Despite the increasing use of one-to-one targeting approaches in digital and social media, segmentation remains very popular and an essential component of the modern marketer’s toolkit.
This prompts the question of what makes a segmentation approach “good”?
Over time, six key criteria have been established as a basis when determining the efficacy of a given segmentation approach.
In a ‘good’ approach, the segments must be…
- Broadly identifiable across multiple research sources, making it possible to build a comprehensive picture of segmental preferences, drivers and channels using multiple research methods.
- Sufficiently large in order to justify separate targeting or development of separate value propositions and marketing programs.
- Stable so that individuals remain assigned to a particular segment for a reasonable period of time and the segment represents the truly persistent characteristics of the individual.
- Differentially responsive to important value proposition, messaging and communications elements.
- Readily visualised, understood and communicated within all levels of management and their agencies.
- Broadly executable across a range of marketing channels and able to function as a framework for optimal multi-channel communications planning, based on segmental channel preferences.
With the increased use of digital, database and geospatial channels like OOH and OTT, executability has become an absolutely critical consideration in choosing a segmentation approach.
In traditional geodemographic systems like geoSmart, Mosaic or Helix, the segments suffer from being too small due to the large number of segments and insufficiently differentiated due to their fitment at street or household level.
Without a factual representation of the lifecycle stage in traditional geodemographic schemes, mums, dads, children and grannies in the same street and within the same household are treated in exactly the same way. This is a real problem in categories like fitness, financial services, retail, telecoms and healthcare, where needs and preferences are strongly influenced by lifecycle stage.
Also, the large number of segments and obtuse names in traditional geodemographic schemes means that they are difficult to use in general management communications.
Behavioural segments are usually built to maximise differentials in customer responsiveness, however, incorporation of historical behaviour means that the assignment of individuals to segments tends to be ephemeral. Also, behavioural segments are generally identifiable and executable only within the environment where they are constructed.
The specific behavioural content of these schemes also often makes them difficult to execute beyond their technical user base.
Attitudinal segments tend to rate strongly in terms of size, stability and differential responsiveness, however, like behavioural segments; they are limited with respect to being indefinable and executable, particularly on databases.