Image Maintenance: WebP, AVIF and Faster Shop Load Time

Images are the quiet revenue killer in shop operation. They do not stand out like a server outage; they just make the page a little heavier, piece by piece. The numbers from the HTTP Archive prove it impressively: on the median mobile page, around 911 KB (Web Almanac 2025 / HTTP Archive) goes to images – by far the single largest resource type, heavier than JavaScript, CSS and fonts combined. The overall median mobile page weight has climbed to around 2.56 MB (Web Almanac 2025 / HTTP Archive), roughly tripling compared with a decade ago. So anyone who wants to keep load time under control has to start with images – no other area puts so much weight on the scale at so little risk.

It is not the individual file that is the problem, but the sheer number. The median image on the web weighs only around 12 KB (Web Almanac 2025 / HTTP Archive) – completely harmless. Yet a median mobile page loads around 15 images (Web Almanac 2025 / HTTP Archive), and the 75th percentile of the largest image per page already sits at 404 KB (Web Almanac 2025 / HTTP Archive). That means every fourth page has at least one motif that alone swallows almost half a megabyte – typically the large hero or product image that users see first. This exact image determines the perceived load time. Without ongoing performance monitoring, such an outlier often goes unnoticed for months while the bounce rate quietly rises.

The most common reason a once-optimised shop becomes slow again is called format drift. It refers to the gradual return of uncompressed images into daily operation: an editor uploads a product photo straight from the camera – 4,000 pixels wide and several megabytes heavy. The shop scales it down in the display but still delivers the enormous original file. A marketing agency sends campaign banners as an uncompressed PNG. A new plugin brings its own images that no one has checked. Every single case looks harmless; together, the load time slowly tips into the red.

Format drift is not a technical but an organisational problem – and that is exactly why one-off optimisations do not solve it. As long as people upload images, it will happen. The answer is an automated image pipeline that intercepts every incoming image, scales it down, compresses it and converts it into modern formats before it is ever delivered to a visitor. This turns a question of discipline into a question of technology. As with ongoing broken-link and redirect maintenance, the rule holds: what is not continuously monitored decays again on its own in operation.

Small files alone are not enough – what also matters is when and in what size an image loads. Lazy loading ensures images are only loaded once they move into the visible area. The native loading='lazy' attribute is available today in all current browsers and needs no additional library. On a long category page with many products, this saves enormously: instead of all fifteen or more images, the browser first loads only the ones actually visible. The exception is the large image above the fold – this one should specifically not be delayed, because it directly determines perceived load time.

Responsive images via the srcset and sizes attributes ensure each device receives exactly the right size. A smartphone with a 400-pixel-wide display then loads a small variant, a large desktop monitor the large one – instead of everyone getting the same oversized file. Combined with modern formats, this delivers the maximum effect: the right format in the right size at the right time. This technique is part of the open web standards (W3C) and is understood by all relevant browsers. Our monitoring services continuously check whether these mechanisms actually take effect in operation and survive updates.

So that image quality is not left to chance, it belongs in the maintenance contract as a fixed component. In concrete terms: a defined image pipeline with agreed upper limits for file size and pixel width, a regular check on whether new uploads and plugins respect those limits, and ongoing monitoring of LCP and page weight. That turns a one-off optimisation into a lasting state. Our maintenance services bundle exactly these components – from the initial setup of the pipeline through the bulk conversion of the legacy stock to the monthly review of the metrics.

The economic benefit is immediate: a faster page improves the user experience, reduces abandonment and has a positive effect on search visibility. At the same time, an automated pipeline relieves the editorial team, who no longer have to worry about formats and compression. Which services are sensibly bundled in such a contract is shown by the overview of maintenance packages and pricing. Anyone who combines image maintenance with further operational topics such as legal texts kept up to date and database and media upkeep keeps the entire shop in shape over time – technically, legally and in terms of speed.