Website Speed Optimization: Where to Start in 2026
The three Core Web Vitals targets, why real-visitor data matters more than one lab test, and where to start fixing.

Key takeaways
- 01Good Core Web Vitals are LCP within 2.5 seconds, INP of 200 milliseconds or less and CLS of 0.1 or less.
- 02These are measured at the 75th percentile of real page loads, so one fast test on one device is not enough.
- 03INP replaced First Input Delay as a Core Web Vital on 12 March 2024, so audits that still report First Input Delay are out of date.
- 04Fix the slowest page type first, such as checkout, product or landing pages, rather than the whole site at once.
Website speed optimization means making pages load and respond quickly for real visitors. Google's Core Web Vitals set the targets: LCP within 2.5 seconds, INP of 200 milliseconds or less and CLS of 0.1 or less, measured at the 75th percentile of page loads. Start with real-visitor data, then fix the slowest page type first.
What the three Core Web Vitals measure
- Largest Contentful Paint (LCP) measures how long the main content takes to appear. Aim for 2.5 seconds or less.
- Interaction to Next Paint (INP) measures how quickly the page responds when a visitor clicks, taps or types. Aim for 200 milliseconds or less. INP replaced First Input Delay as a Core Web Vital on 12 March 2024.
- Cumulative Layout Shift (CLS) measures how much content moves while the page loads. Aim for 0.1 or less.
Use real-visitor data, not one lab test
A lab test runs one simulated load on one device and one network. Field data records what real Chrome users experienced over a period, and the Core Web Vitals targets are set against that kind of data at the 75th percentile. In PageSpeed Insights, read the field data section before the lab score. Search Console's Core Web Vitals report shows similar data across your site. Check mobile and desktop separately, because they often differ.
Find the slowest page type first
Group your pages by type: home page, category or product pages, checkout, blog posts and landing pages. Check each group's field data and start with the group that has the most visitors and the weakest scores. A shared template often causes problems on many pages, so one fix can improve a whole group.
Common causes and fixes
LCP: slow hero images and blocking code
Large, unoptimised hero images, slow server responses and CSS or scripts that block the first render usually delay LCP. Compress images, use modern formats such as AVIF or WebP, size each image to its display area, and load the main image early.
INP: long scripts and heavy tags
INP suffers when the browser's main thread is busy. Long JavaScript tasks, heavy menus and third-party tags, such as chat widgets, tag managers and review scripts, make taps and clicks feel slow. Load non-essential scripts later, remove tags that nobody uses, and split long tasks into smaller pieces.
CLS: reserve space for everything
Images, embeds, ads and cookie banners that appear late push content down. Set width and height, or an aspect ratio, on images and video, reserve space for embeds, and avoid inserting content above the text a visitor is reading.
Check plugins, apps and tags after every change
Each new plugin, app or tag can add scripts to every page. Measure the page before you install it and again after. Keep a short list of what each script does, and remove the ones that no longer earn their place.
Keep speed in the release process
Speed slips quietly after theme updates, new campaigns and new tags. Add a speed check to your release steps, and watch the field data for several weeks after a major change, because it changes slowly.
How Elgrith approaches speed work
We start with field data and the page types that matter most to your business, then agree the target and the fixes in a written scope. We do not promise a PageSpeed score, because a lab score is not what the Core Web Vitals targets are set against. We give a fixed quote after a free call. See our website speed optimization service for the work we do.
Frequently asked questions
Is a PageSpeed score of 90 the goal? Not on its own. A lab score moves with the test device and network. Aim for good field data at the 75th percentile for LCP, INP and CLS.
Does speed alone decide rankings? No. Speed is one part of the page experience, alongside content, links and technical health.
Which fix should come first? Start with the page type that has the most traffic and the weakest field data. Image and third-party script fixes often give the largest early gains.
