Site speed affects SEO through three mechanisms: direct ranking impact, crawl efficiency, and user behavior signals. Google uses page speed as a ranking factor on mobile (since 2018's Speed Update) and desktop. Sites that load under 2.5 seconds for Largest Contentful Paint generally perform better than those over 4 seconds, though speed alone won't vault you past stronger content and backlinks. Crawl budget matters more than most people realize. If your pages take 800ms to respond instead of 200ms, Googlebot crawls fewer pages per session. For large sites—anything over 10,000 URLs—this directly limits how much fresh content gets indexed. We've seen e-commerce clients with 50,000+ product pages where shaving 400ms off server response time led to 35% more pages crawled weekly. User behavior creates the indirect ranking effect. A page that takes 5 seconds to load on mobile sees bounce rates near 90% in analytics. Google tracks pogo-sticking and dwell time as quality signals. Even if speed isn't manually tanking your rankings, users leaving in 3 seconds sends a bad signal. Conversion data backs this up—Amazon's internal studies showed every 100ms of latency cost them 1% in sales. Core Web Vitals formalized this. LCP (load speed), FID (interactivity), and CLS (visual stability) now feed into rankings. You don't need perfect scores—a site with LCP at 2.8s and good content will outrank a 1.2s site with thin content—but crossing into "poor" thresholds (LCP over 4s, CLS over 0.25) creates a real handicap. At Ottawa SEO, we treat speed as table stakes. For local service sites competing in Ottawa or Toronto, getting under 3 seconds on mobile keeps you competitive. For national e-commerce or SaaS, we aim for sub-2s LCP because you're fighting better-resourced competitors. The ROI shows up fastest on high-traffic pages—your homepage, category pages, top landing pages from organic. Fix those first before optimizing a blog post that gets 12 visits a month.