The 30-second version
Slow websites don’t feel like a marketing problem, which is why they go unfixed for years. But a slow site bleeds quietly from three directions at once: visitors on phones give up before the page paints and call whoever loads next, Google folds page experience into rankings so the sluggishness drags your visibility, and every ad click you buy lands on a page that’s busy losing the race. The good news: after tearing down a lot of slow small-business sites, we can tell you the causes are boringly predictable. Here they are in the order we find them, how to diagnose yours in five minutes for free, and how to tell an afternoon fix from a rebuild.
Run the five-minute test first
Go to PageSpeed Insights (Google’s free tool, just search the name), paste your homepage URL, and wait. You’ll get two report tabs; look at mobile first, because that’s where your visitors are and where the pain concentrates.
Plain-language translation of what you’re seeing. The big performance score: above 90 is healthy, 50 to 89 means real friction, below 50 means your site is actively costing you visitors. Largest Contentful Paint is how long until the main content shows up; under two and a half seconds is the goal, and anything past four means people are staring at a blank screen deciding to leave. Cumulative Layout Shift measures the page jumping around while it loads (the reason you tapped the wrong button); it should be near zero. Then run your busiest service page too, since ads and searches often land there, not on the homepage.
Scroll down and PageSpeed lists the specific offenders. You don’t need to understand every line. You’ll usually find your site in the suspects below, and the report’s top items will tell you which ones.
Suspect one: images nobody ever resized
The most common cause, by a mile. Someone uploaded photos straight from a phone or camera: images thousands of pixels wide, several megabytes each, squeezed into a spot on the page the size of a playing card. The browser downloads the whole thing anyway. A hero section with three of these can outweigh everything else on the site combined.
The tells in PageSpeed: “Properly size images,” “Serve images in next-gen formats,” “Defer offscreen images.” The fix is mechanical: resize to the displayed dimensions, compress, convert to WebP, lazy-load anything below the fold. This is the single highest-leverage speed fix that exists, and it’s fully doable in an afternoon.
Suspects two and three: page-builder bloat and plugin pileup
Most slow small-business sites are WordPress running a visual page builder. Builders make editing easy by wrapping everything in layers of markup, styles, and scripts, so the browser does far more work to draw the same page. A simple five-section homepage arrives dressed as a small application.
Then there’s the plugin drawer: an average of one added per problem over the years, each loading its own scripts and styles on every page whether needed or not. The sliders, the popups, the social feeds, the seventeen features nobody remembers activating. Individually small, collectively an anchor. Deactivating the deadweight and consolidating overlapping plugins is afternoon work. The builder itself is structural: you can tune around it, but you can’t tune it away.
Suspects four through six: hosting, render-blocking junk, and no caching
Cheap shared hosting puts your site on a server with hundreds of others; when your neighbors get busy, your pages queue. The tell is a slow “server response time” flag in PageSpeed even on simple pages. Render-blocking junk is scripts and stylesheets that force the browser to stop and process them before showing anything: tracking pixels, chat widgets, font bundles, ancient code from tools you stopped using. And missing caching means every visit rebuilds and re-downloads everything from scratch instead of reusing what could be saved. All three are tunable: better hosting, deferring scripts, and a configured caching layer are each same-week fixes with visible results.
Afternoon fix or rebuild?
Here’s our honest sorting rule. If the site’s bones are sound and it’s carrying extra weight (giant images, plugin pileup, no caching, weak hosting), tuning gets you a long way fast. If the slowness is the architecture itself (a heavy builder stacking scripts on a strained platform), you can spend months tuning and still idle at mediocre, because you’re optimizing the anchor instead of cutting the rope.
We work through this decision in detail in the redesign-or-rebuild diagnostic. And for what the other end looks like: we build static sites with no database, no plugins, and nothing to bloat, which is how our own site posts a 99 PageSpeed score with zero hosting fees. That approach is the foundation of our web design work, because speed isn’t a feature you add later; it’s a consequence of how the thing is built.
Common questions
My site feels fast on my computer. Is the report wrong? Your computer has cached your entire site and probably sits on fast wifi. The report simulates a first-time visitor on a phone, which is who your customers are. Trust the report.
What score should I actually aim for? Above 90 mobile is genuinely good. But chase the experience, not the number: main content visible fast, nothing jumping around, every page snappy on a real phone.
Will fixing speed improve my rankings? It removes a drag rather than adding a rocket. Page experience is one ranking input among many, but the visitor side is immediate: fewer people bailing before the page loads means more calls from the traffic you already get.
Can I just install a speed plugin? Caching plugins legitimately help and are worth configuring. But a plugin can’t shrink your images for you retroactively or un-bloat a page builder. Plugins tune; they don’t transform.