The bottom line
The performance audits we run on inherited sites almost always reveal the same pattern: someone, at some point, did a generic 'speed optimisation' pass that compressed images, minified CSS, and enabled basic caching — and the site is still slow. The reason is that the actual bottleneck wasn't on the front-end. It was a 4-second Time to First Byte caused by an unindexed database query running on every page load. Or it was a render-blocking JavaScript bundle from an analytics tool nobody is reading. Or it was a Cloudflare configuration that was actually caching dynamic content and serving stale checkout pages. Generic optimisation passes apply known techniques to all sites equally — and on sites where the bottleneck isn't in the targeted layer, they make no measurable difference. in Derby JW Digital diagnoses the real bottleneck before fixing anything. Lighthouse and WebPageTest give the synthetic view. Search Console field data shows what real users actually experience (often very different from lab tests). Server logs reveal where time is being spent. Database slow-query logs surface the queries that drag every page. APM tools like New Relic or Datadog identify which code paths consume disproportionate resources. Once the actual bottleneck is identified, we fix at the right layer — server-side caching and database tuning where the origin is slow, CDN configuration where global delivery is the issue, JavaScript bundle splitting where third-party scripts are the problem, image and font optimisation where the browser is bottlenecked. The outcome is documented before/after improvement in Core Web Vitals and real-user metrics — not a vague speed-up promise or a meaningless 90+ Lighthouse score that doesn't reflect actual user experience. Pairs naturally with our managed website hosting (where the server-side performance work lives), bespoke website design (where new builds are tuned from the start), and SEO agency services (where Core Web Vitals contribute to search rankings).
