Google Analytics Help: Avoid These 5 Common Mistakes

by DigitalAlex on April 25, 2008

google analytics helpYou’ve heard it a million times before: garbage in, garbage out. Without the right setup, you could get bad data or, possibly worse, no data at all.

There are plenty of ways to improve your Google Analytics implementation, but here are the 5 most common mistakes I’ve found:

#1 – Mixing urchin.js and ga.js code

Google Analytics comes in 2 flavors: the old code and the new code. The old code is “urchin.js” and the new one is “ga.js”. Not sure which one you use? Go to the page in question, right click and select “View Source” or “View Page Source”. Search for “ga.js” and “urchin.js”. The code is typically located at the bottom of the code.

Google maintains that you can use the codes interchangeably as long as the codes are not on the same page. In my experience, this doesn’t work. I’ve seen referring source and campaign tags stripped when people move between pages with different code.

Bottom line: pick one code and stick with it. For additional help finding pages with rogue code, try Stephan Hamel’s WASP.

#2 – Using the Wrong Google Website Optimizer Code

google website optimizerGoogle Website Optimizer is a great, free website testing tool. It’s a tag based tool, meaning you have to put code on your page just like an analytics tool. Just like Google Analytics, you have to ensure your tags have the proper code and are in the proper place.

One less obvious issue with Google Website Optimizer is that it uses urchin.js, the old Google Analytics code. If you’re using the new code, ga.js, on your website then you’re data is likely going to mess up your site analytics data.

In order to get them to work together, you need to modify the GWO code to use ga.js. Check out this and this article from Google’s help files to get you started.

#3 – Forgetting AdWords Conversion Tags

As I mentioned in my post about how Google Analytics fails paid search marketers, you cannot currently see all 4 lines of your ad creatives within Google Analytics. It only shows the first line. This isn’t very friendly to creative testing.

While the Google Analytics works on it (and they are a great team), you have to make sure you setup AdWords Conversion Tracking. That way, you can see conversions inside of the AdWords interface. You’ll then be able to see what creative drove conversion increase, not just clickthrough increase.

Here are 5 more tips to improve adwords conversion tracking.

#4 – Missing Page Tags

sitescanEvery page needs to be tagged. If you’re missing page tags, anyone who navigates through those “holes” in your analytics can cause all sorts of issues–missing campaign tracking, the appearance that your site is referring to itself, etc.

For a first pass at finding gaps in tagging, I use SiteScan. It has some limitations, but it’s currently the leading free tool to help you find gaps and fix them.

You can also look in the Referring Sites report to see if your site is referring to itself too frequently. Sometimes these self referrals will happen when a session times out, but an exorbitant amount can point to deeper issues.

Also, check out the Traffic Sources report and look at the Conversion tab. Are you seeing tons of conversion for Direct, but much less than expected for other channels like cpc (paid search)? If your campaign tags are setup (see below), then it could be an issue with page tagging.

#5 – Missing Campaign Tags

Google Analytics can automatically recognize AdWords marketing. Everything else requires campaigns tag–adding code to the end of your destination URLs. Check out my article on campaign tagging for more info.

What common stumbling blocks do you run into?

{ 8 comments… read them below or add one }

1 SFGreg April 28, 2008 at 5:48 pm

This is a great, practical article. I did not know about potential js conflicts when using the Optimizer, so this is great to discover, especially since we’re about to switch to the new ga.js and start a round of A/B tests with the Optimizer. Thanks!

2 DigitalAlex April 28, 2008 at 5:50 pm

Hey Greg,

Glad I could be of assistance. Yeah, I didn’t know of those Google Website Optimizer & Google Analytics issues either. Then, I was working with a client to diagnose their implementation and we eliminated every other issue I found and that was the one last thing stopping the data.

Good luck with the testing. Where are you sourcing your ideas? Any lessons learned from previous test? I’m becoming obsessed with optimization… :-P

Cheers,
-Alex

3 DigitalAlex December 4, 2008 at 11:21 pm

Glad I could help. Report back if you find any useful tricks.

4 Jose December 10, 2008 at 3:41 pm

Hi Alex,

Very nice post. Talking about the first issue “#1 – Mixing urchin.js and ga.js code” I have a couple of questions:

1) Have you ever noticed that moving from urchin.js to a mix between both codes, or going from one mixed environment to just using ga.js, alters your count of “Absolute Unique Visitors”, “Visits” or “New Visitors”?

2) Does urchin.js and ga.js use a different cookie or parameter to track recency or frequency?

I have a case where “it seems” that switching from urchin.js to ga.js had an impact in the count of visitors and visits. What do you think?

Thanks!

5 DigitalAlex December 11, 2008 at 12:05 am

Hi Jose,

1.) In my experience, I don’t recommend you mix both codes. It generally seems to wreak havoc with a number of metrics, though I’ve never checked the ones you mentioned.

2.) I know that neither code uses a parameter to track recency or frequency. They use a cookie. Now, I’m not sure if changing Google Analytics codes resets the cookie. Since it’s calling a new javascript file, that’s very possible.

I wager that you’re right. First, make sure the codes were both in the same location, which should be before the close body tag,

After that, make sure that any customization you did with urchin is adequately carried over with ga.js. It’s not always the same.

Hope that’s a bit of Google Analytics help.

-Alex

6 Neales December 19, 2009 at 7:44 am

Changing Google Analytics codes actually resets the cookie. Has anybody considered that? Can that be neutralized?

7 Alex December 19, 2009 at 10:31 am

Hi Neales,

I don’t know of a solution off the top of my head. I’d recommend you check with Justin Cutroni at Epik One/Analytics Talk. He’s very much the expert on the technical pieces of it.

He’s also on Twitter, @justincutroni.

-Alex

8 Sadekya December 23, 2009 at 3:34 am

Your article helped us iron out technical glitches we were facing with Google Analytics especially the issues relating to “New Visitors”.

Leave a Comment

Previous post:

Next post: