No, You Don’t Need a Redesign — You Need a Conversion Framework

Every eighteen months or so, a website starts to feel tired to the people who look at it every day, and someone suggests a redesign. It is a satisfying project. It has a beginning and an end, everyone gets to have opinions, and at the end there is something new to show the board. What there frequently is not, six months later, is more enquiries.

What a redesign actually buys

A redesign is a change of appearance carried out on the whole site at once. That means it is expensive, disruptive, and — crucially — untestable. Because everything moves together, you cannot afterwards say which change produced which effect. If enquiries rise, you will credit the design. If they fall, you will blame the season.

It also, in most cases, leaves the actual machinery untouched. The contact form that fails on older phones is rebuilt in a nicer typeface and still fails. The pricing question everyone asks is still unanswered, in a more elegant font.

The awkward part: looks really do matter

Here the usual version of this argument overreaches, and it is worth conceding properly. “Design is superficial, data is what counts” is a satisfying line and it is not what the research says.

Lindgaard and colleagues at Carleton University, writing in Behaviour & Information Technology in 2006, found that visual appeal ratings formed after 50 milliseconds matched those formed after 500 milliseconds. People judge how a page looks before they have read a word, and the judgement sticks. Appearance is not the icing. It is the first thing that happens.

What decides that judgement is the useful part. Tuch and colleagues, in the International Journal of Human-Computer Studies in 2012, tested 119 real websites and found that low visual complexity and high prototypicality — how closely a page resembles what people expect that kind of page to look like — produced the most positive first impressions, with the effect measurable at 17 milliseconds.

Simple and conventional wins the first half-second, and neither of those is something you need a rebuild to buy.

Read that back and the case against a redesign changes shape. It is not that looks are unimportant. It is that what the evidence rewards is calm and familiar, which is achieved by removing things — and removing things does not require a hundred hours of design work.

The cheap changes with documented returns

Two categories of work have evidence behind them, and neither tends to feature in a redesign proposal.

Speed is the first. In Milliseconds Make Millions, research commissioned by Google and carried out by 55 and Deloitte across 37 brands in 2019, a 0.1 second improvement in mobile site speed moved 3.2% more shoppers from the product listing to the product page and 9.1% more from the product page to the basket. Retail spend rose 9.2% and travel bookings 10%. It is retail and travel data, so treat it as an indication rather than a promise — but a tenth of a second is a remarkably small purchase.

Basic build quality is the second. The 2026 WebAIM Million, an automated scan of the top one million home pages, detected WCAG 2 failures on 95.9% of them, averaging 56.1 errors per page. Because it only counts what software can spot, that is a floor. Unlabelled form fields, unreadable contrast and images with no alternative text are not stylistic preferences; they are the reason a proportion of your visitors cannot complete the thing you wanted them to complete.

Neither of those is a design project. Both can usually be done inside the site you already have, in days rather than months, and both leave you able to say afterwards what changed.

Be careful whose testing advice you take

The standard alternative to a redesign is “run tests” — heatmaps, session recordings, A/B tests, iterate. It is good advice for a retailer with substantial traffic and largely useless advice for everyone else, which is rarely mentioned by the people giving it.

An A/B test works by separating a real difference from random variation, and that needs enough conversions in each variant for the difference to be distinguishable. A site receiving a few hundred visits and a handful of enquiries a month will not reach that point within any useful timeframe. Run the test anyway and you will get a result, feel confident, and have learned nothing — which is worse than not testing, because now you will act on it.

What does work at low volume is qualitative and unfashionable: watching five people actually try to use the site, reading the emails where somebody asked a question the page should have answered, and checking whether your own forms still arrive.

• Submit your own contact form from a phone, on mobile data, and confirm the email arrives
• Ask the last five enquirers what nearly stopped them getting in touch
• Watch one person outside the business try to find your prices
• Check what people search for to reach you, and whether the page they land on answers it
• Run an automated accessibility check on your three most visited pages
• Time how long the homepage takes to become usable on a mid-range phone

That list costs nothing and needs no traffic volume. It also has the advantage that every item on it produces a specific, fixable answer rather than a chart to interpret.

When you genuinely do need a redesign

It would be dishonest to end by saying never. Sometimes a rebuild is the right answer, and the honest test is whether the constraint is in the design or in the platform underneath it.

The distinction is between a website that cannot do what you need and a website you have simply grown bored of. Only one of those is worth five figures.

Where to start on Monday

Take the single page that matters most — usually the main service page — and do three things to it, in this order.

If enquiries move, you will know exactly what moved them, because you only changed one page. That is the part a redesign can never give you, and it is worth more than a new colour scheme.