I’m wondering if anyone has links to papers or other sources of information that might help in understanding time targets (benchmarks or rules of thumb) for client perceived metrics like time to render first pixel, time to render above the fold, and time to first interaction ( among others). Ideally this would be broken out by industry.
I’m interested to understand what’s great, good, normal, poor and the data associated to make these types of qualifications. I know faster has to be better, but how good is good enough and how much does 100ms matter?
It has been a little while but I have some aggregate distributions in my performance calendar post from 2010 - http://calendar.perfplanet.com/2010/the-state-of-web-performance-optimization-2010/
As to how much 100ms matters, that is going to be VERY specific to an industry (if not site). If you have the means to do an A/B test that would be your best bet. If you use something like Google Analytics you can also create some advanced segments based on performance and look at your business metrics but there’s a little more hand-waving there and you may not see something as tight as 100ms.