Should You Convert To The New Google Analytics Code Now?

December 19, 2007

With the new Google Analtyics code (ga.js) now available, a number of people have emailed us over the past few days, asking the question in the title of this post.


In the matrix below, you can see how I visualize the issues.

Should I convert to GA.JS?

We have here a business problem (how badly do you need the new code?) crossed with a technical problem (how easy is the implementation for you?)


The southwest quadrant is the easy one. You are dying for the new features that you will get in the new ga.js (the new Google Analytics), such as event tracking. And your implementation is an easy one. Easy implementations include:

  • You have a staging server so can make lots of mistakes at no risk
  • You are implementing the code for the first time
  • You are redesigning, and so need new code anyway
  • …and then there is always, “What’s the big deal? I have a small site, it’s one domain, if I lose a day of data, life will go on.”

In the northwest corner, we have a similar (but not as compelling) situation – the implementation is easy, but you were pretty happy with the old urchin.js code. So, convert when you can. Like I said to one of our customers today, “Let’s finish this project and then do it after the new year.”

In the northeast corner awaits our most difficult quadrant. There, we no longer have a compelling business case, and the implementation is a hard one. For example, you might have a large, mission-critical site without a staging server. (I know that sounds strange, but I have seen stranger situations.) When this is no longer easy to do, and you don’t really need the cool features you will get, you will be hard pressed to do the conversion. Maybe that means, hard pressed to get it into your budget. Ultimately, you’ll have to convert. I have no inside line on when GA will require everyone to convert. But I do feel confident in saying that you will get some warning of some kind (and I don’t have any inside scoop on that. I just believe it.) My recommendation: procrastinate.

Finally, there is the southeast quadrant. Here, desire for the cool new things you can do with the new version (like making it really easy to have different copies of the ga.js on the same page, so that you and your advertisers can both see how you are doing) will supercede your technical difficulties. To this website owner, I suggest that you let a couple of months pass (after all, you can’t get a lot of the new functionality yet until you have the ability to see those events. So you might as well let other people debug the code.) You could also start to gradually add the code on pages that aren’t mission critical. Remember, you can’t have both the new ga.js and the old urchin.js on the same page. You might start with a test site so that you can discover issues yourself.

So, should you convert to the new Google Analytics (ga.js) code right away? It depends.

– Robbin