Writing in McCall’s magazine about her son’s resistance to the Vietnam conflict in 1966, American author and poet Charlotte E. Keyes tweaked a line from a Carl Sandberg poem and posed a universal question: “Suppose they gave a war and no one came?”. It became a rallying point for Vietnam ‘peaceniks’ and Keyes and her son Gene were branded communists and draft dodgers.
I know how she feels.
In 2008 I was branded a media terrorist for daring to suggest that, by creating sub-sets or extensions to the Living Standards Measure (LSM) 7-10 bands, the South African Advertising Research Foundation (Saarf) had de facto created a 14 LSM model. Well, they’ve been doing it again.
As from the All Media and Products Survey (Amps) 2012AB, further extensions have been applied all the way down to LSM4, giving media planners a de facto 17 level split within the LSM template. Now, in the past Saarf suggested I desist from making such misleading observations because the industry will be confused. So let me just test your level of comprehension.
Is there anyone who doesn’t understand the following statement?
‘Once upon a time Amps sub-divided the South African market into eight socio-economic segments. Now Amps divides the market into 17 socio-economic segments.’
Anybody who is confused by it?
You see, as long as I call them segments, everybody is happy. But, when I say that Amps divides the market into 17 LSMs, then apparently it becomes incomprehensible.
Now I’m sure that, other than dilettantish academic interest, there must have been a reason for this additional segmentation, even if it is not immediately apparent. This reason has not, however, been fully articulated to the industry. I’m assuming the reason these segments were created is so that planners can use and extract value from them.
Naturally, the LSM bands are not airtight pockets and serve to bring together groupings of people from the total population into a continuum of contiguous and sometimes slightly overlapping groups. The good news is that the 17 LSM splits (sorry, segments) invite and allow for a more spontaneous and logical
reclustering of the overall bands into functional market and media segments. One such approach to clustering the LSM bands is found in my revised ‘Muller Cluster Model’.
Based on a mid-point analysis using median household incomes and tipping points in media consumption affinity, such as the transition from SABC radio’s African language stations to other independent commercial radio formats, or the shift from Free to Home (FTH) TV to Pay TV, the model recognises six primary market clusters, each with very distinctive media consumption patterns …
LSM 1-3: Traditional (11.9% of population)
LSM 4-5: Transitional (30% of population)
LSM 6: Middle (22.6% of population)
LSM 7-8 Low: Upper Middle B (15.9% of population)
LSM 8 High-10 Low: Upper Middle A (16.7% of population)
LSM 10 High: Elite (2.8% of population)
This is not to suggest that ‘The Muller Cluster Model’ should be seen as some sort of panacea. It is not an invitation to abandon the core criteria that would normally underpin the target market description. The model does however exhibit a great deal of elasticity and sensitivity in media audience selection and has the additional advantage of improved sample sizes (with the exception of the elite market).
Is clustering data a sound media practice?
If you are using phrases like ‘LSM 4-7’ or ‘LSM 7-10’ in your marketing plans, then you are already effectively clustering the data. The vast majority of TV planning and buying in Mzansi is done using LSM clusters with demographic overlays. My concern is that these clusters are merely assumed historical conventions with no logical foundation in the data itself. The market has moved on and smart media strategists are not only trying to move with it but actually get ahead of it.
On the other hand, you could carry on talking about 10 LSMs. After all, what’s the worst that could happen if they created a market segment and nobody came?