A used motorcycle page can carry forty forecourt photos, a lender’s finance calculator, a part-exchange form, a stock feed from the dealer system and a walkaround video. A parts site adds a make, model and year finder that nearly every visitor touches. Riding gear pages bring 360 spins of helmets and size charts. Each feature earns its place, but each one costs time on a phone.

This guide explains how Google scores that experience and where motorcycle sites tend to lose it. The fixes keep the features and change how they load.

How Google scores a dealer site

There are three Core Web Vitals. Largest Contentful Paint (LCP) asks how soon the main photo of a bike or helmet appears: 2.5 seconds or less is good, over 4 seconds poor. Interaction to Next Paint (INP) asks how quickly a part finder or gallery button responds: 200 milliseconds or less is good, over 500 poor. Cumulative Layout Shift (CLS) asks how much a stock page jumps about: 0.1 or less is good, over 0.25 poor.

Each is judged at the 75th percentile of real visits, with mobile and desktop assessed separately. INP replaced First Input Delay in 2024.

Real-visit data comes from the Chrome UX Report and appears in PageSpeed Insights and Search Console. Lab tools such as Lighthouse and Chrome DevTools test a single load, which is how you find the cause of a problem. Lighthouse cannot measure INP; it shows Total Blocking Time as a proxy.

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

On 23 October 2026 I tested five product pages from UK motorcycle retailers in PageSpeed Insights. Each had one mobile run, emulating a mid-range phone on slow 4G.

PagePerformance scoreLab LCPTotal Blocking TimeLab CLS
Retailer A6313.9 s160 ms0.116
Retailer B5531.2 s110 ms0.000
Retailer C468.7 s530 ms0.001
Retailer D3819.7 s1,030 ms0.000
Retailer E4022.3 s590 ms0.000

Here the issue is loading rather than blocking. The median lab LCP was 19.7 s, against 12.8 s across all 103 valid pages in my wider test. Four of the five took over 10 s. Two pages needed more than 11 s just for their first content to appear. Lighthouse flagged images without set width and height on all five, which fits the forecourt photo problem below, and late LCP discovery on three. The median score was 46 and the median blocking time a comparatively modest 530 ms.

Two pages had Chrome UX Report data for their own URL. One failed Core Web Vitals with a field INP of 656 ms, past the 500 ms “poor” line. The other had too little data for a verdict. Of the three with whole-site (origin) data only, two passed. A throttled lab phone runs much slower than most real visits, and Google does not assess lab LCP. Five pages is a small sample rather than a ranking of dealers.

Forecourt photos and the used bike gallery (LCP)

Used stock photos are often taken on a phone or camera on the forecourt and uploaded straight to the site. Unless the platform resizes them, a buyer on mobile may download a file many times larger than the screen needs.

The Web Almanac 2024 performance chapter gives a sense of how common this is across the web. In 2024 data covering all kinds of site, 8% of mobile LCP images were over 1,000 KB. That figure is not specific to motorcycle dealers, but large unedited uploads are an easy way to join that group.

Google’s guide to optimising LCP sets out the fixes:

  • Resize on upload. Generate a set of widths and serve them through srcset, in WebP or AVIF.
  • Load the first photo eagerly and give it fetchpriority="high". Never lazy-load it. Lazy-load the other thirty-nine.
  • Do not hide it behind a slider script. Google warns against keeping the LCP element hidden until JavaScript has loaded, which some gallery sliders do with a fade-in.

Stock feeds and pages built by script

Many dealers display stock through a feed or widget supplied by their dealer management system. Sometimes the whole listing, including the first photo, is built in the browser by JavaScript, or sits inside an iframe served from another domain.

The Web Almanac found that 35% of mobile LCP elements in 2024 could not be found in the initial HTML. When the main photo arrives that way, the browser cannot start downloading it until the script has run. web.dev’s LCP guidance recommends server-side rendering or prerendering over client-side rendering. Where an image must be referenced from script, a <link rel="preload" as="image" fetchpriority="high"> in the page head lets the download start early.

Serving the photos from the stock system’s own domain adds another cost. Google’s advice is to serve critical resources from the same origin as the page. If you can, have the feed write stock into your own pages and image storage, with a CDN in front.

Finance calculators and layout shift (CLS)

Finance calculators are often third-party iframes that resize once they load. Google’s CLS guidance lists widgets that resize among its causes of layout shift, alongside images without dimensions, web fonts, inserted content and content loaded asynchronously.

On a dealer page the effect is familiar: the buyer starts reading the specification, the calculator expands, and the text drops down the screen.

  • Give the calculator a fixed height that matches its loaded size, or show a static representative example and load the calculator when the buyer taps it.
  • Set width and height on every stock photo. The Web Almanac found that 66% of mobile pages in 2024 had at least one image without set dimensions, with a median of two per page.
  • Reserve space for part-exchange forms, review stars, warranty banners and cookie notices.

Content that moves within 500 ms of a tap, click or key press does not count. A calculator the buyer opens is fine. One that grows two seconds after the page loads is not.

Part finders and parts search (INP)

INP measures the delay after clicks, taps and key presses throughout the visit, usually reporting the slowest. A make, model and year finder is a series of taps, and search-as-you-type is a series of key presses. Scrolling through results does not count.

Part finders can be slow when each choice triggers a lot of work. Some download the full model list for every make, or filter thousands of parts in the browser after each selection. web.dev divides an interaction into input delay, processing time and presentation delay, and links to guidance on long tasks, web workers and large page structures.

  • Show the selected make or model at once, then fetch the next list.
  • Request only the models for the chosen make, not the whole range.
  • Filter large parts lists on the server, or move the work to a web worker.
  • Wait for a pause in typing before searching, and limit the first set of suggestions.
  • Render results in pages rather than building thousands of rows at once.

Walkaround videos, 360 helmet spins and configurators

Walkaround videos and exhaust sound clips sell bikes. A video can be the LCP element, timed from when its poster loads or its first frame appears, whichever is sooner. Give each video a small, compressed poster image. For embedded players from video sites, show a still with a play button and load the player when the buyer taps it.

A helmet spin is typically built from dozens of photos. Load the first frame with the page and the rest only when the buyer starts to rotate it.

A 3D motorcycle configurator deserves the same treatment: show a still image first and load the model on request. You can try this in the motorcycle configurator demos.

Testing a motorcycle site

  1. In Search Console’s Core Web Vitals report, compare used stock pages, parts pages and gear pages. They may fail for different reasons.
  2. Run one page of each type through PageSpeed Insights, starting with the field data section.
  3. View the source of a used bike page. Is the first photo’s address in the HTML, or does it only appear after scripts run?
  4. Download the first photo and check its file size and pixel width against the space it fills on a phone.
  5. Load a page with a finance calculator and watch the specification text. If it moves, reserve the calculator’s space.
  6. On a mid-range Android phone, pick a make, model and year in the part finder and note any pause after each choice.

None of this changes how the site looks. If you would like it done, see Core Web Vitals, WooCommerce speed, caching and scripts or Shopify speed optimisation. The rest of the series covers selling motorcycles online and motorcycle SEO.

Sources