Why not do what “kl” suggested doing (near the top of the comments section)? Are there any drawbacks?

Someone should tell google to turn on gzip compression for the text/javascript mime type, or to switch the suggested code to application/javascript. It would probably save them terabytes per month in traffic and make peoples sites go faster

i.e. Google’s “Instructions for adding tracking” don’t make it clear how important it is to stick to the example tracking code – if at all.

The problem is that while that will, indeed, get most of the content to render before requesting ga.js, it holds up the visual indicators of “doneness” in the browser. Further, it causes any scripts that expect to be run onload to be significantly delayed. Depending on the site, the UI consequences of this lag can be mild or severe. I’d just prefer not to have to deal with it at all.

Regards

Sure, you could miss it, but I’m not sure I’d call it a “visit” if the user never even got to the point where all the resources loaded. I’d hope that the goal of all content is to provide value to the user first and to the publisher second, so by that rubric, missing a couple of clicks on analysis is no big loss if the user gets a better experience. I can see valid reasons for your personal calculus being different, though. I’m only tracking a personal blog, after all.

Regards

The (small) benefit of the original script is that missing a track is unlikely because the page is forced to wait?

Wait? We have a right hand?

Regards