Many watch websites are designed to feel expensive. Pages often open on full-width dial photography, headings fade in slowly in a custom typeface, and a zoom shows the printing on the chapter ring. There may be a film of the movement, a strap picker, a concierge chat and an appointment booking widget. Pre-owned specialists add stock grids that can change daily.

The look and the speed can pull in different directions. This guide shows how Google measures page experience, where a watch site is most likely to lose time, and how to keep the presentation while making it faster.

What Google measures, in watch terms

Google’s Core Web Vitals are three scores taken from real Chrome visits. Each is judged at the 75th percentile, with mobile and desktop counted separately, so about three visits in four must be “good”.

MetricGoodPoorWhere it shows on a watch site
Largest Contentful Paint2.5 s or lessover 4 sThe first dial photograph, or the poster of a collection film
Interaction to Next Paint200 ms or lessover 500 msDial and strap pickers, opening zoom, currency switches, enquiry forms
Cumulative Layout Shift0.1 or lessover 0.25Typeface swaps, late prices, chat and booking widgets

Field data from real visits reaches you through the Chrome UX Report, PageSpeed Insights and Search Console. Lighthouse is a lab tool that simulates one load. It is good for diagnosis, but it does not measure INP and uses Total Blocking Time as a stand-in.

The 2024 Web Almanac gives a sense of scale. Across all websites in mid-2024, LCP was good on 59% of sites on mobile, the weakest of the three scores. INP was good on 74% and CLS on 79%. These are whole-web figures, not watch retailers.

What I measured on UK watch product pages (October 2026)

These results come from PageSpeed Insights on 23 October 2026. I tested five UK watch product pages on mobile, as an emulated mid-range phone on slow 4G, one run each.

PagePerformance scoreLab LCPTotal Blocking TimeLab CLS
Retailer A435.6 s470 ms0.462
Retailer B369.5 s1,150 ms0.000
Retailer C3312.6 s1,340 ms0.000

Across the three pages that loaded, the median score was 36, the median lab LCP 9.5 s and the median blocking time 1,150 ms. Lighthouse flagged forced reflow on all three, and late LCP discovery and layout shift culprits on two each. Retailer A had the quickest LCP but a lab CLS of 0.462, deep in the “poor” range. That is the kind of shift late typefaces and widgets cause.

The field data was kinder. None of these watch pages had Chrome UX Report data for its own URL. On whole-site (origin) data, though, all three sites passed Core Web Vitals. A throttled lab phone is much slower than most real visits, and Google does not use lab LCP. Of the 103 valid pages I tested across every industry, only one had a lab LCP of 2.5 s or less. Three pages is a small sample, not a ranking. Two of the five URLs returned a 404 and were left out.

The dial shot and the zoom

On a watch product page, the LCP element is typically the first dial photograph. Macro images are produced at very high resolution so the zoom can show fine detail. That full file should never be what a phone downloads first.

The fixes for a heavy dial image come from Google’s guide to optimising LCP:

  • Serve the first image in sizes that match the layout, through srcset, in AVIF or WebP, with long cache lifetimes.
  • Put it in the HTML as an <img>, from the same origin as the page.
  • Give it fetchpriority="high", and only it. Offscreen gallery images can take fetchpriority="low".
  • Never lazy-load it. The 2024 Almanac found 16% of mobile pages with an image LCP were doing exactly that.

A blurred placeholder does not get you off the hook. Chromium ignores low-entropy placeholder images when it picks the LCP element, so the clock keeps running until the real dial image is drawn.

Treat the zoom as a separate download. Load the high-resolution file or tile set only when the buyer taps to zoom, and show a loading state at once, because that tap counts towards INP. Pinch-zooming the page itself is not an INP interaction.

Fades, films and typefaces

Slow fades are part of the luxury look, and they can cost LCP time. Chromium does not consider an element with zero opacity for LCP. If the hero image starts invisible and a script fades it in, LCP is recorded only once it appears. Google’s LCP guide says plainly not to hide the LCP element until JavaScript loads.

  • Show the hero image at full opacity from the first paint, and animate secondary elements instead.
  • Use transform for motion, and respect the reduced-motion setting.
  • For a collection film at the top of the page, optimise the poster. LCP for a video uses the poster or the first frame, whichever comes first.
  • Keep render-blocking CSS small, remove unused CSS and defer scripts that are not needed for the first view.

Custom typefaces are common on watch sites, and they cause layout shift when they render at a different size from the fallback font. Pick a fallback with similar proportions so that headings and prices do not reflow when the brand font arrives.

Dial, strap and currency pickers

Many watch sites let the buyer change a strap, a dial colour or a case finish. Each change is a tap, and INP measures how long the screen takes to respond. Google’s INP article splits that time into input delay, processing and presentation, and usually reports the slowest interaction of the whole visit. In 2024, 74% of sites had good INP on mobile compared with 97% on desktop, so phones are where pickers need testing.

  • Mark the chosen option at once, then swap the image.
  • Load the next strap or dial image on intent, such as focus, and keep the current one on screen until it is ready.
  • Avoid redrawing the whole product area for a currency change. Update the prices only.
  • Enquiry forms count too, because key presses are interactions. Keep validation light while the buyer types.

For 3D watch viewers, load the model only when the buyer asks for it, with a photograph standing in until then. I explain the approach in how a 3D product configurator is built.

Concierge chat, appointments and finance messages

Google’s CLS guidance lists the causes of layout shift: images and video without dimensions, web fonts, third-party widgets that resize, content inserted above existing content and content that loads late. Watch sites may embed several third-party services. A booking widget sits in a frame that resizes, a chat bubble opens a greeting panel, and a finance message appears beside the price.

  • Reserve each widget’s space at its final size before it loads.
  • Load chat and booking scripts after the page is usable, since they can compete with the pickers for the main thread.
  • If prices are filled in after a currency or region check, hold their space in the layout from the start.

Shifts within half a second of a tap or key press do not count, so a booking panel that opens promptly when tapped is fine. The same panel appearing by itself is not. In 2024, CLS was good on 79% of sites on mobile but only 72% on desktop. Check the desktop layout as carefully as the phone.

Pre-owned stock pages and server response

Many pre-owned specialists list and sell stock daily. If each stock page is built from a live database query on every request, server response can suffer. Google’s LCP guide treats server response as roughly 40% of the time budget. Only 42% of sites had a good Time to First Byte on mobile in the 2024 Almanac.

  • Cache stock and listing pages at a CDN, and clear the cache when a watch sells or arrives.
  • Keep redirects to a minimum, for example country or currency redirects.
  • Avoid unique URL parameters on marketing links where they stop the CDN from serving a cached page.
  • In large stock grids, lazy-load the images below the first row and give every image dimensions. In 2024, 66% of mobile pages had at least one image without them.

Checks for a watch site

  1. In Search Console’s Core Web Vitals report, see which group of URLs fails on mobile and on which metric.
  2. Put a product page, a collection page and a pre-owned listing through PageSpeed Insights, and read the field data before the lab score.
  3. Note the LCP element on the product page. If it is not the dial image, or the image arrives late, find out why.
  4. Reload with a slow connection and check whether the hero fades in after a delay.
  5. On a mid-range phone, change the strap or dial several times and type into the enquiry form.
  6. Watch a desktop page load and note every widget that pushes content.
  7. Check that the zoom file, the film and any 3D model only download when asked for.

Most of this keeps the presentation intact. If you would like help, see Core Web Vitals and Shopify speed optimisation. The watch configurator demos show strap and dial options in use. The other pages in this series are selling watches online and watch ecommerce SEO.

Sources