Planners are flat earth thinkers when it comes to demographics
The Financial Times has a good quality reader survey that got me thinking. In social science research, and in the applied end of the practice in advertising and marketing, there’s an entrenched interest in demographics. It could be argued that media planning in particular is a blend of demographics, spots and frequencies. But demographics only take you so far. Planners would do well to note that the science of demographics has shifted far away from the JICNARS structure which still acts as the template for most demographic analyses in the media, and work by industrial sociologists such as John Goldthorpe is a major advance on how we should think about the labour and education parts of demographics.
But planners and their paymasters are reluctant to embrace new concepts in demographics, and I think this results in a big gap in our understanding of what happens in people’s heads and hearts when they interact with media. The FT survey is a simple enough mix of demographics and job role questions, but adds a little more nuance than is typical in its focus on why the respondent reads different parts of the FT, and how they would describe the FT.
Digital metrics media are a litmus test of old style demographics, as the mass availability of the web and smartphones means that in the developed world at least, demographics are becoming a progressively weaker predictor of usage and perception of encountered media. Planners will have to work long and hard to think about measuring usage and real-time perception of media.
Posted by Ken.
motiveworks 
Definitely agree that the social classifications we all commonly use are outdated – as I recall, Priests were ranked at the top of the AB segment!
And for some categories, segmentations based on attitudes, lifestyles or other variables outside of demographics are much more appropriate eg. your technology attitudes as a determinant of what phone you’ll buy rather than just gender or income.
But when you do an attitudinal segmentation, more often than not, you’ll find the attitudes linked directly to demographics.
So younger, more affluent males are always going to disproportionately make up your “technology leaders” attitudinal groups.
In an ideal world, you’d layer on attitudinal measures on top of demographics.
But that said, I’ve found that basic demographics still have consistently been one of the strongest and useful ways of segmenting people’s behaviour – measures like gender, age, lifestage, earnings and education.
I think we’re in agreement here. My sense is that how we think about earnings and education needs to change. So the Erikson-Goldthorpe schema looks at how much autonomy you have at work, how your destination class can be as important as your origin class. People working on internships, for example, have zero money, but don’t have a working class mindset linked to piece rate work. Third level education is now the norm for about 65% of school leavers, as about 20-od per cent about 30 years ago.
I’d like to think we will start to look at proper statistical models, including structural equation models and regression models, of how much of preferences we can explain based on demographics, and how much on transient things like mood, versus newer workplace variables such as role autonomy and predictability of work.