Measuring website redesign success

The honest way to evaluate whether a redesign worked is to compare the same specific metrics before and after, over a comparable time period. Do not rely on a general impression that the new site “feels better”. Impression is not evidence. This cluster does not treat it as such.

What to compare, specifically, to measure success after a website redesign, starting with conversion

Conversions, not just traffic. The metric that actually answers whether a redesign achieved anything for the business is simple: does it produce at least as many, ideally more, genuine enquiries or conversions as the old site did over a comparable period? This is covered in the broader context of what to track on what to actually measure on a small business website. Traffic alone, per this site’s general position on vanity metrics, does not answer this question.

Organic search visibility, isolated from the general traffic number. A redesign is one of the specific moments organic rankings are genuinely at risk (see redesigning without losing rankings). Search Console’s performance data — clicks, impressions and average position for the site’s main queries — deserves its own before-and-after comparison. Don’t just fold it into a glance at total sessions.

Page-level performance for the pages that mattered most on the old site. A specific service page or the homepage may have carried meaningful traffic and conversions before the redesign. That same page’s performance is the most direct like-for-like comparison available, even if its URL, design or content changed. It beats a site-wide average, which can mask a genuine problem on one important page.

Mobile-specific performance, given how much traffic typically arrives on mobile. A redesign that improved desktop metrics while mobile usability or speed quietly regressed is a genuine partial failure that a site-wide average could hide.

Establishing the baseline before launch: what actually needs measuring, and why that measurement can’t be reconstructed afterwards

The comparison in this page only works if the pre-redesign numbers were captured deliberately, not reconstructed from memory after the fact. Keep the same GA4 property rather than creating a new one for the new site. A new property discards the historical record entirely. Done that way, the pre-redesign baseline already exists in the property’s own history and does not need to be separately archived. What does need to be actively pulled and saved before launch is Search Console’s performance data for the specific queries and pages that mattered most. Search Console’s own data retention window is limited. A baseline not captured in time cannot be recovered later. Decide which few metrics count as “the baseline” as a specific pre-launch step, not an afterthought. Don’t wait until after launch to work out which numbers would have been useful to compare against.

Using GA4’s own comparison feature for measuring website success, rather than checking two reports separately

GA4 supports adding a comparison date range directly to most standard reports. It shows the current period and a chosen prior period side by side on the same charts and tables, instead of requiring two separate report exports to be manually lined up. Set this comparison explicitly to the pre-redesign period. Don’t leave it on GA4’s own default, which is usually just the immediately preceding period of equal length. That keeps the evaluation anchored to the actual before-and-after question, not a generic recent trend.

Performance metrics, Core Web Vitals and engagement belong in this evaluation too

A redesign frequently changes a site’s technical performance, for better if the rebuild addressed genuine bloat, or for worse if new visual features were added without a performance budget. Core Web Vitals and general page speed, covered in the mobile performance context on mobile page speed and why it differs from desktop, are worth comparing before and after, alongside the conversion metrics above. A redesign that improved visual design while quietly making the site slower has made a genuine trade-off. That is a trade-off a conversion number alone will not always reveal immediately.

What qualitative signals add that the numbers alone do not

Alongside the measured comparison, get direct feedback from whoever actually handles incoming enquiries — a receptionist, a sales team, the business owner themselves. Whether enquiry quality or the kinds of questions being asked have changed is a genuine, if unquantified, signal worth recording alongside the analytics comparison. It is not a substitute for measured data, and should never be presented as if it were. But a redesign that measurably increased enquiry volume, while the team reports the new enquiries are consistently a worse fit for the business, gives a more complete picture than the volume number alone.

Why the comparison window matters

Comparing the week immediately after launch against the week before it risks measuring normal post-launch fluctuation, not a genuine result. Search engines take time to re-crawl and re-evaluate a changed site. Visitor behaviour can also be temporarily unsettled by unfamiliarity with a new layout, regardless of whether the redesign is actually an improvement. A more reliable comparison waits for a period of stable post-launch data — generally several full weeks at minimum. It should compare against the equivalent period a year earlier, as well as the immediate pre-redesign period, to account for any seasonal pattern in the business.

A practical comparison table: conversion rate, organic visibility and engagement

Metric	Compare against
Conversions (enquiries, bookings, sales)	Same period pre-redesign, and same period a year earlier
Organic clicks and average position (Search Console)	Same period pre-redesign, monitored weekly through the transition
Performance of the specific pages that mattered most before	The equivalent new page, even if URL or design changed
Mobile-specific conversion rate and engagement	Desktop performance on the same site, to catch a mobile-specific regression

What a genuinely honest redesign success result looks like

Not every redesign improves every metric, and an honest evaluation says so plainly where that is the case. A redesign that improved usability and brand perception while search traffic dipped temporarily during re-indexing, for instance, is a mixed but explicable result — not a failure to be hidden or a success to be overstated. The evaluation exists to inform the next decision, not to justify the last one.

What to do next

Set the specific metrics and comparison periods above before the redesign launches, not after. That way there is a genuine pre-redesign baseline to measure against. What to actually measure on a small business website covers the broader set of metrics this evaluation draws on. The pricing dataset is the methodology this cluster’s own honesty-about-evidence standard is modelled on.