LCP (Largest Contentful Paint) includes Background Images.
That means if your web design includes a GB Img (not an image for content!), then it impacts the LCP, and thus your CWV score.
Worse - the potential negative of that is far greater than Cumulative Layout Shift (CLS)
People wonder why I give G a hard time, and why I doubt their judgement in regards to metrics and figures.
THIS is a shining example of Why!
Why the hell should the usage of a background image negatively impact your performance score?
More importantly, HOW is that worse than things jumping around on the page?
Surely text/buttons/links moving and jostling is a poorer experience than when a background image renders?
What about interruptions and content-blockers?
Push requests, modals etc.?
(coming soon?)
Again, I would have thought Those were more hurtful to a UX than a flaming background image!!!
So ... what options do you have?
Well, the first is to cut your background image.
Switch to a colour/gradient etc.
Alternatively, use Preload and compress the shit out of that image!
Unfortunately, due to the current weighting - most optimisations are unlikely to balance as they really should (good job G!).
So, we may need to cheat!
Use a a single colour Gif bg image (same/larger dimensions than your actual BG Img) - should come in at less than 2Kb.
Then use a little JS to change the BG img on window load to change to the one you want.
TaDa!
Your BG Img now has only a tiny footprint.
Doesn't fix G's stupidity, but will help reduce the impact that your big, pretty background image has on "user experience" ;)
Another option to consider is more technical, but look at connection type data, and serve different images based on visitors network/bandwidth type etc.
A major PITA, but best for users.
• • •
Missing some Tweet in this thread? You can try to
force a refresh