! Deployment branch, test database: 30683 milliseconds ! Average of 10 test runs per system ! Empty browser cache for each run, but populated symfony cache which is a reasonable assumption
Use a CDN (content delivery network) ! Compress components with gzip ! Add Expires headers ! Put JavaScript at bottom ! Use cookie-free domains VERDICT: Minimal gains, hard to test, not worthwhile!
down! ! Removing the intermediary page will save time by eliminating the following: ! JavaScript redirect time ! Processing required by Symfony for the redirect page ! HTTP request/response for the redirect page and assets ! Tweaking the minor issues reported by YSlow may help speed up the site as well, but are negligible in comparison and also hard to measure between dev and prod envs.
to the Welcome page and loading the async conversion JavaScript files there. ! Only load the conversion JavaScript if the referrer is the Join page, otherwise conversion counts will be incorrectly duplicated. ! Load each conversion JavaScript in the <head> to make sure they load before the page loads, so that the user can’t click away before the conversion is recorded.
! saved round trip time for HTTP request/response ! saved JavaScript redirect time ! saved Symfony processing time ! Large time discrepancy between dev and testing benchmarks imply that database optimizations could help reduce page load times as well. ! Caveat is that this performance may not be replicated on production… still needs testing.
I didn’t have to do it manually! ! Time elapsed was measured using system time between clicking on the Sign Up button and having the Welcome page fully loaded, mimicking user experience. ! Firebug requires adding two separate page load times and did not take into account browser processing time.