Everyone worries about losing rankings in a redesign. Far fewer worry about losing conversions, which is the one that shows up in revenue first. The pattern is familiar: the new site is unquestionably more beautiful, everyone is pleased at launch, and six weeks later enquiries are down a third and nobody can say exactly why. It is avoidable, and the prevention is mostly done before any design starts.
Baseline the old site properly
You cannot tell whether a redesign worked without knowing what the old one did, and “we get about twenty leads a month” is not a baseline.
Record, for at least the three months before launch: conversion rate by page, by device, and by traffic source; the number of leads and their quality as sales sees it; the path people actually take to convert; and the pages that produce enquiries even though nobody thinks about them. Screenshot the current site while you are at it, because in nine months nobody will remember what it looked like when the argument about what changed begins.
Segment it. A site that converts at 4 percent on desktop and 1 percent on mobile is telling you where the real work is, and an average hides both numbers.
Find out what is carrying the conversion
Every site has elements doing unglamorous heavy lifting, and they are rarely the ones the team is proud of.
It might be a phone number in the header that half of your customers use. A specific FAQ answering the objection that kills deals. An ugly comparison table. A trust badge. The sentence in the footer about being family run since 1994. Those items get removed during a redesign for looking inelegant, and the loss never gets attributed because nobody inventoried them.
So inventory them. Walk the current conversion path and list every element that answers a question, reduces a risk, or provides a route to contact. That list is a constraint on the new design, not a suggestion. Redesign how they look, not whether they exist, and check the proof elements in particular, since they are the most commonly dropped for aesthetic reasons.
Beware the beautiful failures
Three specific patterns cause most redesign conversion drops.
Fewer words. Design driven rewrites cut copy in half because whitespace looks sophisticated. But long consideration purchases need information, and stripping the page to a headline and a button removes the answers people needed to say yes. Length is not the enemy, waffle is.
Buried actions. The old site had a crude but visible enquiry button on every section. The new one has one tasteful call to action at the bottom of a very long scroll. Anyone who decides early now has nowhere to click.
Elegant forms with more fields. Redesigns often add a budget dropdown and a company field because the new layout has room. Every field is a toll, and the completion loss is measurable.
Also watch for slower pages hiding under a nicer surface. A hero video and three web fonts can undo whatever the design gained, especially on the phones where most of your traffic lives.
Do not change everything at once, if you can avoid it
The cleanest way to protect conversion is to avoid a single simultaneous change of design, copy, structure, and platform, because when the number moves you will have no idea which part did it.
Where traffic allows, ship the new templates to a share of visitors first and compare against the old, which is the only version of this that produces real evidence. Where it does not, sequence the change instead: structure first, then design, then copy, with a few weeks between, or roll out template by template.
Launch in a way you can reverse
Keep the old site available in a state you can restore quickly, and make sure the analytics on the new site are working before it goes live rather than after, because a week of missing data at launch is a week you cannot compare. Verify every goal and event fires on the new templates; renamed buttons and changed form endpoints break tracking silently, and “conversions dropped” is sometimes just “tracking broke”.
Then watch for six weeks, not six days. Traffic mix, device split, and seasonality all wobble in the first fortnight after a launch, and the honest verdict needs a month at least.
If something did drop, resist redesigning again. Find the specific page and the specific step, compare it against your inventory of what carried conversion before, and put the missing thing back in the new site’s language.
Want your current site baselined before you touch it? hello@beconfidency.agency.
If you want a redesign judged on the numbers rather than on opinions, that is exactly what our website redesign service is for.