It says “published”. Is it really?
What saves, what slows down, what breaks. Every finding is measured.
Headless Chrome · Measured at 4 widths · Remediation plan
What changes.
Context
The site was online and, from the outside, looked finished.
What I learned
That the worst failure is not the one that breaks, it is the one that lies.
How it was built.Decisions, evidence and details
Context
The site was online and, from the outside, looked finished. The audit was not there to redesign it: it was there to measure how well it does its job for someone opening it on a phone over mobile data, which is how most people open it.
Findings
- The admin panel saved changes in the browser itself and answered «published»: whoever used it believed they were updating the site, and they were not.
- 858 MB of PDF files served to the public, with downloads of up to 6.9 minutes over a 4G connection.
- A 4.67 MB homepage and a main content load time of between 6 and 9 seconds.
Recommendation
- A first phase of roughly one day of work, prioritised by impact over effort.
- Homepage from 4.67 MB down to around 0.9 MB.
- Main content load time from 6-9 seconds down to a range of 1.5 to 2.5 seconds.
What I learned
That the worst failure is not the one that breaks, it is the one that lies. A panel confirming a publication that never happened destroys trust in the whole system, and nobody reports it as a bug because on screen everything went fine.