Speed without the theatre.

Start with the customer's task

Can the page show its main content promptly? Do controls respond? Does the layout stay still enough to read and tap? These are better questions than whether an impressive animation appears on your development computer.

Image dimensions, appropriate image variants, sensible loading order and restrained JavaScript are useful implementation decisions to examine. Measure the important pages, not just an empty test route.

Know which evidence you have

The Core Web Vitals guidance identifies good thresholds of LCP at or below 2.5 seconds, INP at or below 200 milliseconds and CLS at or below 0.1. Field assessment uses the 75th percentile of visits, with mobile and desktop evaluated separately.

A laboratory test can help find problems before launch. It does not turn into field evidence just because its score is green. A new or low-traffic site may not yet have sufficient public field data. Record that as unavailable, not passed.

Make a sensible budget

Our design default is to justify every heavy resource by the customer benefit it provides. The main makeover image has a purpose; five competing background videos have a much harder case to make.

Use consistent test conditions when comparing changes. Record the tested build, route, viewport and network assumptions. Otherwise you may be comparing two different experiences and drawing a very confident conclusion from the wrong race.

Do not flatten the brand

Performance work should preserve useful visual communication. Compress an image appropriately before removing the very example that explains the business. The goal is an effective experience with a sensible delivery cost, not a page stripped of meaning.

A practical checklist

• Record the route, build and test conditions.
• Separate lab results from real-user field data.
• Prioritise resources that explain the offer and support the task.