Website Optimization

I would start with a full [SEO audit](/seo-audit/) before touching any files. Most people dive straight into compressing images or minifying code, but that can miss deeper issues. On a recent project, a site crawl revealed that 30% of pages were returning 404s because a plugin update had broken internal links. Fixing those brought more value than any image optimisation. Step one: run a crawler, check Search Console for index coverage and manual actions, and measure real-user [Core Web Vitals](/core-web-vitals/) from the Chrome UX Report. Only then do I prioritise fixes. I group issues by impact: fix [duplicate content](/duplicate-content/) and broken redirects first, then tackle page speed. For performance, I compress images to WebP, enable caching headers (Cache-Control for 7 days on static assets), and serve from a CDN like Cloudflare. I review the [robots.txt](/robots-txt/) file to ensure critical pages aren't blocked, and check that [canonical tags](/canonical-tag/) point to the correct URLs — misconfigured canonicals can dilute ranking signals. After technical foundations, I look at content quality and internal linking. Google's SEO Starter Guide still emphasises relevance and authority. A fast page with thin content won't hold rankings. Finally, I verify mobile usability on a real device, not just DevTools. That order — audit, technical foundations, performance, content — has consistently given me the best ROI. For example, on a site with 10,000 pages, fixing the crawl issues and canonical errors improved indexed pages by 60% in two months, while performance optimisation alone would have missed the root cause.

Biggest watch-out: treating optimisation as a one-off. I've seen sites revert within months because no one monitors new content or updates. Another: chasing a perfect Lighthouse score while ignoring real user metrics. A score of 100 doesn't matter if your field LCP is 3 seconds due to server response time. Also, be careful with over-optimisation. I once saw a site that deferred all CSS, causing a flash of unstyled content that hurt CLS. Test on a real Android device, not just an iOS simulator. Mobile-first indexing means your mobile pages must be fully optimised — that includes touch targets and font sizes, not just speed. Watch out for lazy loading above-the-fold images — that actually hurts LCP. I also caution against ignoring content: a fast site with poor content won't convert. Third-party scripts are another trap. A single tracking script can add 1 second to LCP. Audit every script and consider asynchronous loading. Beware of redirect chains: every [301 redirect](/301-redirect/) adds latency, so keep chains to one hop. Also, don't rely solely on automated tools. Manual checks for page experience and usability catch what crawlers miss. For instance, a site I audited had perfect validation but the search box was hidden on mobile. That's not an optimisation win. Finally, ensure you measure the right metrics — field LCP, CLS, and INP from CrUX, not just lab data. If you're optimising for the wrong goal, you'll waste effort.

I used to believe website optimisation was purely about page speed. I spent weeks minifying CSS and compressing images on a large ecommerce site, only to realise the real issue was a heavy third-party widget that added 2 seconds to LCP. I also treated it as a one-off project — optimise and walk away. Six months later, the site had regressed because new content added without optimising images dragged the page weight back up. Another mistake: copying tactics from another site without checking whether they applied. I added lazy loading to above-the-fold hero images, which actually hurt LCP because the browser waited to load a critical element. Now I test each change against field data before rolling out. I also measured vanity metrics like "time to interactive" when the job-to-be-done was faster checkout. I've learned to start with the user's goal and work backward. Early on, I ignored mobile optimisation, assuming desktop fixes would carry over. They don't. A responsive layout with slow mobile performance is still a failure. Finally, I overlooked internal linking as part of optimisation. Adding contextual links from high-authority pages spread link equity better than any technical fix. I also once spent a month optimising a site that later lost rankings because I forgot to update the sitemap after a site migration. That taught me to always double-check indexation after any optimisation round.