A PageSpeed score of 95 or higher, on both mobile and desktop.

Usually, yes. The fixes that move the needle most, like image optimisation, deferring scripts, and CDN configuration, can often be done without rebuilding anything. We audit first, then tell you whether tidying up the current site or rebuilding it works out cheaper for you.

Google ranks pages using mobile PageSpeed no matter where your audience actually browses. So a weak mobile score still drags down your search rankings, including for the desktop visitors you care about.

We use WebPageTest with throttled profiles that mimic Indonesian 4G, and we measure on real mid-range Android handsets. That's a long way from lab numbers taken on a fibre connection inside a US data centre.

A Lighthouse score of 95+ is a clear, verifiable bar that maps to genuinely fast pages on real devices, not a vague target anyone can argue about. It also gives your team a number to defend against regressions over time. We ship against that floor and verify it before handover, so 'fast enough' is never left to opinion.

Core Web Vitals are three Google metrics: LCP (how fast the main content loads), INP (how quickly the page responds to interaction), and CLS (how much the layout shifts unexpectedly). Google factors them into rankings, and they closely track how fast the site feels to a real visitor. We measure all three on real devices and won't ship a failing vital.

The biggest factors are usually unoptimised images, render-blocking scripts, heavy third-party tags (analytics, chat, tracking), and how the server delivers pages. Fonts, the amount of JavaScript shipped, and whether content is cached at the edge also matter. We audit each of these and fix them in the order that gives you the most speed for the least effort.

Where your site is served from decides how quickly the first byte reaches the visitor. We serve from a CDN edge through Vercel or Cloudflare Pages, so pages come from a location near the user rather than a single distant server. Combined with static generation, this keeps Time to First Byte under 100ms even for visitors far from the origin.

Both. Lab tools like Lighthouse and WebPageTest give us a controlled, repeatable score during the build. We also look at field data (Core Web Vitals from real Chrome users) in Search Console once the site is live. Lab numbers catch issues early; field data confirms the experience your actual audience is getting.

Not if they're loaded carefully. We defer or async render-blocking scripts, load analytics after the first interaction, and load chat widgets only when someone opens them. Each script is weighed against what it actually costs in speed, so you keep the tools you need without paying for them on every page load.

That's the part most teams miss. We set image guidelines, configure the CMS to compress and resize uploads, and where it fits we add a CI check that blocks any deployment that would drop the Lighthouse score below the agreed number. So the site stays fast as your marketing team adds pages, not just on launch day.