The Goblin has reluctant feelings about site speed. It is, on the one hand, a real ranking factor and a real source of lost sales. It is, on the other hand, a topic so thoroughly hijacked by tool-vendors and dashboard-merchants that the average shopkeeper now believes their shop must score 100 in PageSpeed Insights or be punished.
This is not true. Let us, briefly, sort what matters from what does not.
What actually matters.
Google uses three measurements, collectively called Core Web Vitals. Their names are unfortunate but their meaning is not difficult.
- Largest Contentful Paint (LCP). How long it takes for the main thing on the page to appear. On a product page, this is usually the photograph. Target: under two and a half seconds.
- Cumulative Layout Shift (CLS). How much the page jumps around as it loads. If your customer is about to click the Buy button and the page shifts so they accidentally click a banner, that is the layout shift Google is measuring. Target: under 0.1.
- Interaction to Next Paint (INP). How long the page takes to respond when the customer taps something. Target: under 200 milliseconds.
That is the entire list. If your shop scores well on those three, the rest of the “score” in PageSpeed Insights is largely cosmetic.
What does not matter as much as people think.
The overall PageSpeed score, oddly enough. The score is a weighted aggregate that includes Core Web Vitals plus a number of supplementary measurements that are useful for tuning but not for ranking. A shop with a 70 score that nails the three vitals will, in practice, outrank a shop with a 95 score that does not.
The Goblin says this because shopkeepers spend an inordinate amount of money chasing the last fifteen points of a score that nobody is grading them on.
The disappointing role of your apps.
The single most common reason a small shop is slow is the cumulative weight of installed apps. Each app adds JavaScript to the page. Each piece of JavaScript must be downloaded, parsed, and executed before the page is interactive. The page-speed cost of any one app is usually small; the cost of twelve apps is usually catastrophic.
The Goblin’s rule: if you cannot, off the top of your head, name what an installed app does, uninstall it. If the shop still works the next day, you have your answer. He has watched shops shed three or four apps and gain a full second of LCP in the process.
The unglamorous fixes that actually work.
Four, in descending order of impact.
- Compress your photographs before uploading. Most product photographs uploaded to Shopify are between five and ten megabytes. They should be between two hundred and five hundred kilobytes. The internet has tools for this; many themes will do it automatically; the Shopify CDN will help; but nothing helps as much as compressing the file before it ever reaches the server.
- Uninstall the apps you do not use. See above. The Goblin is unmoved by your defence of any specific app.
- Use the theme’s built-in fonts where you can. Custom fonts are lovely, slow, and not worth the cost on every page of a small shop. If you must have a custom font, host it locally and load it with
font-display: swap. - Lazy-load images below the fold. Most themes do this by default. Verify yours does. If it does not, the documentation will tell you how.
That is the whole list. The Goblin will, if commissioned, name the specific images on your shop that are oversized, and the specific apps he believes are doing nothing.
Site speed is not a mystery, and it is not a contest. The shopkeeper who fixes the three measurements above and stops there is doing better than the shopkeeper who pays for an “optimisation service” that promises to chase the last five points of a score nobody is grading.
— The Goblin, who would prefer not to be timed.