Website redesign: 7 mistakes that lower conversions

The new site is live. The design is beautiful, the animations are smooth, the photography looks great. The whole team is thrilled. You launch, you post about it on LinkedIn — and then you look at Google Analytics.

Conversions are down. Enquiries are down. The phone rings less. And you think: how is that possible? We invested a fortune in a new site, and it’s objectively better.

But “better” for you and “better” for your customers are two different things. A redesign is one of the riskiest operations in online marketing. Not because a new site can’t be better — but because most redesigns are driven by how the site looks, not by how it works.

Here are the seven most common reasons new sites earn less. Then how to do a redesign that actually delivers.

1. Losing learned patterns — people already knew how to buy on the old site

Your existing customers knew their way around the old site. They knew where the pricing was, how to fill in the enquiry form, where to click for contact. They built those patterns over months or years — one visit, one order at a time.

A redesign breaks all of them. Even when the new navigation is “more logical”. The problem isn’t logic — the problem is that the visitor now has to think. And the moment somebody thinks instead of acting, conversions fall.

What to watch

Navigation structure. Move key items into a different category or rename them and people lose them. “Services” is suddenly “Solutions” and the customer doesn’t know whether it’s the same thing.
The conversion path. If it took two clicks to reach an enquiry on the old site and four on the new one, you have a problem.
Position of key elements. Was the phone number top right? Leave it there. People look there automatically.

This doesn’t mean changing nothing. It means knowing what worked on the old site — and deliberately keeping it.

2. Designing for beauty instead of function

Probably the most common cause of a post-redesign conversion drop. The company hires a capable design agency or freelancer, the result looks superb — but nobody asked why visitors come to the site and what they want to do there.

The designer gets the brief “make us a modern site”. And does exactly that. Big photos, lots of white space, minimalist menu, hidden text. The result wins design awards and hurts conversions.

What it looks like in practice

Large hero sections with no content. A full-screen photo with a single word. It looks cool, but the visitor has no idea what the company does.
Navigation hidden behind a hamburger menu on desktop too. On trend? Yes. Functional for commerce and services? Definitely not. A hidden menu means fewer clicks to key pages.
Copy sacrificed to layout. The copywriter writes a selling text, the designer halves it because “it doesn’t look right in the layout otherwise”.

The result: a site that projects professionalism but never leads the visitor to act.

How to do it right

Design must serve the content and the conversion goals, not the other way round. That means first defining what the page has to communicate and where it should take the visitor, and only then the visual treatment. The guide to website copywriting covers how to write copy that converts. On landing pages, where the conversion goal is unambiguous, this goes double — how to structure a landing page that captures leads is a separate piece.

3. Ignoring the mobile experience

Most redesigns are still designed primarily on desktop. The designer works on a large monitor, the client approves on a laptop, and the mobile version “adapts somehow”.

Meanwhile mobile traffic makes up more than half of total traffic on most sites. In some sectors — restaurants, local services, e-commerce — it’s over 70 %. And still the mobile version gets treated as an afterthought.

Typical post-redesign mobile problems

Buttons too small or too close together. Elegant on desktop, untappable on a phone.
Forms not adapted to mobile. Fields that are painful to type into. Missing input types (numeric keyboard for phone, email keyboard for email).
Content that disappears on mobile. Some redesigns hide key sections on mobile — testimonials, pricing, FAQ — so the page “doesn’t look crowded”. Those were exactly the sections that could have decided the conversion.
Slow loading. New designs often bring larger images, more animation and heavier scripts, which take several times longer on a mobile connection.

What to do

Design mobile-first. Start with how the site works on a phone, then add for larger screens. And above all: walk the conversion path on mobile for real — go from landing on the site to submitting an enquiry, on your own phone, ideally on a slower connection.

4. Removing the elements that were working

A redesign tends to “tidy up” and “simplify”. Which sounds sensible — but often removes or hides something that demonstrably brought conversions.

What typically gets removed (and shouldn’t)

Testimonials. The old site had a strip of client logos and three quotes under the hero. The new site moved them to a separate “References” page almost nobody clicks through to.
Trust signals. Certifications, guarantees, number of completed projects. Visible on the old site, gone “in the redesign”.
The FAQ section. “We can write that better” — and then it never gets written at all.
The phone number in the header. The designer said it disrupted the visual cleanliness. Customers were calling precisely because they could see it.

All of these reduce a visitor’s uncertainty and help them decide. Removing them doesn’t simplify their path — it makes it harder, because the information they need to decide is gone.

How to handle it

Before removing anything, look at the data. Analytics, session recordings — whatever you have. If people clicked testimonials on the old site and then converted, testimonials stay somewhere visible. Decide on data, not on the designer’s aesthetic instinct.

5. No tracking — and no way to compare

This mistake doesn’t show up immediately, which is exactly why it hurts more later. The company launches a new site and then discovers it has no data to compare against.

What it looks like

Analytics was never set up properly on the old site (or the move to GA4 lost the history).
Analytics on the new site gets configured a week after launch.
Nobody tracked the old site’s conversion rate, so it’s unclear whether the new one converts better or worse.
Forms have no submission tracking, so nobody knows how many people complete them and how many abandon.

The result: subjective impressions instead of hard data. “It feels like it’s working better” isn’t a metric.

What to set up BEFORE launch

Document the old site’s key metrics. Conversion rate, bounce rate, time on page, form submissions, most-visited pages.
Set up tracking on the new site before it goes live. Analytics, form tracking, conversion events. Verify all of it works.
Define what you’ll measure. Not “total traffic”, but specific conversion actions: enquiry submitted, phone click, add to cart.
Run both versions in parallel at least briefly (see point 6 on gradual rollout).

Without data you evaluate a redesign on feel. And feelings will mislead you.

6. A big-bang launch instead of a gradual rollout

Most redesigns launch as “we switch off the old site on Friday and switch on the new one on Monday”. A big event, big excitement — and big risk.

If something goes wrong (and something nearly always does), there’s nowhere to retreat to. The new site is out, the old one is gone. If conversions dropped because of a broken form, missing content or slow loading, you find out weeks later — losing customers throughout.

What to do instead

A/B test the key pages. Launch the new design on one or two pages first (a landing page for one service, say) and compare conversions against the old design. Only when the new design demonstrably works better do you extend it.
Roll out by section. You don’t have to change the whole site at once. Start with the homepage, then product pages, then the blog.
Soft launch. Run the new site on a subdomain or for a limited group and collect feedback before opening it to everyone.

This takes more planning and technical work. It’s worth it — because redesign mistakes are far easier to fix on one page than across a whole site.

7. No testing with real users

The last and possibly most critical mistake: nobody asked actual users. The site was approved by the director, the marketing manager and the designer. All three know it perfectly. None of them sees it through the eyes of a customer looking for the first time.

Why internal approval isn’t enough

The curse of knowledge. You know where everything is, what each service means and how to reach an enquiry. The customer doesn’t.
Different priorities. The internal team worries whether the colours match the brand manual and whether the photos are good enough. The customer worries whether they understand what the company offers and what it costs.
Blindness to problems. Look at something every day and you stop seeing its problems.

How to test with real users (without a budget)

Ask five people from your target audience to complete a specific task on the new site. For example: “Find the price of service X and submit an enquiry.” Watch where they falter.
Guerrilla testing. Sit a colleague from another department (or a friend with no connection to your company) in front of the site and ask them to narrate what they see and do. Ten minutes will tell you more than a month of internal discussion.
A session recording tool. You’ll see where people click, where they scroll, where they leave. It costs a few euros a month and saves you a repeat rebuild.

Five user tests surface most of the critical problems. Not perfect, but incomparably better than none — what user testing reveals is worth a read before you commit to a rebuild.

How to do a redesign that actually raises conversions

Now you know what not to do. Here’s the approach that works.

1. Start with data, not Figma

Before you open a design tool, find out:

What works on the old site? Which pages convert best? Where do people spend the most time? Where do the customers who actually pay come from?
What doesn’t? Where do people leave? Which forms have low completion? Where do visitors get lost?
What do customers say? Why did they choose you? What was missing on the site? What nearly put them off?

This is the foundation of a redesign. Without it you’re guessing.

2. Define measurable goals

“We want a more modern site” isn’t a goal. Goals look like this:

Raise enquiry form conversion from 2 % to 3 %
Bring bounce rate on product pages under 50 %
Increase submitted enquiries by 20 % in 3 months

With clear goals you can tell whether the redesign did its job. Without them you’ll be discussing feelings.

3. Keep what works

Not everything on the old site is wrong. Find the elements that demonstrably bring conversions and deliberately keep or improve them — don’t delete them.

4. Test before launch, measure after

User testing before launch. A/B tests after. Data every week. Readiness to fix quickly what doesn’t work.

5. Plan the redesign as a process, not a one-off project

The best sites don’t come from a single redesign. They come from continuous optimisation — small data-driven changes, tested with real users, measured over time.

Not sure whether your site converts as well as it could?

A redesign is a large investment. Before committing, it makes sense to know exactly what isn’t working on the current site — and whether it needs a redesign at all, or just targeted fixes.

Kaen goes through your site continuously — copy, UX, technical state and forms. Every problem comes with a concrete recommendation and a measurable hypothesis. Not general advice, but steps you can ship now or hand to a developer.

You may find you don’t need a new website. You may just need to fix three things on the one you have.