About a week ago, Google Engineer Aaron Hopkins discussed other methods [die.net] for reducing the page load time ... ideas such as HTTP pipelining (something a user can enable in their browser), spreading objects among multiple domains (something a web admin can do), and serving fewer, larger files as opposed to many small ones (again, something in our control). Fascinating stuff, especially the reasoning given behind each idea.

While these all sound like great methods to provide our visitors with a better experience, is there a way to simply measure the results?

At first I thought to myself, the bots! Google's webmaster console now shows the average time, in miliseconds, it spends downloading each of your pages over the last 3 months ... so there's one measurement. Then, there's Xenu's Link Sleuth -- it also shows the duration, although in seconds, for loading pages from whichever site it's hitting (note: I run Xenu on my top-priority sites bi-weekly).

But ... I am wrong on both counts, amn't I? While using mod_deflate will have an effect on the html sent to Googlebot or Xenu, that's the only change, as Googlebot doesn't retrieve every image every time it loads every page, does it? Xenu I know only retrieves an image once in a session.

Browsers don't work this way, do they? To the point, when I get these changes taken care of, I won't be able to use Google's console or Xenu's Link Sleuth to see much of a measurable change, will I? So, without being able to measure the improvement, it's more difficult to justify expending time and energy toward it.

So ... if you intend to move your site in this direction -- toward using less bandwidth, and toward decreasing the load time for your visitors -- how do you intend to measure your success?