3 Alternatives to Demographic Profiling
In our webinar Beyond Gen Z: A Guide to Modern Segmentation we dissected demographic and cohort based segmentation techniques. After looking at each in turn, we discovered a number of problems with the traditional paradigm. The most substantial of these findings was the realisation that demographic profiling tells us little about a consumer’s identify and values. As the term consumer begins to describe a temporary state of mind rather than a lifelong commitment, identifying macro factors that affect purchase decisions becomes more difficult.
Greater social mobility, globalised brands and greater access to information is changing buying habits worldwide. Shoppers have a greater ability to change their preference dependant on product category, lifestyle and worldviews than ever before. We live in the age of the individual consumer. Traditional demographic segmentation methods are not sufficient to keep with the changing pace of modern culture.
Even advanced models such as MOSAIC that group consumers based on lifestyles are only an approximation of behaviour. As data analysis & market research take leaps and bounds forward, it is slowly becoming possible to understand the people behind the numbers better than ever before.
Social and Community Influence
The model of social and community influence borrows heavily from the modern explanation of media communications. The basic premise is that audiences are segmented dependant on their ability to influence other audiences — therefore creating a natural hierarchical structure. Industry commentators, influencers, media representatives and senior employees make up the smallest group. Brand messages, ideals and values are filtered through this group to commentators, management and others with public influence. Finally, the final (largest groups) are the active public and passive public.
This is a simplistic example of how the theory can be applied. In practise, companies can define upwards of 10 vertical segments and even create grids of influence based on product or category interest. The power of this form of segmentation lies in the fact that it can be integrated with real data sources, thus grouping consumers by brand interaction rather than predicted likelihood of interaction. In addition, it aids in identifying and targeting audiences based on their relative engagement, sentiment and product interest.
In truth, not all customers are created equal. It is an uncomfortable truth that many are loathe to admit. But accepting this fact, and targeting customers. consumers and shoppers differently based on that understanding is the key to successfully increasing brand engagement across all demographics.
Time & Event Based Segmentation
The second method of segmentation we have chosen is the time and event based model. In practise this is difficult to achieve and is only now being slowly adopted. Digital media bridging the gap and making this more of a reality for brands. In particular, event based segmentation is particularly useful when applied to digital TV viewership, social media listening, geo-fencing and online only events.
Perhaps the best example of event based segmentation is provided by the 2015 Superbowl, where advertisers targeted audiences by geographic location, and viewership of the event. Taking this one step further, it is possible to track attendance or viewership of particular ‘events’. The future of television and digital media advertising, for example, hinges on event based segmentation. Smart filtering will show adverts to viewers dependant on which adverts they have previously seen. Such technology allows for continuous storytelling and brand building without repetition.
Another example of this in action is HubSpot’s smart content, which displays different content to website visitor’s dependant on any number of pre-defined factors. From age to country, lifecycle stage to previously viewed content. The challenge for brands that use event based segmentation is to cross the digital divide and translate these ideas into the physical world.
Finally, the most controversial alternative to traditional demographic profiling is individualised segmentation. Though it is the most difficult to achieve, for businesses that are able to harness this model, the rewards are the greatest. Individual segmentation harnesses all sources of data collection and research available to a business to build a unique, tailored profile for each consumer. From web page and email interactions, to in-store purchases, customer service enquiries and more. The core challenge of building these unique profiles is compiling every data source into one profile.
But once the data is there, the next step is to build machine learning algorithms that search through the tremendous amount of information and find common trends. In this way it is possible to analyse patterns of behaviour that transcend demographics. Instead patterns are based on the exact data gathered, unique to each individual. Patterns may be identified by previous purchase history, email interaction, customer service usage — or any number of factors.
Tesco were the first company to pilot an individualised segmentation approach. The Clubcard system became a central hub of data, where transactions, promotions, in-store activity and online activity could all be stored. In fact, this system gave the retail giant so much data clout in the data science industry thatDunnhumby was formed — a data science and analysis agency owned wholly by Tesco.
None of these alternatives are perfect — they are very much still in development. However, demographics are no longer a perfect measure of markets either. We are in a state of flux as we continue to develop and innovate towards a more integrated model of segmentation. There are, of course, risks involved in adopting new methods. But the reward is a better understanding of your consumer, how they interact with your brand and how to remain close to them throughout life.
Have you experimented with any alternatives to demographic profiling? What have been your experiences — has it worked for you, or is the technology still a work in progress? Let us know in the comments below and join the discussion.