Skip to content
SectionPerformance and measurement
Reviewed2026-08-01
Words1,341
Sources4
Performance and measurement

Images, fonts and page weight on clinic sites

Where clinic page weight actually comes from, and the specific changes that reduce it without touching the design.

Short answer

On most clinic sites, images and web fonts account for the majority of page weight. Images should be served in a modern format, at the dimensions they are displayed, with width and height set, lazy loaded below the fold and given high priority above it. Fonts should be limited to two families and the weights actually used, preloaded and served with a display strategy that avoids invisible text.

Page weight is not an abstract quality score. It is the number of bytes a patient's phone has to receive on a mobile connection before the page is usable, and on clinic sites it is dominated by two things: photographs and typefaces.

Where image weight comes from

The typical failure is a photograph uploaded at its original camera dimensions, resized by CSS to fit a container, and served in full to every device. A four megabyte original displayed at 800 pixels wide has wasted almost all of what it transferred.

  • Serve at display size. Generate the sizes you actually use and serve the right one with srcset and sizes.
  • Use a modern format. WebP or AVIF with a JPEG fallback where necessary. Google's image documentation covers the practical guidance.
  • Always set width and height. This reserves space and prevents layout shift, and it costs two attributes.
  • Lazy load below the fold, never above it. Lazy loading the hero delays the metric you were trying to improve.
  • Compress deliberately. Most photographs are visually indistinguishable at a quality setting well below the default.

Clinical imagery, handled properly

Before and after photographs and other clinical images carry regulatory requirements as well as technical ones. Treat the regulatory question first: whether the image may be published at all, with what consent and with what accompanying information. Only then optimise the file. Consent for clinical photography is a clinical record matter, not a marketing one, and should be documented as such.

Fonts

A clinic site commonly loads two or three families at five or six weights, several of which appear nowhere on the site. Each weight is a separate file and each is a request on the critical path.

  1. Audit which families and weights are genuinely used, then delete the rest from the load.
  2. Two families is enough for any site. Three is a decision that needs defending.
  3. Preload the fonts used above the fold.
  4. Use a font display strategy that shows fallback text immediately rather than leaving text invisible while the font loads.
  5. Choose a fallback with similar metrics so the swap does not move the layout.
  6. Self-host where it simplifies the connection sequence, or preconnect where you do not.
Invisible text is a real cost

A page that renders no text until a font arrives is a page that is blank for the first seconds of a mobile visit. Patients leave during that window, and the metric that captures it is the one you were trying to improve. Fallback text visible immediately is nearly always the right trade.

Video and embeds

An auto-playing background video on a clinic homepage is the single most expensive design decision available. If video is required, use a poster image, load the player only on interaction, and never autoplay a large file on mobile. The same applies to map embeds, booking widgets and review carousels: load them when they are needed, not before.

Third-party weight

Third-party scripts often exceed the site's own code. Build a list: chat, analytics, tag management, consent, booking, reviews, heatmaps, advertising pixels and anything else. For each, record what it does, who reads its output and what it costs. Anything with no owner comes off, and most clinic sites have several.

SURFACE 08

Your own website

HIGH CONTROL
What controls it
  • Every byte a crawler receives, and the speed it arrives at
  • The entity claims: name, address, identifiers, sameAs, service list
  • Canonical URLs, hreflang if used, and the internal link graph
  • Whether the page answers the question it was built to answer
What does not
  • How the answer is displayed once it leaves your server
  • Whether a third party republishes an outdated version of your facts
How to test it
  • Crawl the whole site and compare the URL list against the sitemap
  • Validate every JSON-LD block against the Rich Results Test
  • Measure field data, not only lab scores, for Core Web Vitals

The site's own code

Page builders generate a great deal of markup and stylesheet for a small amount of visible content. The realistic clinic response is not a rebuild but a settings audit: turn off the modules and animation libraries the site does not use, and stop loading assets on pages that do not contain the components they belong to.

Setting a budget

A budget makes the decision explicit before it is made badly. A workable one for a clinic page: under one megabyte transferred for the initial view on mobile, under one hundred kilobytes of JavaScript executed before interaction, two font families with at most four weights, and one hero image under two hundred kilobytes. These are starting values to argue with, not published standards, and having any budget is better than the alternative.

Testing what patients actually get

Test at 390 pixels wide with the network throttled, and on a real mid-range device if you have one. Record the transferred bytes, the number of requests and the time until the main content is visible. Repeat after each change. A single number per page, tracked over time, keeps performance visible rather than periodic.

Alt text is not a performance issue, and matters anyway

Every meaningful image needs alternative text describing what it shows, for screen reader users and for anyone whose images fail to load. Decorative images should carry an empty alt attribute so they are skipped. WCAG 2.2 sets the standard, and for a healthcare provider this is part of providing an accessible service rather than an optimisation.

Where the weight comes back

Page weight is not fixed once. It creeps back through ordinary content editing: a new hero uploaded at full resolution, a testimonial slider added by a plugin, a tracking pixel added for a campaign that ended a year ago. Every clinic site that was optimised eighteen months ago and never checked since is heavier now than when the work finished.

The counter is a monthly measurement rather than a project. Record transferred bytes and request count for three pages: the homepage, the busiest treatment page and a location page. Three numbers, once a month, in the same document as everything else you track. When one of them jumps, you know within four weeks rather than at the next audit.

Fixing it at the point of upload

The most reliable fix is upstream of the site. Whoever adds content needs a resizing step in their routine and a maximum file size they are asked not to exceed. Most content management systems can generate the sizes automatically, and configuring that once removes the problem permanently for everyone who edits the site afterwards, including the people who will never read this article.

No commercial links on this page

This article contains no affiliate links, no sponsored placements and no links to any agency, supplier, clinic or commercial brand. Nobody paid for it, nobody previewed it and no directory advertiser had sight of it. This publication does not rank or recommend agencies anywhere on the site.

Nothing here is medical or legal advice. Regulatory material summarises published guidance. Read the source before relying on it. Our full position is in the editorial standards.

Sources

Primary documentation and regulators only. We do not cite opinion surveys as though they were measurements of a system.

Frequently asked questions

What image format should a clinic site use?

A modern format such as WebP or AVIF for photographs, at the dimensions displayed, with a fallback where support requires it. The format matters less than serving the right size.

Should we lazy load all images?

No. Lazy load below the fold only. The hero image should load eagerly with high priority, because it is usually the element the loading metric measures.

How many web fonts is too many?

Two families covers any clinic site. Beyond that each additional family and weight is a request on the critical path, and most sites load weights that appear nowhere in the design.

Is a background video ever justified?

Rarely, and never autoplaying at full size on mobile. If it is required, use a poster image and load the player on interaction.

Does image alt text help search?

It describes the image for people who cannot see it, which is its purpose, and it gives search engines context for image results. Write it as a description, not as a place to repeat a keyword.

The surfaces move. We write when they do.

A short briefing on documented changes to the surfaces clinics are discovered on. One labelled sponsor slot per issue, no rankings, no invented numbers.