Hello All, I have a question!

I was wandering if there are any reports on the loadtime difference caused by implementing the Google tracking script?

Are there any Benchmarks (response times/pageload times etc.)?
Is there amethod for testing the difference caused by GA?

Would it be usefull to host the ga.js script on your own server and update it maybe daily or weekly? I believe the communication between user and server takes the longest, would this reduce it somewhat?

Have any tests like this been done? Are there links to the results?
Is there a good method for testing the performance change by using GA?

Thanks,

Scot

Hosting the GA tracking code on your wn site; intersting idea, but I wouldn't do it.

Usually one places the GA code (or any other tracking code) at the bottom of the page. The rest of your content will thus load before the tracker is called. This hould allow for fast load times.

Also:

1) Google servers are probably better that yours, so why downgrade by hosting je js on your own server.
2) If you place it on your own server, the visitor will have to download it again, while when you load it from googles server it will already be in the cache, because many sitees use the same ga.js on the google server. This is actually faster.
3) If you host it on your own server, it will probably take longer to process, beacue browsers are limiters on the amount of request they can place at one single server at a time.

So I think hosting it on your own site, would not be a good idea!

To my knowledge ther is close to none extra overhead on the server side. The main load time problems are on the client side, but also this is very little as the 3 lines of Java wouldn't cost anymore resources than a simple copyright notice would cost.
So it could take a little time to load the google bity, but I've never experienced any load lagg with the rest of my content. It displays very fast and only the little progress bar/throbber in your browser top/bottom will be active a little longer. Meanwhile the vistor can view your page, while google catches up and loggs the visit.

It might be best practice to place your GA code at the bottom of your page. The always say directly above the body tag on their installation guide, but please ignore that and place it at the bottom. Just to be safe...

Hope this helped!

I haven't seen anything noticeable...

I think the call to Google (if it would not be in the users cache) is done on a seperae thread, than which the user is susing to download your page content. Maybe the DNS lookup would take some time, but this is neglectable. Also, because you are using two threads and are indeed downloading a little more data it may slow down just slightly, depending on connection speed. But the GA.js file is so small this will really, really not matter on the typical page.
I say a typical page, because this of course depends on the download size of your page, because everything is relative. On an all-text page with 3 lines, the ga.js will be a relatively heavy download. On the typical page; with images this will be overshadowed by the download of the images.

A problem could arise if you have other scripts being called on the onLoad event, as the google tracking function will have to be called first. Sticking the Google tracking code at the bottom of your page will probably overcome this. Alternatively you could us a workaround by using various "when DOM loaded" events. Theyh are usually more responsive, becuse you don't have to wait for your own scripts to load.

Hope this helped!

What exactly is the "Google Tracking Code"?