It's intuitively true that not loading assets when they aren't needed will be beneficial to users with unstable/low-bandwidth connections. However, we should be able to demonstrate this.
We have https://grafana.wikimedia.org/dashboard/db/mobile-2g with which we can see there are 1.355MiB of images in the Barack Obama page and track first render time on this page via webpagetest.
- What part of the story is missing from the dashboard?
- What do we need to capture and measure in order to demonstrate benefit of this work
- Do we have existing instrumentation to help us measure the impact, if any, of this work? Is it possible to add any?
- Who will do this? Performance team or reading?