The 5 Most Common Website Mistakes We See from Design Agencies (And How to Avoid Them)

We are often the second agency on a project. Someone has already paid for a website, it looks perfectly respectable, and it is not doing what they hoped. So we open it up and look.

The same handful of problems keep turning up. What is striking is that they are almost never failures of taste or craft — the designers involved were good at their jobs. They are gaps in what the project was asked to deliver. Nobody was made responsible for them, so nobody did them, and nothing in the approval process was ever going to catch it.

Here are the five we see most, and the question that would have caught each one.

1. Nobody watched anyone use it

The site was reviewed by the agency, then by the client, then by the client’s colleagues. All of them already knew what the business does, where the important pages are, and what the navigation labels mean. None of them were the audience.

This is the cheapest gap to close and the one most often missing. Jakob Nielsen of Nielsen Norman Group argued as far back as March 2000 that a first study with five participants finds 85% of the usability problems in a design, and that you need at least 15 users to find all of them. Five people. Not a lab, not a panel — five people from outside the project, asked to do one real task while somebody watches quietly. If nobody outside the project has tried to use the site, the sign-off was a review of a picture, not a test of a product.

Ask for it by name in the brief: how many people outside our two teams will attempt a real task on this site before launch, and what happens to what they find? If nobody will do it, do it yourself.

2. The home page never says what the business does

This one is almost always caused by the client, and a good agency should push back on it. The home page opens with a mood statement — “Transforming ambition into impact” — because that felt more confident in the workshop than “We install commercial kitchens across the South East”.

Visitors do not have the patience for it. Writing in September 2011 on an analysis of more than 205,000 pages, Nielsen Norman Group concluded that users often leave a page within 10-20 seconds, and that to hold anyone’s attention for longer you have to communicate your value proposition within 10 seconds. The mood statement spends that budget saying nothing.

The fix costs nothing and takes an afternoon: say plainly what you do, who for, and where, in the first thing anyone reads. Save the ambition for the second screen, once you have earned it.

3. Speed was treated as something to sort out later

Performance is the classic phase-two item. The design is approved on aesthetics, the build follows the design, and by the time anyone measures the result, the heavy hero video and the six web fonts are load-bearing parts of a layout the client has already fallen in love with.

It is far cheaper to set the target first. Google publishes the numbers: its Core Web Vitals guidance asks for a Largest Contentful Paint within 2.5 seconds, an Interaction to Next Paint of 200 milliseconds or less, and a Cumulative Layout Shift of 0.1 or less, all measured at the 75th percentile of real page loads. Those are not aspirations; they are a pass mark you can write into a contract.

And speed is worth real money. In Milliseconds Make Millions, research commissioned by Google and carried out by 55 and Deloitte across 37 brands, a 0.1 second improvement in mobile site speed moved 3.2% more retail shoppers from the product listing to the product page and 9.1% more from the product page to the basket, while retail spend rose 9.2% and travel bookings rose 10%. A tenth of a second. Ask what your performance budget is before you approve a design, not after.

4. The whole thing was approved on a desktop

Everyone in the meeting is looking at a large screen. The mobile version gets a quick scroll near the end, and the small compromises made to fit the design onto a phone — the stacked navigation, the hero text that shrinks to nothing, the table that scrolls sideways — never get the same scrutiny.

The counter-mistake is now just as common. Plenty of agencies design mobile-first and treat desktop as an afterthought, which is equally lopsided: StatCounter put mobile at 49.36% of global web traffic in August 2026, with desktop at 49.11% and tablets at 1.54%. The split is close to even. Both versions are the real website. Review them on the devices they will actually be used on, in the meeting, on the day.

5. Accessibility was never in the brief

Not deliberately excluded — simply never mentioned. Nobody asked what standard the site should meet, so nobody built to one, and the colour contrast, form labels, alt text and keyboard navigation ended up wherever the theme happened to leave them.

This is not a rare oversight. The WebAIM Million survey for 2026 found detected WCAG 2 failures on 95.9% of the top one million home pages, averaging 56.1 errors per page, up from 94.8% the year before. Those are only the failures automated software can catch, so the true picture is worse. A site with a specified standard in its brief is already unusual.

Name the standard — WCAG 2.2 AA is the sensible default — and ask how it will be checked. It costs very little when it is built in and a great deal when it is retrofitted.

Why these survive a professional review

It would be easy to read that list as a charge of laziness. It is not. There is a structural reason these particular faults get through, and it is worth understanding if you want to avoid them.

They are all invisible in the artefact that gets approved. A design is signed off as a picture, and none of these five problems show up in a picture. Worse, the research suggests a conventional-looking page will be judged rather favourably regardless: Tuch and colleagues, writing in the International Journal of Human-Computer Studies in 2012, found across 119 sites that low visual complexity and high prototypicality produced the most positive first impressions, measurable at 17 milliseconds. A site can be slow, unclear, inaccessible and untested, and still make a good first impression. Nothing in the approval meeting will tell you otherwise.

Change the questions and you change the brief. Change the brief and most of this list never happens in the first place — which is a considerably better outcome than hiring a second agency to find it afterwards.