Measuring Social with Advanced Segments

Following on the heels of my last blog post, which covered the #1 mistake marketers make with tracking social media, this time we’re going to talk about how to create an advanced segment to segment out your social media traffic. Not just the sites that you have links out on but also other social sites, like Digg, Reddit, and Technorati.

When I analyze traffic patterns from referring domains, I like to tease out the social sites. If your site caters to a social audience and is participating in different types of social media, this is especially important. However, even if you’re not, your visitors might be. And they may be finding you on these sites.

Case in point: I once worked for a publishing company that received a tsunami of traffic from Digg long before they ever broke the ribbon on a single social network. This was because one of their tutorials was written by an illustrator who worked on the move A Scanner Darkly, and he covered how to do that technique in Illustrator. Of course, it was evidence of increasing social activity like this that wooed the company into the social hemisphere of the Web.

So how do you segment your social traffic out if you’re using Google Analytics? With a feature in called Advanced Segments.  You’ll find it chillin in the upper-right corner of your screen.

Setting Up Your Segments

Setting up advanced segments isn’t as scary as it may sound. I think the use of the word “advanced” can throw some people a little. You’ll see, one user interface (UI) issue aside, creating segments is very intuitive.

Step 1: Click on the All Visits drop-down menu, then click the Create a new advanced segment link on the left side of the Advanced Segments menu.

Step 2: When you’re into the advanced segments hardhat area, you’ll see on the left side of your screen a list of dimensions and metrics. You could go rooting through those categories and subcategories. But if you’re too cool for school, you’ll want to just start entering the dimension or metric that you want and let Google do the heavy lifting.

Tip: One way to keep metrics and dimensions straight is if you can measure it with a number of some type (percentage, number of pages, time, etc.) it’s a metric; everything else is a dimension. Hat tip to John Marshall from Market Motive for that little heavy revvy. ;)

Step 3: Drag your dimension, Source, over to the dotted box that says dimension or metric inside. Then, from the Condition drop-down menu choose Matches regular expression. In the Value text box, you’ll want to enter the regular expression that encompasses all of the social media sites you want to track. There are so many to choose from, and new ones are popping up all the time, so your chances of catching all the social butterflies with one jar are pretty slim. But you can at least track the ones that make the biggest impact on your bottom line.

Step 4: Separate each of the social networks with a pipe character. It can be found above the \ character just north of the Enter key (Return on a Mac). The pipe character simply means or. So in this case you’re telling Google, “Corral all visits from Facebook or Twitter or Digg (ad nauseam) and funnel them into this segment we’re calling Social Media Traffic.” For our site, the regular expression we use is:

facebook.com|twitter.com|digg.com|stumbleupon|del\.icio\.us|netvibes|bloglines|linkedin|technorati|faves.com|aim.com|friendfeed|blinklist|fark|furl|newsgator|prweb|msplinks|myspace|bit.ly|tr.im|cli.gs|zi.ma|poprl|tinyurl|hootsuite|ow\.ly|delicious\.com|reddit|ht\.ly

I cheat quite a bit here. Technically, each of these should include the .com (or .net or whatever). And the period that precedes com or net should be preceded by a \ to “escape” the . character, which in regex world is also a wildcard that allows for any one character to be used. However, the chances of one random character coming before aim and com above are pretty slim jim. But I also show you how to comment some of them out (e.g., ow\.ly). That just tells Google that you want it to treat the period as a character and not regular expression. I also type them out in a document and then paste the list in since the text box you’re working with is pretty spartan.

The best way to know if your regex is sufficient is to put the string into your Referring Sites report filter and check out what comes up. To do this, go to Traffic Sources > Referring Sites, and paste it into the filter below your report.

When I tested ours recently, I discovered that my use of aim wasn’t sufficient because we had traffic from www.aimclearblog.com in our social segment because Marty Weintraub came to BlueGlass FL and had a wrap up of the conference on his blog. (Hi, Marty!) So I had to add to add the .com. Not that Marty isn’t social or anything … :)

There are times adding the .com still isn’t sufficient. For example, let’s say you get traffic from www.insidefacebook.com. Including the .com after facebook wouldn’t cordon this traffic off.  In cases like this you’ll need to use another regular expression character called the caret (^) in front of facebook, which says, “This source has to start with facebook.” It would look like: ^facebook.com.

Tip: Voice of experience speaking here … I recommend that you do a search/replace in Word to winnow out any rogue spaces. They can be hard to spot when you’re sporting multiple lines like I did when I created this segment.

Step 5: Name your segment, then test it to see how many visits you get for it. You can also choose to apply it to other profiles, before creating the segment, if you have access to more than one account/profile that you would want to apply this segment to. A very convenient option for agencies. Now you can click the Create Segment button.

Pet Peeve: From a usability standpoint, this UI is a disaster. When you go through a workflow, it makes most sense to drill down through options or work your way to the right. In this window, you have to jump around because you can’t test the segment until you name it (which is below the Test Segment button and to the left). Then you have to reach down to add your segment to more profiles and THEN click the button just to the right of your segment name. And if you save your segment before testing it, the window closes and you have to reopen it to test. I have few complaints about Google Analytics, but that is definitely one of them. It took me a while to get used to doing the hokey pokey to create, test, and mass apply a segment.

Step 6: Go back into your segments and select your newly created social media advanced segment and analyze the bajeezies out of it. Which social sites are you getting the most traction in? What are those bounce rates of your social visitors? Do they stick around? How many pages do they look at? Do they convert? How much revenue did you make off of these visits? Do they come back? Do they use search for their wish list using your site search? If so, what are they looking for? (Maybe you have a popular product that’s hard to find. Or maybe, like one client, you find out something a lot of your visitors want and make tens of thousands of dollars because you met the demand you had no idea was there.)

You can also compare visits from your Social Media Traffic segment (or whatever you call yours) vis-a-vis your other segments, such as organic, PPC, email, etc.

Cheat and Use Ours

If you want to use this segment for your Google Analytics profile, simply click this link, which will open up the window you see in Step 5 with everything filled in as you see here. Then, if you have multiple profiles, you can apply them in one fell swoop.

Analyze, Test — Rinse, Repeat

Clearly, segmenting out your social traffic will give you an idea of the quality of visitors who descend upon your site from social sites. Keep in mind, even if your social media traffic doesn’t convert well for your macro conversion (main conversions), they might be doing smaller tasks that build up to the bigger conversions, like signing up for your newsletter or following you on Twitter. That’s why you need to make sure you make your goals robust enough to capture these macro and micro conversions, or you might underestimate your social traffic.

And if you find that traffic from social sites exhibits certain propensities — like they gravitate toward posts about social media or funny one-offs — run some A/B or multivariate tests to see if you can find the right bait to reel them in.

Can’t get enough analytics? Here are some more posts that swing analytics for your perusal. Do you have any other uses for advanced segments or have any questions about how to use them? If so, share them below or hit us up on Twitter (BlueGlass or Annie) or Facebook.