Segmentation is one of the key weapons of the success marketer.

It’s allowed them to get the right messages and products to the right people. It’s a core part of the marketer’s tool kit.

So what can marketers learn from segmenting their social audiences?

Few marketers have a clear picture of who is in their audiences and how they might apply the tools of segmentation to their social strategies.

A social media manager can use segmentation to understand exactly what content to create and who to engage with based on what actually matters to their overall marketing strategy.

Here are three ways you could segment a social media audience.

  • By demographic.
  • By who influences them.
  • By the hidden affinities within the audience.

To demonstrate, we looked at the public Twitter data of several well-known brands.

By gender or age

We looked at the audience of @nikeuk, the sportswear brand, which has more than 117k followers, and worked out the ways in which men engaged differently to women.

Men outnumbered women about five to two in this Nike audience. This male segment talked about football, sport and gaming, in particular. The smaller female segment talked about sport but also TV shows and Christmas.

Last week (Dec 11th to Dec 17th), for example, the most outperforming hashtags amongst the men who followed  were #mufc (Manchester United), #spoty (Sports Personality of the Year), and #14daysoffifa (about the launch of Fifa 14).

The celebrities who most influenced the male audience were Wayne Rooney, Jack Wilshere and Phil Neville.

For the smaller female segment, the most outperforming hashtags were #Xfactor, #Spoty and #mcfcafc (the Manchester City v Arsenal match where City won 6:3). The celebrities who most influenced this audience were Jessie J, Ashley Cole and Paula Radcliffe.

By who influences your audience

Identify the top 100 or 10,000 influencers who influence your audience, and figure out what they care about. These are the people or accounts who get the other members of your audience to engage.

For example, we looked at the 1.1M followers @Nokia, the main Nokia account. We discovered 1,700 influencers who influenced more than 100 or more other followers of the @nokia account. These top 1700 twitter accounts could, in theory, reach more than 50m accounts across twitter in total.

These influencers were located in different geographies (about 50% in the US, 10% in India) – and had very different content and brand affinities to the audience of followers as a whole, (#mtvstars anyone?)

By working out what binds them together

In this approach, you look at how users behave and interact and use a clustering algorithm to determine coherent groups based on infinities.

For example, we looked at very successful Twitter account of the luxury US store, Bergdorf Goodman (@bergdorfs).

Our algorithm looked at millions of tweets sent by the members of this audience, and uncovered areas of commonality. In face we found four distinct segments (and a ragtag group of miscellenous people that didn’t fit into any one category).

  1. Jewelry lovers represented 4.6% of the audience. These people talked about fashion shows, NY fashion week and 2014 Spring Summer collections.

  2. Fashion-forward women represented about 14.9% of the audience. They also talked about NY Fashion week, but unlike the first segment were also heavily engaged around Paris, London and Milan fashion week.

  3. Women talking about family-oriented and fashion issues represented 11.9% of the audience. They talked extensively about the Style TV network an upscale New York kids clothes stores.

  4. Celebrity lovers represented about 15.3% of the audience. This group engaged less around fashion or clothes content and much more around celebrities like OneDirection.

Segmentation is an essential part of every marketers toolkit. Social media provides for exceptionally high quality opportunities to segment,to strategically build your audiences with the right people and get the right messages to those people.

How you do that will be the subject of a follow-up blog post.