These are some of the techniques I use for Advanced Segmentation in Google Analytics to spot weaknesses and opportunities to get better results from sites I analyse. Which techniques do you use?
Pattern Recognition by William Gibson is one of my favourite book titles of recent years is (it's not a bad read either if you like the Sci Fi genre). I like the title since it seems to encapsulate a lot of the analysis we do in online marketing and especially through web analytics.
Analytics is on my mind at the moment since this Thursday I'm presenting an Econsultancy Breakfast Briefing on Getting the Most from Google Analytics.
Configuring Google Analytics
In the briefing, we'll start by taking a brief look at configuring Google Analytics correctly. 70% of businesses say that they have Google Analytics configured correctly according to the recent Econsultancy Online Measurement and Strategy survey report. But I'm not so sure this is the case since there are many, many options for Google Analytics setup to answer the questions marketers ask like "how do I track visits from social networks".
But back to pattern recognition. I believe that the most important patterns to recognise are the relative differences in performance of whatever you're measuring - it's all relative as they say, whether it's relative differences in bounce or conversion rates from different traffic sources or for landing pages for example.
Advanced Segmentation Patterns
One the great tools for looking at relative differences in performances is the Advanced Segmentation in Google Analytics, so as a preview of part of the briefing I thought it would be useful to look at some of the most best ways to use Advanced Segmentation in Google Analytics.
I'm always interested to learn alternative approaches from readers of my blog posts, or tweets, so please add your recommendations on which standard or custom Google Analytics Advanced Segmentation in the comments or via a Retweet.
Advanced Segmentation options in Google Analytics
If you're thinking of the best way to use Advanced segmentation to get better results from your site, I'd recommend reviewing these 7 options. My examples use a combination of standard and custom segments. For example, here is a custom advanced segment to isolate social network referrers...
1. Segmentation by Referrer / Traffic source
I use the default segments paid versus non-paid traffic a lot to understand how well different keywords are performing. I also often setup a branded vs non-branded keyword segment, perhaps using a registered expression to capture different variants of the brand name.
Custom segments isolating a particular partner / referring site such as Twitter or social networks can also be helpful. It is generally best to isolate social network as a filter as this excellent post shows.
Campaign tracked segments can also be useful. For example, Econsultancy have configured Google Analytics with campaign tracking so that the relative popularity of e-newsletter content can be seen.
By looking at the relative % of the e-newsletter segment you can see that the newsletter has been relatively successful in driving visits to the third report.
OK, you may be able to see that through your email package, but you can't see the click streams or relative differences so well.
2. Segmentation by Visitor Type
This is a straightforward one - segment by returning visitors to see which content is popular with returning visitors against first time visitors.
Or a bit more sophisticated, use Page depth to isolate engaged customers - Avinash Kaushik has suggested using Page Depth > 3. Then you can ask questions like where do my more engaged customer refer from, where do they enter, are they more likely to convert. More sophisticated still you can create custom segments, for example one I used recently on an analytics project was Paid, First time visitors with Goal Completion > 0.
3. Segmentation by Content Viewed
People who see one sort of content versus another can be used for ad-hoc analysis to see which content influenced conversion, although there are other ways to do this like $Index if you have goal conversion setup.
4. Segmentation by Landing Page Type
Since segments are applied on previously collected data, they are useful to isolate the impact of landing pages although Navigation Analysis reports such as Navigation Summary and "How visitors found your content will show this.
5. Segmentation by event
This is by Sales (Visits with Transactions) or Conversion Goals (Visits with Conversion Events)
You can also segment by customers who have bought or visited several times to see how your returning visitors use the site, where do they enter for example.
6. Segmentation by Platform
You can segment by different browser and screen types to understand preferences for different types of users.
An interesting at the moment is the new standard segment for iPhone. Econsultancy can use this to see the type of content accessed - the content drilldown report is the best Google Analytics report to apply this to since it aggregates all content.
For example, this is the content preferred by Econsultancy iPhone users.
7. Segmentation by Location
I had an interesting discussion with a retail marketer recently when she described how she used their analytics tool to look at how visitors from different countries browsed the site differently - visitors from different country tended to use different options on the main nav which could help inform how to setup country microsites or email content.
Thanks to Avinash Kaushik for inspiration in using Advanced Segments. Avinash has done a great job in promoting and giving examples of the benefits of Advanced Segmentation when it was launched nearly 2 years ago.
Attend the event
You can register to attend the Econsultancy Breakfast Briefing on Getting the Most from Google Analytics.