Website Redesign Without Losing Conversion

The Hidden Risk of Website Redesigns

Redesigning a website is one of the riskiest decisions a digital business can make. According to various industry studies, between 30% and 50% of redesigns result in a conversion drop in the first months.

Why? Because most redesigns are based on aesthetic opinions, design trends, or the need to "modernize" the image, without measuring what currently works well or validating changes before implementing them.

When a Redesign Makes Sense

Not all conversion problems are solved with a redesign. These are the legitimate scenarios:

• Unsustainable technical debt: the current stack prevents improvements or is too slow.
• Business model change: the site structure doesn't reflect the current offering.
• Unacceptable mobile experience: the site wasn't designed for mobile and patches aren't enough.
• Data supports it: usability audits and tests point to structural problems that can't be fixed with incremental changes.

If your site converts reasonably well but "looks outdated," a complete redesign probably isn't the best investment. Incremental, data-driven changes are safer and more efficient.

The CRO Framework for Redesigns

Phase 1: Audit the Current State

Before designing anything, thoroughly document what works and what doesn't:

• Conversion metrics by page: conversion rate, bounce rate, time on page for each important URL.
• Heatmaps and recordings: where users click, how far they scroll, where they get frustrated.
• Historical A/B tests: what changes worked and which didn't. Validated learnings must survive the redesign.
• User feedback: surveys, reviews, support tickets mentioning website problems.

Phase 2: Define What to Keep

From the audit, extract a list of elements that shouldn't change:

• CTAs with good click-through rates.
• Conversion flows that work (checkout, forms).
• Trust elements (testimonials, logos, certifications) in validated positions.
• Navigation structure that users understand.

Phase 3: Design with Hypotheses

Each change in the redesign should be formulated as a testable hypothesis:

"We believe changing the navigation from hamburger to horizontal bar on desktop will improve product discovery rate because 65% of users don't interact with the current menu."

Phase 4: Launch in Phases

The most serious mistake is launching the complete redesign all at once. The CRO approach:

1. Launch the new homepage and measure for 2-4 weeks.
2. Launch product pages and measure.
3. Launch checkout and measure rigorously.
4. Launch the rest with confidence that critical pages are already validated.

If you detect a significant conversion drop at any point, you can revert that specific section without affecting the rest.

Mistakes That Destroy Conversion

Changing URLs Without Redirects

A redesign usually means new URLs. Without 301 redirects for all old URLs, you lose accumulated SEO positioning and create 404 errors that frustrate users.

Removing Trust Elements

In the rush to simplify, many redesigns remove testimonials, client logos, security badges, or FAQs that were playing a crucial role in the purchase decision.

Overloading with Animations

Modern animations improve aesthetic perception but can hurt loading speed and distract from the main CTA.

Ignoring SEO

A redesign is an opportunity to improve technical SEO, but also a risk to destroy it. Ensure titles, meta descriptions, heading structure, alt text, and Core Web Vitals are maintained or improved.

Pre-Launch Checklist

Before going live with any redesign phase:

• All 301 redirects configured and tested.
• GA4 tracking verified on all new pages.
• Conversion events working correctly.
• Loading speed equal to or better than the previous version.
• Tested on mobile, tablet, and desktop across major browsers.
• Trust elements present in validated positions.
• Forms functioning and sending data correctly.
• Complete checkout flow tested end-to-end.

The Alternative: Iterative Redesign

Instead of a complete redesign, many companies get better results with an iterative approach:

1. Identify priority problems with data.
2. Redesign only the affected components.
3. Test each change with A/B testing.
4. Implement what works, discard what doesn't.
5. Repeat.

This approach is visually slower — your site evolves gradually instead of changing all at once — but it's significantly safer for conversion and generates cumulative learnings.

---

At Boost, we support website redesigns with a CRO approach so the new site doesn't just look better, but converts better. If you're planning a redesign, talk to our team or analyze your current site with Scan&Boost.

Adrià Vidal — Boost