Document Complete vs Fully Loaded

Web Performance Discuss Test Results
1

Hi,
I am a newbie. What is the difference between Document Complete and Fully Loaded? My test results usually come in at 5 seconds for doc complete and 8 secs for Fully Loaded. So, 3 sec difference.
Thanks!
Todd

2

Document complete is the point where the browser onLoad event fires and is generally when all of the static page content has loaded. Fully loaded is the point after onLoad where network activity has stopped for 2 seconds (usually any activity out here is from javascript on the page doing something dynamic).

3

Why do you say “…after network activity has stopped for 2 sec…”?
Is it always 2 seconds before determining if the fully loaded time is completed?

TRX

4

Yes, give or take a few hundred ms for the polling interval. The 2 seconds isn’t included in the measurement, the measurement ends at the last network activity before it was determined that the connection was idle.

5

Thanks for the clarification!

6

Hi Patrick

Thanks for the detailed answer, We are trying to measure User Experience time i.e when does the page becomes visible and usable by the user, hence could you let me know if we should track the Document load time or Fully loaded time ?

My understanding is as the document completes - the page becomes usable and that should be my User Experience time, there could be some Async calls happening post Document Complete, which user need not wait for.

Please let me know your thought on the same.

Thanks
Sailesh

7

I know this is from many months ago, but in case others are looking at the same question:

That “Document Complete” number actually indicates when your browser has finished loading everything, and it usually indicates when it starts rendering. How long it takes to render the page for the user depends on the speed of the user’s machine, and how much junk is being loaded along with the page. If the page has a ton of includes (such as CSS and JavaScript,) you will have a significant delay before the page finishes displaying for the user, just because the user’s browser is rendering the CSS and running the JavaScript.

That’s normally disregarded in metrics, because it varies by machine, but if you have a lot of requests in your page, and a lot of CSS complexities or multiple JavaScript libraries, you may well have issues for users who have slow machines or machines with little available memory (not always the same thing.)

Document Complete is still the best place to see how fast a page is, but look at the number of requests, too. The “Speed Index” is also helpful, as are the Google PageSpeed numbers.