Website Speed Optimisation for Any Platform: Page Speed and Core Web Vitals

// Who is website speed optimisation for?

Website speed optimisation is for any site that loads slowly and is losing visitors, sales or search visibility because of it, whatever it is built on. Slowness rarely has a single cause; it usually comes from a slow server response, heavy images and scripts that load one after another. That is why we measure which layer weighs most before changing anything.

• Search Console's Core Web Vitals report flags URL groups as "Poor" or "Need improvement"
• The field data section of PageSpeed Insights shows LCP or INP in red
• Time to first byte (TTFB) is above 0.8 seconds and the page sits on a blank screen
• Content jumps while the page loads and buttons move under the user's finger
• The page freezes for a moment after tapping a menu, filter or basket button
• Paid traffic arrives but visitors leave before the page has loaded

We use the same measurement method for a brochure site, a custom-built portal or a shop with thousands of products; on shops, category, product and checkout templates are treated separately.

// Why can a site score 100 in the lab and still fail on field data?

Because the two measure different things: Lighthouse is a lab test that emulates one device and one connection, while field data (CrUX) is the experience of real Chrome users on your site over the previous 28 days. Google uses field data in Search Console and in its page experience assessment.

So a site can score 100 in Lighthouse and still fail in the field. The most common reason is INP: the lab test never taps the menu, but when a real user does, a heavy JavaScript task locks the page. Older phones and weak mobile connections are invisible in the lab as well. We measure both and report them side by side.

The thresholds apply at the 75th percentile of visits, so at least three in four visits must stay within them. Details are in Google's Core Web Vitals guide and the PageSpeed Insights documentation.

// What does the work include?

The work runs on two layers: the server decides how quickly the first byte arrives, and the front end decides how quickly that content becomes usable on screen. The table below shows what a one-off page speed optimisation includes and what stays outside it.

If the hosting itself is the bottleneck, the measurement shows it and we recommend a move as a separate step under Website & Hosting Migration.

// How does the process work and how long does it take?

There are four steps and a typical site takes 3 to 5 working days in total; for large shops and custom applications the timeline is written into the quote after measurement. We work on a copy of the site (staging), take a backup first and release changes to the live site during low-traffic hours.

// Which platforms do we work on?

We work on custom PHP applications, Node.js and Next.js projects, static HTML sites, front ends fed by REST APIs, e-commerce sites and WordPress. Each platform has its bottleneck in a different place, so we apply measurement rather than a fixed recipe.

• Custom PHP: slow database queries, uncached page generation, OPcache and PHP version
• Node.js / Next.js: server-side rendering time, static generation and caching strategy, size of the JavaScript bundle sent to the browser
• Static HTML: image dimensions, render-blocking CSS and fonts, cache headers
• E-commerce: category filters, product images, third-party scripts in the basket and checkout

If your site runs on WordPress, see our WordPress Speed & Performance page for plugin, theme and database specifics. On closed platforms such as Wix, Shopify or Squarespace there is no server access, so our scope is limited to images, scripts and content structure; we say so in the first audit. If speed should be handled together with crawlability and indexing, the work can be combined with Technical SEO.

// What determines the price?

Four variables set the price: the platform, the number of page templates to measure, the level of server access and how many third-party scripts the site carries. On a static brochure site the work mostly stays at image and CSS level, whereas a custom application needs work on database queries and server code. That is why we quote from measurement rather than a fixed price list.

The first audit is free and its report is shared with you; the quote is fixed after the audit and does not rise later. A target score is written into the quote, and if we do not reach it you are not charged the difference. Because field data builds up over a 28-day window, we carry out a free check measurement three months after delivery.

// Examples from our projects

On the sites we build, speed is part of the design rather than a setting added afterwards. The scores below are PageSpeed Insights measurements taken on 4 October 2026; the cards below link to the reports themselves, and you can test the same pages yourself today.

• Ergoterapistimle — Appointment-focused website for a paediatric occupational therapy centre: mobile 92, desktop 99 (median of 5 runs)
• AS Medya — Animated WordPress corporate site for a digital marketing agency: mobile 100, desktop 99
• Işık Konstrüksiyon — Multilingual WordPress site for a construction and architecture firm: mobile 100, desktop 100

Lab scores fluctuate from run to run, so we measure several times and report the median; your report uses the same method. An animated site scoring 100 on mobile shows that speed does not require giving up visual richness.

Any improvement made without measuring what slows the site down is a guess; we measure first and change things second.

// Who does the work, and when is it not needed?

The work is led by Caner Zep Çelik, Founder & Technical Consultant of Zep Bilişim. Since 2020 he has worked on more than 100 projects for clients in the United Kingdom, Turkey and Germany; he carries out the measurement and diagnosis himself and does not subcontract the work.

If all three metrics are green in your field data, website speed optimisation is unnecessary, and we tell you so in the first audit rather than selling work. If only the admin area is slow, visitors are not affected and a smaller fix will do. If the design and codebase are fundamentally outdated, optimisation has limits; in that case a Website Redesign may be the better investment.

// Frequently Asked Questions