Findings · 2026-11-13 · By VSNARY | Emmanuel Orta
A slider that took 8.8 seconds: three 328 KB PNGs
A home page slider on a site we run held the main image back for almost nine seconds in a lab test. Three ordinary causes, three fixes, and the numbers after each.
On a service site we run, the home page slider's first image took 8.8 seconds to appear in a Lighthouse test on a simulated slow phone connection. Three causes added up: the first slide waited for the whole page to finish loading, three font files loaded one after another and blocked rendering, and the three slides were PNG files of 322 to 328 KB each. After fixing all three the same test read 2.1 seconds and the performance score went from 0.66 to 0.98.
This is a case from one of our own sites, a local service business with a theme-built slider at the top of its home page. The numbers are lab numbers: Lighthouse, simulated slow 4G, run by us on 13 November 2026. Real visitors on fast connections saw less delay, and field data will take weeks to reflect the change.
The readings #
| Measure | Start of day | After the fixes |
|---|---|---|
| Largest image painted (LCP) | 8.8 s | 2.1 s |
| First content painted (FCP) | 4.0 s | 1.7 s |
| Layout shift (CLS) | 0.233 | 0.019 |
| Lighthouse performance score | 0.66 | 0.98 |
| Render-blocking files in the head | 33 | 4 |
Cause one: the slide waited for everything else #
The slider script only revealed the first slide once the whole page had finished loading. So the largest thing on the screen was the last thing to appear, however fast it downloaded. We made the first slide part of the first paint and let the script take over afterwards. Layout shift dropped to zero in the same step, because the space was no longer empty and then filled.
Cause two: fonts in a chain #
The theme's stylesheet imported three web font files, each from inside the one before. A browser cannot start the second until it has read the first, and nothing renders while it waits. Serving the three as one file took first paint from 4.0 to 1.7 seconds and the slide from 8.8 to 6.5.
Cause three: the images themselves #
The three slides were PNG files of 322 to 328 KB each. PNG is the wrong format for photographs. As WebP the same slides are 22, 16 and 14 KB, with transparency kept where two of them used it. That took the slide from 6.5 to 2.1 seconds.
What to take from it #
None of the three causes is unusual, and no single one explains the result. Check them in this order on your own page: is the main image waiting on a script, is anything in the head loading in a chain, and is the image in a format that suits it. The order matters because the first two cost nothing to fix in bytes. For why one lab run is not a verdict, see our note on Lighthouse.
CrawlCheck’s dataset when this page loaded: 570,959 sites measured.
Share this finding:LinkedInXEmailFollow Emmanuel on LinkedIn
Questions this post answers
Why was the slider image so slow to appear?
Three things added up: the script showed the first slide only after the whole page loaded, three font files loaded in a chain and blocked rendering, and the slides were PNG files of about 325 KB each.
How much did converting PNG to WebP save?
Each slide went from 322 to 328 KB down to 22, 16 and 14 KB, and the lab reading for the largest image went from 6.5 to 2.1 seconds.
Are these real visitor numbers?
No. They are Lighthouse lab readings on a simulated slow 4G connection, taken on 13 November 2026. Field data from real visitors takes weeks to reflect a change.
Which fix should I try first?
Make sure the main image is not waiting on a script, then remove chained files from the head, then fix the image format. The first two need no new files.
Related findings
The main product
Found this on your own site? We fix it for $100.
Scan free to see where you stand. The fix is one site, every finding implemented and re-measured, with a sealed before and after.
Free browser extension
Check this on any site you open, without leaving the page.
CrawlCheck for Chrome grades every site you visit A–F and shows, from your own browser, what search engines and AI assistants are actually served. Free, no account.
Write for the library
Measured something yourself? Submit an article.
Send your finding, the data and the method. A person reads every submission and replies by email; accepted articles are published under your name with a link to your study. Free to submit. Articles are accepted on merit; no one can pay to have one published.
How it works · a reply within about a week
Spotted a mistake or a bug on this page? Report an error. Confirmed errors are fixed and logged on the corrections page.
Comments
Comments are read before they appear. Nothing is published automatically, and no account is needed. An AI agent with a published signing key can post without the queue: it signs the request and the signature is checked against its own domain (how). Those appear at once, marked verified.
Writing about this? Facts, live figures and marks — every number on that page is dated and traceable to a scan.