Site speed fixes ranked by effort against payoff: images, scripts, apps, fonts and hosting. Why speed moves conversion and where to start this week.
Speed is the most measurable conversion lever a store has, and the evidence is unusually direct. Site speed optimisation isn't a developer vanity metric. Google's mobile research with SOASTA found bounce probability rises 32% as load time stretches from one second to three. Deloitte and Google's "Milliseconds Make Millions" study found that a 0.1 second improvement in mobile site speed lifted retail conversion by around 8%. It's checkout revenue with a latency problem.
This is the infrastructure pillar piece. Here's the fix list, ranked by effort against payoff, so you can spend a week well instead of a quarter badly.
Two numbers matter, and they come from different places:
The three Core Web Vitals: LCP (how fast the main content shows, target under 2.5s), INP (how fast the page responds to taps, target under 200ms), CLS (how much the layout jumps, target under 0.1). For ecommerce, LCP on product and collection pages is almost always the money metric.
| Fix | Effort | Payoff | Notes |
|---|---|---|---|
| Compress and resize images, serve WebP | Low | High | The top fix for most stores. Hero images are the usual LCP element |
| Remove dead apps and scripts | Low | High | Every installed app you stopped using is still loading JavaScript on every page |
| Lazy-load below-the-fold images, never the hero | Low | Medium | Lazy-loading the LCP image makes things worse, a common theme mistake |
| Reduce font weight count, preload the main font | Low | Medium | Two families, limited weights. Each weight is a download |
| Audit third-party scripts (chat, reviews, heatmaps, pixels) | Medium | High | Load them after interaction or delay them. Tag managers full of legacy tags are silent killers |
| Fix layout shift: set image dimensions, reserve banner space | Medium | Medium | CLS fixes are fiddly but stop the rage-tap misclicks |
| Theme-level code work | High | Medium to high | Worth it when the theme itself is the bottleneck |
| Hosting / CDN changes, replatforming | High | Varies | Last resort. See the migration guide before touching this |
On hosted platforms the biggest recurring offender is app accumulation. Worked example: a store installs a countdown timer for one campaign, a popup tool for one season, two review apps during a switchover. Two years later all four still inject scripts on every page view. Quarterly app audits belong in website maintenance as standing work: list every app, mark the ones earning their load, remove the rest, and check the theme code afterwards because many apps leave script tags behind after uninstall.
Yes. Core Web Vitals are a ranking signal, assessed on field data. The bigger effect is indirect: faster pages bounce less and convert more, and Google's research quantifies both.
Under 2.5 seconds in field data on your product and collection templates. Under 2 seconds is a strong result on a hosted platform.
Lab tests run on a throttled simulated device. Field data reflects real visitors. Manage the field data and use lab runs only to diagnose.
Qwrki runs the speed sprint as standing work inside the operating layer, not as a one-off project. We pull field data per page type, schedule the image, app and font passes, then re-test on the rolling window so regressions get caught before they cost checkouts. Book a call and we'll start with a free read of your product and collection templates.
Shopify vs WooCommerce, compared on what actually differs: who carries the operational burden, the real cost shape, flexibility, and when each one honestly wins.
A website migration SEO checklist: redirect mapping, pre-launch benchmarks, launch-day verification and the monitoring window that catches problems early.
Screaming Frog and Ahrefs will dump 4,000 issues on you. About 40 actually matter. A real technical SEO audit is the work of separating the two — here's the shortlist.
