Telmar News & Insights

Audience Segmentation Techniques Media Planners Need To Know

August 15, 2019

Audience segmentation is key to effective and targeted marketing strategies that can help to improve reach and frequency. Segmentation tools allow you to subdivide your target into groups of shared characteristics and differences. Subdivide your target into groups using things like behavior, lifestyle, life stage, attitudes, and demographics. Segmentation tools can capture a staggering amount of data. Parsing, understanding, and making this data actionable, however, can be a challenge. 

Sophisticated data integration techniques can make multiple sources of data work together for better audience analysis. Data integration can bring primary data and syndicated (general market) data together or bring two separate syndicated data sets together like, consumer research, and media audience research. Below are four techniques that combine and contextualize data and research to help media planners successfully conduct audience segmentation analysis.

4 Audience Segmentation Techniques

1. Correspondence Analysis

Correspondence analysis is used to identify and understand relationships between markets, brands, and media. It identifies the factors that differentiate between people in a market, as well as identifies potential market gaps. Correspondence tools help to quickly see these relationships by generating a correlated pictorial perceptual map of your cross-tab results. 

Take, for example, a correspondence analysis that quickly segments brands to reveal which of them align with consumer attitudes such as “strongly concerned about the environment,” “strongly concerned about getting the best price,” and “strongly concerned about making a fashion statement.” Correspondence can also be used to identify the most discriminating or important lifestyle statements in preparation for using a cluster analysis tool.

Related Content: Cross-Channel Advertising Best Practices

2. CHAID Analysis

CHAID analysis reveals the most significant variables and combinations of variables that comprise your target population. The best combinations of variables, analyzed as discrete groups, can then be worked with as predictive sub-targets for reaching the root target population. As an example, let’s say your broad marketing target is “high-end performance car buyers” and for certain media channels you wish to use, you are not able to evaluate their audience data for this target.

CHAID analysis will help you discover which combination of variables (that you know are media audience measured) can be used as the most predictive ones to reach high-end performance car buyers. Running a

CHAID analysis with various demographic variables against “high-end performance car buyers might reveal a predictive target of Males 35-64 with a household income of at least $150K who live in a particular region as heavily comprising the best prospects in your target. That finding can then be used for media planning and buying where data on high-end performance car buyers is not available.

3. Cluster Analysis

Cluster analysis segments a target audience into multiple smaller target groups (clusters) whereby the people within each cluster are maximally associated with each other (homogeneous) and the groups themselves are maximally differentiated (heterogenous).

We know that all high-end performance car buyers are not alike. Cluster analysis will identify and create target segments that can each be analyzed and acted upon with refined marketing and media tactics for more effective reach and efficient media spend.

Related Content: Best Practices for Data Visualization

4. Factor Analysis

Factor analysis is well known for its ability to reduce a large data set down to a smaller one for easier handling. It achieves this by determining which variables in a large data set have the strongest underlying dimensions that are inter-correlated and groups these variables into buckets called “factors.” Each factor can become a new variable within the data set and effectively replace all the variables that it represents.

Data reduction is cool, but even cooler is how factor analysis can reveal latent (hidden) information in your data by providing you with answers to questions that were never asked. Let’s say you’re audience targeting and the target is a “smartphone super users” and you have data from a survey that asks a whole bunch of questions about all the features the respondent uses on their smartphone.

The survey did not ask any questions that identify the level of expertise of the respondent. Trying to identify experts based on one or two advanced features does not work because most people know how to use some number of advanced features. 

Factor analysis rescues this situation by being able to use all the questions asked about features to discover the underlying dimensions that are inter-correlated into a factor that can be identified as “Expert User.” From this factor a new variable can be added to the data set as “smartphone super user” and it can then be used for segmentation analysis and in all the other audience segmentation tools.

Conclusion

Powerful segmentation tools can be tremendously helpful in gaining a deeper and richer understanding of your audience targeting. Insight into audience targeting can fine-tune your tactics, overcome targeting obstacles, and achieve a more cost-efficient use of your budget to reach more potential customers.

It all starts with the data. If your data is not rich enough to fulfill your needs, using data integration can yield greater targeting value from the data that you already have.