Blog
Image weight by platform: 41 stores measured
We measured image bytes on 41 real storefronts across 14 platforms. WooCommerce sent 2.7 times more bytes per pixel than Shopify.
Last reviewed:
On this page
A WooCommerce storefront in our sample sent 2.7 times more bytes per pixel than a Shopify storefront. We found that by weighing images on 41 real stores across 14 platforms.
That is a byte measurement. Not page speed. Not image quality. The sample and the limits stay beside the finding.
Why we measured instead of reading comparisons
We needed to decide which platforms should get a connector. Published comparisons gave us conflicting answers, including on whether Shopify served AVIF.
So we measured storefronts. The first result overturned the assumption in our plan: one platform described as doing little had eight ready-made widths and served WebP. Building from the comparison would have cost months.
A platform that already handles its images has no problem for our connector to solve.
What bytes per pixel measures
Bytes per pixel is the size of each delivered image divided by its pixel count, summarized as the median for a storefront. It lets us compare encoding weight without rewarding a store just for using smaller pictures.
It does not measure quality. A low number can mean efficient encoding or a damaged image. It is not a savings estimate either. We left “what NimboCDN would save” out of the table because that would require a model, not a direct measurement.
The measured table
Medians across stores. n counts stores, not images. Rows with one or two stores are shown, but they do not describe a platform.
| Platform | Stores | Bytes per pixel | Already modern format |
|---|---|---|---|
| Shopify | 12 | 0.0933 | 100% |
| VTEX | 6 | 0.0815 | 97% |
| WooCommerce | 4 | 0.2508 | 48% |
| WordPress | 3 | 0.2012 | 0% |
| Nuvemshop | 2 | 0.0894 | 47% |
| Wix | 1 | 0.1081 | 91% |
| PrestaShop | 1 | 0.1952 | 48% |
| Unrecognised | 2 | 0.6427 | 0% |
Two columns matter. Bytes per pixel shows delivered weight. “Already modern format” shows how often the platform had already moved beyond JPEG or PNG.
How the sample compares with the wider web
The HTTP Archive Web Almanac reports bits per pixel across more than ten million pages. Our figures are bytes per pixel. Multiply by eight and the units match.
| Measured on our sample | Bytes/pixel | Bits/pixel |
|---|---|---|
| WooCommerce storefronts | 0.2508 | 2.01 |
| WordPress sites | 0.2012 | 1.61 |
| Shopify storefronts | 0.0933 | 0.75 |
The Web Almanac puts the median image at 2.1 bits per pixel and the median JPEG at 2.0. WooCommerce in our sample sits at 2.01. Shopify sits at 0.75, below the report's 1.4 median for AVIF.
This does not make Shopify “faster.” It says its storefront images carried fewer bytes per pixel in this sample.
What this does not prove
- Most platform samples are small. Shopify has 12 stores, VTEX 6, and WordPress plus WooCommerce 7. A one-store row is only that store.
- We measured storefronts. Product pages, checkout, and dashboards can behave differently.
- This is not a savings estimate. It says what arrived, not what another service would remove.
- Bytes per pixel is not quality or load time. Those need different measurements.
What the result changed
It closed more platforms than it opened. All 12 Shopify stores already used modern formats. Wix, Squarespace, and Webflow both delivered reasonably and left no place for an outside connector to change image delivery.
WordPress and WooCommerce showed the clearest gap among the platforms with a usable sample, and both allow a connector. That supported the market we were already serving. It did not prove every WordPress store has the same problem.
For the practical next step, read what images can and cannot fix on a slow WooCommerce store, or rule out the other causes of a slow WordPress site.
Weigh your own storefront
Your store is not a median. Paste one page into the website image size checker. It weighs every image and encodes the heaviest ones for real. Today's number is measured. The possible after is estimated, with its error shown.
If the site runs on WordPress, install the connector. No account. No credit card. The home page starts shipping lighter right away. If you prefer to work inside the media library, compare ShortPixel and Imagify.
Questions
Why use bytes per pixel instead of kilobytes per image?
Kilobytes per image mostly reflects image dimensions. Dividing delivered bytes by pixel count removes that size difference and makes encoding weight easier to compare across stores.
Is a lower number always better?
No. A low number can come from efficient encoding or from damaging the image. Bytes per pixel measures weight, not perceived quality. Quality needs a separate comparison at a matched perceptual score.
Are seven WordPress and WooCommerce stores enough?
They are enough to report the gap we saw, with the sample attached. They are not enough to claim every WordPress or WooCommerce store behaves the same way. That is why the count stays beside the finding.
Does this mean Shopify is faster than WooCommerce?
No. We measured image bytes, not page-load time. In this sample, Shopify storefront images carried fewer bytes per pixel. Load time also depends on hosting, theme, scripts, and everything else on the page.
Can I reproduce the measurement?
Yes. The method and date are published with the numbers. It covers storefronts measured on September 3, 2026. A store that changes its theme or image settings later may produce a different result.