Performance profile · Jewellers
Jewellers: macro detail without the page weight
Customers zoom into a setting on a phone, so you need detail and low weight at the same time — and the usual answer, uploading bigger files, makes the page slower for everyone who never zooms.
The short answer
On a jeweller's site the controlling number is how long a product page takes to paint its main image, because that image is the product and a customer judges the piece and the shop in the same second.
Measure it on the product page, throttled to a mobile connection, and read TTFB separately from render time. A cached product page on our platform serves a UK visitor comfortably below 200ms, so anything slower than that is almost always the images rather than the server. The fix is WebP output, dimensions matched to their containers, and a zoom that fetches its large file only when someone asks for it — on hosting from $2.42/mo.
By the HostingFast team · Reviewed 12 August 2026
24/7
Engineers on shift, any hour
1-click
One-click WordPress
Free
Certificates at no charge
Daily
Daily backup cadence
There is a real tension in a jewellery site that most retail does not have. Detail sells the piece — the setting, the hallmark, the finish of a claw — so the photography has to survive being zoomed. But the same file that rewards a zoom punishes everyone who never zooms, and most visitors never do.
Handled badly, that tension produces a product page carrying twelve 4000-pixel JPEGs, all fetched before the page settles, on a phone on mobile data. Handled properly it produces a page that paints quickly and requests the large file only on demand.
Which of those you have is a measurement, not an opinion. Here is how to take it, and what to do with the answer.
Measure the product page on a throttled connection
Test a product page, not the homepage, and throttle it. PageSpeed Insights runs a mobile profile by default and reports the field data from real Chrome visitors, which is the version that counts. WebPageTest gives you the waterfall, and on a jewellery site the waterfall almost always shows the main image finishing long after the rest of the page is ready.
Read TTFB on its own first. If a cached product page answers a nearby visitor under 200ms and the page still feels slow, you have a payload problem and no hosting upgrade will touch it. If TTFB is sitting above roughly 600ms, stop looking at images and look at caching, PHP version and the database.
Serve the right file to the right screen
One image per product is the mistake. Use responsive sources so a phone downloads a phone-sized file, let the platform convert to WebP, and match the delivered dimensions to the container rather than scaling a 4000-pixel master in CSS. That single change usually halves the weight of a jewellery product page.
Keep the zoom, but make it lazy. A viewer that fetches its high-resolution file on click gives a serious buyer the detail they want without charging the other ninety-nine visitors for it. Lazy-load everything below the fold too, including the related-products carousel.
The checkout is a different machine
Cart, checkout and account pages exclude themselves from the cache, necessarily so. For a high-value transaction that matters more than usual: a buyer entering a four-figure card payment has already decided, and a stalled payment step is the one moment where a won sale is lost.
That request is raw PHP and database work. PHP 8.x, Redis object caching and NVMe storage are what move it. SSL is not theatre either — the certificate is included on every plan and reissues itself, so the padlock a nervous buyer checks for never quietly lapses.
Verify, do not assume
After each change, retest. Three lab runs, keep the median, and expect field data to lag by weeks. Check that the cache is genuinely serving: 'x-litespeed-cache: hit' in the response headers on a second load is the evidence, and a permanent miss usually means a plugin cookie is marking every visitor as dynamic.
Do the work on a staging clone first. Image plugins are the classic case of a change that improves a benchmark and breaks a gallery, and finding that out on the live shop during a commission enquiry is an expensive way to learn it.
Detail on demand, weight only when asked
Independent jewellers, bespoke goldsmiths and designer-makers selling their own collections need photography that survives a zoom and a page that does not punish the visitor who never zooms.
NVMe storage, LiteSpeed caching in the server and built-in WebP conversion make that a configuration problem rather than a compromise. Every tier, including the entry plan.
- WebP conversion built into the platform, no plugin licence
- LiteSpeed caching in the web server, verifiable in the headers
- Redis object caching for the uncacheable checkout path
- Free SSL that reissues itself before it can expire
Why HostingFast
Standard on every plan
Image weight handled at the platform
WebP conversion is built in at no extra cost, so a gallery of full-resolution masters stops being the reason the product page waits.
A product page that starts fast
A cached page served from London reaches UK and Irish visitors on a 5–15ms round trip, comfortably under 200ms of TTFB.
The padlock never lapses
Free SSL on every plan, reissued automatically. The one trust signal a high-value buyer checks is the one you never have to remember.
Headroom for the commission form
Enquiry posts and checkout run uncached. PHP 8.x, Redis and NVMe are what make them quick, and they are standard rather than an upgrade.
Staging for image experiments
Clone the shop, swap the gallery plugin, measure. Push it live only once the number has moved the right way.
Upgrades without a migration
Plan changes happen in the client area with no move and no downtime, from shared accounts through VPS to dedicated hardware.
Quick Start
From order to online
- 1
Take a mobile baseline of one product page
Field vitals from PageSpeed Insights, waterfall from WebPageTest, TTFB noted separately. That is your before figure and you will need it.
- 2
Rebuild the image pipeline before anything else
Responsive sources, WebP output, container-matched dimensions, lazy-loading below the fold, zoom fetched on demand. Retest after each one.
- 3
Then measure the paths that never cache
Add to basket, reach the checkout, submit a commission enquiry. Time each. That is where PHP version, Redis and NVMe show up.
Built In
Loaded onto every plan
- WebP image conversion included at no extra cost
- LiteSpeed caching compiled into the server rather than bolted on
- NVMe SSD storage on every tier
- Redis object caching available for uncached requests
- Per-site PHP version switching from the control panel
- Staging environments for testing before you ship
- Free SSL on every plan, renewed automatically
- Daily backups with a restore you run yourself
- In-place plan upgrades, no migration when you change tier
- 99.9% uptime as the target, monitored around the clock
Frequently Asked
What people ask us most often
How large should product images be uploaded at?
Large enough for the zoom, then served small. Upload one master, let the platform generate responsive sizes and WebP versions, and match the delivered dimensions to the container rather than scaling a 4000-pixel file in CSS. The zoom viewer requests the big file only when a visitor clicks, so detail costs nothing to the majority who never do.
My product page paints slowly but the server responds instantly. What now?
That is a payload problem and no hosting change will fix it. A sub-200ms TTFB with a slow paint means the browser is waiting on bytes: images first, then fonts, then whatever the theme loads before it renders. Work the image pipeline, retest, and only then look at scripts.
Does a high-value checkout need anything different?
Not different, just measured. The checkout never caches, so it runs at whatever speed PHP and the database can manage — that is the number to watch. The trust side is already handled: SSL is included and reissues itself, so the padlock cannot silently expire on you.
The cache header says miss on every load. What causes that?
Usually a cookie. A plugin that sets one for every visitor makes each request look logged-in, and the cache correctly steps aside. The other common causes are caching never switched on in the plugin, or the test page excluding itself. The plugin's environment report normally names the culprit in one screen.
Keep reading
Insurance Brokers — Hosting Guide
Where a broker's quote form spends its time, and why it never caches.
Accountants — Hosting Guide
What an accountancy site's load time is really made of, measured.
SSL Certificate (Glossary)
The term defined, plus what it costs you the day it lapses.
WordPress Hosting — Built for Jewellers
WordPress on LiteSpeed with Redis behind it and a staging clone to test on.
Domain Names
Register or move a name — first year free with an annual plan.
Changing hosts? Run through our checklist first.
A straightforward sequence for a switch your visitors never feel: which files move first, how to shift email across without losing a single message, the right moment to repoint DNS, and the two mistakes behind almost all the downtime we get asked to rescue.
Weigh the page, then fix it.
Built-in WebP, LiteSpeed in the server and NVMe on every tier from $2.42/mo — renewing at the same figure.
View WordPress Hosting — Built for Jewellers plans