We all know the worthlessness of the data puke. Our analytics tool can give us so much data. Then we realize that we have no idea what to do with it all.
Goals and key performance indicators (KPIs) are so important. What metrics should you use to see if your Web site is meeting its goals? Your Web site as a whole will have goals and KPIs, but your marketing efforts will (or should!) as well.
This is where segmentation comes in. Segmentation allows you to see past the overall averages/trends and focus on specific segments of your Web site and how your Web site performs for those particular segments.
In Google Analytics, there are a few ways to segment data, including profiles, filters, and advanced segments. Google Analytics Help has an entry on the difference between filters and segments and LunaMetrics has a great post on the topic as well.
In this post, I’ll be focusing on advanced segments.
Here’s a quick example – a marketing campaign. Let’s say you’re running an email campaign. You’ve tracked your links from the email. Great start. You have your click-throughs, you have your bounce rates, you have your goal conversions. Great.
With Google Analytics (and most other tools) you can easily build a segment for users coming in via a specific campaign (or source, or medium, or trackable URL for that matter).
So you build your segment (users coming in from a specific campaign) and then you analyze the behavior of only that segment. Talk about beyond the click-through!
First, what do you want those users to do?
- Do those users do what you want them to do? (view a page, click a link, register for a course, etc.)?
- What are the most popular pages of that segment?
- What’s the bounce rate of that segment?
- What’s the time on site of that segment?
- Is that segment using your knowledgebase or internal site search? What terms are they using?
Let’s also say that the campaign you’re running is to get your applicants to submit their applications. You segment out the users coming in through that campaign. You might be surprised to realize that these users visit X page often – maybe it’s your tuition page, or the financial aid page, but … maybe it’s a certain page geared more toward current students! Why are the prospects/applicants interested? Find out!
Another example – user experience. You test across browsers, right? Of course. But do you really know how cross-browser-friendly your site is? You can find out by comparing browser segments.
Here’s a simple example:
Of course, first and foremost, look across date ranges. Does this trend stay relatively the same? If it does, this site might say that the above site might be pretty cross-browser-friendly.
As opposed to this site:
Again, look across date ranges. Does this trend continue? If it does, there *may* be a cross-browser issue here. Notice the time-on-site discrepancy between Internet Explorer and Firefox? This may be an issue, especially since the majority of the traffic is from Internet Explorer.
This should make you dig deeper. Look at other metrics and see if the trend continues. Do more cross-browser testing, especially of key areas of the Web site. Don’t assume, though, just from looking at the numbers. Dig deeper and find out for real.
Does your Web site target mostly in-state users? If so, maybe looking at in-state vs. out-of-state is a good segment comparison for you.
Another great segment to use is an “engaged user.” You can set up a segment for users who look at more than 3 pages or spend more than 3 minutes on your site. Does behavior for this group differ from your visitors who only see 1, 2, or 3 pages or are on the site only a minute?
The list is endless and it all depends on the goals of your site.
Looking at onsite behavior for different segments will really give you so many valuable insights that, without segmentation, may have been impossible to see.