Analytics Basics: Segmentation

This article, written by Neil Mason, was originally published on Clickz.com and is republished here with permission.

Over the past few weeks I have had the opportunity to try out the new Advanced Segmentation features that are being rolled out in Google Analytics. Although it’s a beta version of the new capability it allows Google Analytics users to take a deeper dive into their data and to discover underlying patterns in visitor behaviour.

In the past couple of years since I first started writing about segmentation in this column (www.clickz.com/3615916) it has become easier for digital marketers to look at different groups of website users. Developments in the technologies mean that there is almost no excuse for not beginning to segment your website visitors and to begin to understand how different types of visitors behave on your site. Having said that there are still many organizations that only look at the top line numbers and treat all visitors with a “one size fits all” approach to marketing.

So what is segmentation and how can it be used? In marketing terms segmentation is the process of identifying groups of individuals that have something in common. Those individuals then belong to the same segment. Importantly, what those individuals in that segment have in common is different to what other individuals in other segments have in common. So a simple example would be to segment users by the number of times they have visited the website. You could classify them into “buckets” such as new users (the first time they have visted), light users (i.e. have visited 2 to 3 times) and heavy users (i.e. have visited 4 times or more). The number of buckets in this instance would be determined by looking at the distribution of visits per visitor on the site and making an appropriate decision.

The point of making segmentation such as the one described above would be to understand whether there are any differences in the behaviour of these different groups which could then lead to some kind of differentiated marketing message or user experience being developed. If there is, then there is the opportunity to improve the expected return on investment. Let’s imagine that you have just run some kind of email marketing campaign and experienced a 10% uplift in sales. If the campaign was a generic campaign that you sent to all your users on your database, then what probably happened was that some segments of users really responded to your campaign and others didn’t. The reality is that you might have experienced a 20% sales uplift in some segments and 0% uplift in others. The idea behind segmentation is that if you can design a series of different campaigns that are relevant to each of the different segments then you should achieve say a 20% uplift in sales across all segments.

With most web analytics tools these days it’s possible to carry out basic segmentation of the data. It’s up to the analyst or the user to determine and decide what are the useful segments to look at. In GA (Google Analytics) there are some segments that are already set up that cover some of the obvious behaviours that might be of interest, like new vs returning visitors, paid for vs non-paid for traffic but the analysts will want to explore and create their own segments based on their own understanding of the website, and also the kinds of issues that are being faced in the business. For example it might be that there is an issue in the business about how many people look at a quote for a hotel room but then don’t book. In that instance the analyst may set up three different visitors segments as follows:

• Browsers (people who visited the website but didn’t reach the quote process)
• Quoters (people who got a quote but didn’t book a room)
• Bookers (people who booked online)

Once the segments have been created, the analyst can look for differences between these segments in terms of how they get to the website and what they do when they get there. Are there differences in the type of channel they come in on or the keywords they use? Do they look at different types of content? Do they use different tools and applications on the site such as the on-site search? If so what are the kinds of keywords do they use on the on-site search engine?

As you can see, once the ability to create segments becomes a reality, the analysis possibilities become endless. Different tools offer different capabilities but the principles remain the same for the curious analyst. The challenge though with segmentation is not so much in doing the analysis but taking action as a result. The benefits from taking insight into action, are building and deploying more targeted marketing programmes as a result.