What is customer segmentation?

Customer segmentation allows businesses to target specific groups of people so that they can reduce risks and sensibly allocate their marketing resources. It is a grouping together of your customers based on demographics, behavior, and so on. The creation of user personas can aid marketing activities, allowing you to appeal to the right group of people, which is great for advertising and content marketing. There are many elements to customer segmentation, and Google and Facebook can be used for understanding demographics, geography, and interests.

Google Analytics Demographics

In 2013, Google launched their demographic report within Google Analytics, and the collective cheer of marketers resonated across the globe. The information that Google includes is gold dust to a marketer, and it’s a great starting point for customer segmentation.

As with any dataset, it’s important to understand its potential inaccuracies and see it as one part of a larger puzzle. The data within Google’s demographics report is extracted from DoubleClick third-party cookies, so users without that cookie will not have their data reported. Google will also apply thresholds to prevent individual users from being identified, so sometimes the reports will only be based on a reduced percentage of users. Nonetheless, the data can be extremely useful.

To set up the demographic reporting, you need to:

  • Change the Google Analytics tracking code on your pages from ga.js to dc.js
  • Head to your account settings and enable tracking

For the full details of this process, head over to Google’s Help Center.

Understanding the Data

Once you’ve set it up, you’ll start seeing some fascinating results. Obviously, it would be foolish of me to include my own company’s data, so the examples below were kindly given to me by a friend who works for a company that sells a music based app.

First of all, simply head to Demographics – Overview.

CS1

We can see from the above that the business has a younger demographic and the gender split is even. When comparing different months, the results are very similar, which to my mind lends them more value; wildly differing results would suggest an error in the reporting. This is obviously dependent on the business and seasonality, so you will have to use your own judgment.

Use the top-left drop-down box to view different metrics, which include:

  • % New Sessions—find out if you’ve managed to successfully target a new demographic.
  • Avg. Session Duration—discover which group of people spend the most time on your website.
  • Bounce Rate—is a certain group of people leaving your website in droves?
  • Pages/Session—same as above, but allows you to check which pages.

Next, I’m going to head to Demographic – Age and then set Gender as a secondary dimension.

This allows me to see that the even gender split is apparent across different age groups.

CS2

Interests

A vital part of customer segmentation is understanding your customer’s interests outside of your niche. Head over to the Interests section, and you will find three main categories:

  • Affinity Category: groups users based on a collection of activities, e.g. Movie Lovers.
  • In-market Segment: this lets us see what other products our users are likely to purchase, e.g. Consumer Electronics.
  • Other Category:  this allows us to see which categories our users are also interested in, e.g. Arts and Entertainment.

CS3

Again, you will see the option to use the metrics listed above (% New, Avg. Session Duration, and so on) to understand different elements of your traffic.

You can also dig further into this to see the interest of your top performing groups. I could see that most of the app downloads were by females aged 25-34, so I followed this path: Overview – Audience – Gender – Female – 25-34.

This let me view the favorite interests of this particular group:

CS4

By now, you are probably beginning to build up some useful customer personas. Either export the data to Excel or start making some notes, as it will help to have the data you need at hand.

Facebook’s Graph Search

The ability to search through all the data in social media is thought by many to be the next technological “golden goose.” Social search is often referred to as the “next Google,” and it is why Facebook’s current income revenues are so disproportionate to their high valuation. Facebook launched their Graph Search in 2013. It’s still in the early stages and it is the first real attempt at “social search.”

Only those in the US currently have access to Graph Search by default, but there is an incredibly simple trick to accessing it if you live outside of the US. All you have to do is head to your general account settings and change your language setting to English (US). It’s really that simple! Some people have reported that it takes up to 24 hours to update while others have said it works instantly.

CS5

I’m going to take a look at MusiXmatch, as they offer a similar service to the company we have been using as a working example. So, let’s take a look at a few search queries to find some useful data:

Phrase searched: “Interests liked by people who like MusiXmatch

CS6

Now, let’s take a look at pages they like.

Phrase searched: “Pages liked by people who like MusiXmatch

CS7

We can also use more than one competitor in the search. This should lower the chance of any anomalies.

Phrase searched: “Pages liked by people who like MusiXmatch and Shazam

CS8

There is the potential for an almost endless amount of searches, as we can increase the variables:

Phrase searched: “Interests liked by people who like MusiXmatch and are older than 18 and younger than 30 and live in London”

CS9

The interests or data you find can also be used in the search. To get you thinking, here are just a few examples of what you can use in searches:

  • Location
  • Groups
  • Page Name
  • Topic of Interest
  • Band
  • Actor
  • Director
  • Drinks
  • Books

Digging deeper into Facebook

Simon Penson wrote a great post on Moz back in September of 2014, which mentioned a technique for using Facebook’s advertising feature to turn this qualitative data (albeit drawn from a large pool) into quantitative data. Unfortunately, this technique is no longer applicable. Facebook has been playing with how they calculate their potential audiences, so his calculation is no longer effective.

However, there is another technique we can still use, which involves digging a little deeper into Facebook’s application program interface, or API. Klipfolio has written a great post on this, so I suggest heading over there. The disadvantage of Klipfolio’s technique is that it is simply looking at interaction with your own profile, but it’s still useful for customer segmentation.

Where to go from here?

Google and Facebook’s data can be extremely useful for understanding your customer base. You should also consider psychographics, but that’s another blog post! Facebook’s Graph Search is still in its infancy—it’s fun to play with and can give some useful results, but once that data can be extracted in a more useful way, it will be invaluable. However, it is still currently helpful for use alongside Google’s data, and will allow you to create customer personas and target your customers effectively.

Was this article helpful?
1 Star2 Stars3 Stars4 Stars5 Stars (5 votes, average: 5.00 out of 5)
John Philips
John Philips

John worked for years as a freelance marketer and developer, before joining the team at SSLs.com. He’s an expert in marketing analytics, so he feels most comfortable surrounded by spreadsheets and graphs.