The web design and development metrics that matter are the ones that link the site to business results: conversion rate on key pages, task completion, Core Web Vitals, error and uptime rates, and how quickly the team ships fixes. Visual polish and page counts are not metrics. A redesign should be judged on whether visitors find what they came for faster and act more often.
Many web projects are signed off on taste. Someone senior likes the homepage, the site launches, and nobody checks months later whether it performs better than the old one. Defining metrics before design starts changes that.
A website is a tool with a job: generate leads, sell products, answer support questions, recruit staff. Measuring tells you whether the design does that job. It also settles disagreements. When two stakeholders argue about a layout, a clear metric and a test end the argument faster than opinions.
Metrics also protect the project after launch. Sites decay as content is added, plugins pile up and scripts multiply. A baseline lets you spot when performance slips.
Which business metrics should a website be judged on?
Start with outcomes. Pick the actions that represent value for the business and track them as conversions in GA4.
- Conversion rate per key page: product pages, pricing, contact, sign-up.
- Lead or sale volume from organic and direct traffic, which the site influences most.
- Revenue per session for ecommerce.
- Form completion rate: starts versus submissions, which reveals friction in the form itself.
- Micro-conversions: newsletter sign-ups, brochure downloads, clicks to call.
Compare these for the same months before and after a redesign, and segment by device. A mobile conversion rate far below desktop usually points to a mobile design problem rather than an audience difference.
What do Core Web Vitals tell you?
Core Web Vitals are Google’s published measures of loading, responsiveness and visual stability. The thresholds for a good experience are Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds and Cumulative Layout Shift under 0.1, measured at the 75th percentile of real visits.
Check field data in Search Console under Core Web Vitals and in PageSpeed Insights. Field data comes from real users in the Chrome UX Report; lab data from Lighthouse is useful for debugging but does not match what visitors experience.
Common causes when vitals fail
- LCP: oversized hero images, slow server response, render-blocking CSS or fonts.
- INP: heavy JavaScript, third-party scripts, large event handlers on clicks.
- CLS: images without width and height, late-loading banners, web fonts swapping.
Set a performance budget in the project, such as a maximum page weight and script count, and check it before each release.
How do you measure usability, not just speed?
Speed is necessary but not enough. A fast site with confusing navigation still loses visitors. Usability metrics show whether people can complete tasks.
Run task-based tests with five to eight users per round: ask them to find a price, book a demo or locate a return policy, and record success, time and errors. Tools such as Maze or Lookback make remote testing practical. Session recordings and heatmaps from Microsoft Clarity or Hotjar show where real visitors hesitate, rage-click or abandon.
- Task success rate for the three to five most important tasks.
- Time on task, compared between old and new designs.
- Rage clicks and dead clicks on key pages, reported in Clarity.
- Site search queries with zero results, which reveal missing content or labels.
Which development metrics show build quality?
Development quality shows up in how the site behaves over time. A well-built site stays fast and stable as content grows; a fragile one breaks with each update.
- Uptime, monitored with a service such as UptimeRobot or Better Stack.
- Server error rate: 5xx responses in logs or monitoring.
- JavaScript errors per session, from a tool like Sentry.
- Accessibility issues found by automated checks such as axe, plus manual keyboard testing.
- Deployment frequency and time to fix a reported bug.
Accessibility is worth singling out. The WCAG 2.2 guidelines at level AA are the common target; automated tools catch only part of the issues, so include manual checks for keyboard navigation and screen reader labels.
How do SEO metrics fit into a redesign?
Redesigns can lose search traffic if URLs change without redirects or if content gets cut. Track organic clicks and impressions in Search Console for the key pages before launch, keep a URL map from old to new, and set 301 redirects for every changed address.
After launch, watch the Pages report in Search Console for new errors, and compare organic clicks week by week. A drop in the first weeks is common while Google recrawls; a drop that persists after a month needs investigation.
An example measurement plan
For example, a regional law firm redesigning its site might set targets of a higher contact-form completion rate on mobile, all Core Web Vitals in the good range for its top twenty pages, no loss in organic clicks after eight weeks, and zero 404 errors from old URLs. Each target has a source: GA4, Search Console and a crawl. The specific targets would come from the firm’s own baseline.
How should results be reported to clients?
Keep reports short and tied to the goals agreed at the start. A monthly one-page summary with a handful of metrics, their trend and one recommended action works better than a long export. Use Looker Studio to pull GA4 and Search Console data into one view.
Report problems honestly. If mobile conversion fell after launch, say so, explain the likely cause and propose a fix. Clients trust teams that surface issues before being asked.
What should you avoid measuring?
Some numbers look like progress and mean little. Total page views rise when navigation is confusing and people click around looking for something. Average time on site rises when pages are hard to scan. Bounce rate in older analytics setups depended on tracking quirks; GA4 replaced it with engagement rate for a reason.
Number of pages launched, number of components built and design award entries are outputs, not outcomes. They belong in a project log, not in a report about whether the site works.
- Replace page views with conversions and task success.
- Replace time on site with engaged sessions and scroll depth on key pages.
- Treat any metric that rises when users struggle as a warning sign.
Where to start this week
Record your baseline today: conversion rate by key page and device from GA4, Core Web Vitals status from Search Console, and organic clicks for your top pages. Save the numbers with the date.
Then pick the one page that matters most to revenue and watch ten session recordings in Clarity. Note where people stall. That list becomes the first set of design changes worth testing.
Keep reading: Web Design and Development · Web Design