling, a stranger can say what you do, where, and for whom. Test it: show the homepage to someone for five seconds and ask them.
- Priced, or at least ranged. On the live Manchester web design result, one studio in ten shows a price. The searches for “website design cost” run at 880 a month because nobody answers. A visible price is the strongest trust signal a small business can publish.
- Obvious next step. One primary action per page — call, quote, buy, book — and it is the most prominent thing on the screen, with the phone number in reach for people who would rather ring.
- Found for what you sell. Search Console shows the queries a site appears for. A site found only for its own name is a business card, not a website. Each service needs its own page, titled in the customer’s words.
- Editable by you. Prices, opening hours and offers change weekly. If every change is a ticket to a developer, they stop happening and the site rots.
- Secure and maintained. HTTPS, updates applied, backups taken, uptime monitored. Invisible until it is not.
- Honest about who is behind it. A named person, a real address or area, real reviews you can click through to, and no rating or claim that cannot be checked. Since April 2025 that last point is consumer law as well as good practice.
What the measurable ones look like on live sites
| Site | First content painted | Full load | Layout shift | Requests to render |
|---|---|---|---|---|
| propavingmanchester.com (paving, Manchester) | 0.56s | 0.68s | 0 | 11 |
| beds.ie (17,000-product store) | 0.42s | — | 0 | — |
| cfcweddinglists.com (gift-list platform) | 0.78s | — | 0 | 30 |
| vehicles2lease.co.uk (van leasing) | 0.89s | — | 0 | — |
| rochsweets.com (1,000-line sweet shop) | 0.86s | 1.27s | 0 | 71 |
Source: Browser Navigation Timing, Resource Timing and Layout Instability APIs, read on the live homepages on 4 September 2026. Dashes are figures not recorded in that reading rather than bad results.
Those are not exceptional numbers; they are what a site does when images are sized properly, the code is lean and nothing loads that is not needed. The comparison is the average small business WordPress site with a page builder and twenty plugins, which commonly takes four to eight seconds on a phone. The visitor does not know why. They just leave.
Test your own site in ten minutes
- Open PageSpeed Insights, enter your address, read the mobile tab. Note the Largest Contentful Paint and Cumulative Layout Shift figures.
- Load the homepage on your phone on mobile data, not wi-fi. Count the seconds.
- Show the homepage to someone who does not know your business for five seconds. Ask what you do and where.
- Try to find a price. Time it.
- Try to enquire from a service page without going back to the homepage.
- Search Google for your trade plus your town. Are you there? Now search your business name. Are you there? The gap between the two answers is your findability problem.
- Open Search Console → Queries. What proportion of your clicks are your own name?
- Change your opening hours. Can you do it yourself in two minutes?
- Check the padlock, and check the footer for who is behind the site.
Most sites fail three of those. The three that fail are almost always speed, price and the next step, and none of the three needs a redesign to fix.
What does not make a website good
- Animation for its own sake. Motion that helps you understand what a page is doing is good design. Motion that makes you wait is a cost.
- A video in the hero. It is the single heaviest thing you can put on a page, and it plays under the text people are trying to read.
- Stock photography of handshakes. It signals that you had nothing real to show.
- A chatbot bubble over the checkout. Helpfulness that intercepts a purchase is a conversion problem.
- Awards, badges and star ratings that do not link to where they were earned.
- More pages. “How many pages should a website have” gets asked, and the answer is one per thing a customer needs to know. Ten pages that each say one thing beat forty that say nothing.
So is design worth paying for?
Yes, but not for the reason the design industry gives. The value of a designer is not that the site looks nicer. It is that the nine things above are decisions, and they conflict: the beautiful full-bleed image is the slow load; the clever navigation is the hidden price; the immersive scroll is the unstable layout. Design is the discipline of making those trade-offs in the customer’s favour. That is worth paying for, and it is also why our site prices are deliberately low: the build is where the decisions get made, and the growth work afterwards is where they get proven.
