By Dr Paul Haupt
The SAARF LSM or SAARF Living Standards Measure, developed by the South African Advertising Research Foundation (SAARF) has become one of the most widely used marketing research tools in South Africa. Paul Haupt, MD of SAARF, comments, "Unfortunately, it has become so relied upon that it's very often being misused and has therefore become the victim of its own success".
Developing the LSM
To understand what can be achieved by using the SAARF LSM as a marketing segmentation tool, you need to know why it was created. When thinking about any population, it is important to understand that although people are very diverse, they do have certain commonalities. What is required from a market segmentation tool is to create an index that will differentiate between people with different behaviour patterns and group together those people with similar behaviour.
The development of the LSM Index was stimulated by a series of events; the most important being that the then commonly used market segmentation tool of 'urban/rural' was losing its power as a differentiator. The gap between the urban and rural markets was narrowing and the habits of both markets were becoming increasingly similar.
To counter this, SAARF introduced a new grouping according to community size, namely "metropolitan", "cities/large towns", "small towns/villages" and "rural". This was an attempt to further segment the market by level of sophistication so that marketers could better define their markets. It was just another demographic however, and it was soon realized that what was needed was a set of descriptors based on more than one variable, which would provide a more powerful segmentation tool than any single variable on its own.
At the time, the late Eddie Schulze had been working on a system of classification for Unilever, based on whether people shopped at a supermarket or not, and then going further, to split the supermarket shoppers into those who owned commodities such as cars, television sets and radios. From this thinking, emerged South Africa's first multivariate market segmentation index - the LSMs, which in addition was also a move away from segmentation based purely on demographics.
The LSM index was designed to profile the market into relatively homogeneous groups. It is based on a set of marketing differentiators which group people according to their living standards, using criteria such as degree of urbanisation and ownership of cars and major appliances (assets). Naturally, the LSM bands are not airtight pockets. LSMs bring together groupings of people out of the total population continuum into contiguous and sometimes slightly overlapping groups. Essentially, the LSM is a wealth measure based on standard of living rather than income - in fact, income does not appear anywhere within the LSMs at all. Interestingly enough, variables such as income, education and occupation were tested as part of the first LSM but did not add anything to the strength of the measure.
In the new SAARF Universal LSM®, the population continuum is divided into ten groups, from 1 at the bottom end, to 10 at the top. The LSMs are calculated using 29 variables taken directly from the SAARF All Media and Products Survey (AMPS®).
2001 SAARF UNIVERSAL LSM® DESCRIPTORS
|1||Hot running water||16||Less than 2 radio sets / household #|
|3||Microwave oven||18||Rural outside Gauteng/W.Cape #|
|4||Flush toilet in/outside house||19||Built-in kitchen sink|
|5||No domestic in household||20||Home security service #|
|6||VCR||21||Deep freezer #|
|7||Vacuum cleaner/floor polisher||22||Water in home/on plot #|
|8||No cellphone in household #||23||M-Net/DStv subscription #|
|9||Traditional hut||24||Dishwasher #|
|10||Washing machine||25||Electricity #|
|11||PC in home #||26||Sewing machine #|
|12||Electric stove||27||Gauteng #|
|13||TV set||28||Western Cape #|
|14||Tumble dryer #||29||Motor vehicle in household|
# The 14 new variables that appear in the 2001 SAARF Universal LSM®
SAARF 2004 SU-LSM Descriptors
Four variables of the 2001 LSM descriptors were excluded and four new variables were included, indicating development and a changing market place i.e.
|Traditional hut||House/cluster house/town house|
|Western Cape||1 cell phone in household|
The full list of AMPS 2004 LSM descriptors are as follows:
1. Hot running water
3. Microwave oven
4. Flush toilet in house or on plot
5. VCR in household
6. Vacuum cleaner/floor polisher
7. Have a washing machine
8. Have a computer at home
9. Have an electric stove
10. Have TV set(s)
11. Have a tumble dryer
12. Have a Telkom telephone
13. Hi-fi or music center
14. Built-in kitchen sink
15. Home security service
16. Have a deep freeze
17. Water in home or on stand
18. Have MNet and/or DStv
19. Have a dishwasher
20. Metropolitan dweller
21. Have a sewing machine
22. DVD player
23. House/cluster/ town house
24. 1/more motor vehicles
25. No domestic worker
26. No cell phone in household
27. 1 Cell phone in household
28. None or only one radio
29. Living in a non-urban area
Setting the record straight
Some users still think LSMs are the 'polite' way of talking about race. However, this was never the aim and race has never been used as part of the LSM at all. It is true that initially LSMs did correlate highly with race, with most blacks falling into LSM 1 to 6, and the higher LSMs being multiracial. This, however, is because LSMs reflect the reality of South Africa - they weren't created to obscure it but to reflect it. As South African society corrects itself post 1994, race's impact, as a differentiating variable, will decline, as will its correlation with LSMs, as is already evident.
Another misconception is that LSMs can be used as a psychographic or attitudinal measure. LSMs can tell a marketer that those in LSM 10 for example, have more commodities than others. It doesn't tell you their income, or whether they are predisposed towards spending money. To say that a product is being targeted at LSM 10 is to miss the point.
The SAARF Universal LSMs® are thus not an alternate label for income. Income is actually very often a misleading variable on which to base a marketing strategy, especially if the predisposition of the person towards spending is not known. A miserly businessman who earns a fortune may sleep on a mattress on the floor and warm up TV dinners in the microwave. A pensioner may have all the "modcons" which place her household into LSM 10, but whether she has a high disposable income or a propensity to spend is not a given.
SAARF regularly receives phone calls from people who say 'My target market is LSM 9 and 10 - please tell me who they are?' This is enough to make you cry! The SAARF Universal LSM® should not be used in isolation. Human beings are much too complex to be described using a single differentiator such as LSMs. Users of AMPS data know that when combining LSMs with other descriptors such as language, income, life stage and so on, powerful segmentation of the market can be achieved.
Some users also complain that the groupings are too wide. They don't realise that there's nothing stopping them from re-segmenting the LSMs to suit their markets. They could further segment LSM 6 for example, just as SAARF used to split the old LSM 7 and 8 into sub-groups to help users who found these groups too large to handle.
Another perceived problem is the weighting system. From 1991 to 1993, LSMs were based on 13 variables, which were given different weights each year. For instance, not having electricity would be an important variable one year and less important the next. In 1993 it was decided to standardise the weighting to allow for year-on-year trending. In 1995, the variables were increased to 20 and included limited personal data, which necessitated another weighting system. The 95 LSMs were kept the same until AMPS 99. For AMPS 2000A, the factors describing LSMs changed again - hence 2000 was a new benchmark year for LSMs and an intermediary step to the new SAARF Universal LSM.
Whether the 1993, 1995 or the 2000 weightings are used, the LSMs are no more or no less useful. They are all accurate but different from each other. Users must just be aware that they can't do trending from the one set of weightings to the other.
A large part of the marketing and advertising industry tends to use LSMs as the alpha and the omega of determining target markets. Of course it would have been nice if the LSMs could have been as important as that - as unequivocal. But the reality is that they are only one possible descriptor of market groupings, and to use them as the be-all and the end-all of market segmentation does more damage than good to the tool. In short, as with any research, a bit of common sense is called for.
Bigger, better SAARF Universal LSM®
Following extensive analysis of its much-used LSM during the latter half of 2000 and the first half of 2001, SAARF decided on the creation of an improved Living Standards Measure and simultaneously with the release of AMPS 2001A, launched the new SAARF Universal LSM®.
SAARF decided to make these improvements because the LSM used for AMPS/RAMS and TAMS were not comparable and this had lead to much confusion in the past. There was also a conflict between the way the original LSM was devised and the need for such a measure to be able to show trends and be sensitive to societal changes. The LSM variables from 1995 onwards included personal variables, which were questioned by the industry because it introduced certain biases.
Following extensive analysis of the LSM, the joint SAARF Councils voted to implement the new SAARF Universal LSM® as from AMPS/RAMS 2001A, based on universally applicable variables only. This means that all respondents can answer all the LSM questions leading to a new universal index applicable to all adults of age 16+, without introducing a bias. An example of this would be the bias created by the old “supermarket shopper” variable, which sometimes lead to a husband and wife not being in the same LSM category.
The SAARF TAMS panel implemented the SAARF Universal LSM® simultaneously, which means that whether you use AMPS, RAMS or TAMS now, the same LSM will be used throughout, bringing to an end the confusion of the past.
What is the SAARF Universal LSM?
The new SAARF Universal LSM® is similar to the older version, but starts out with 10 groups. As South African society develops, the SAARF Universal LSM® has the ability to be extended beyond group 10, and 11, 12, etc. will be added as time goes by.
Out of the original list of 20 variables used to determine a person’s LSM category in the previous LSM, 15 household variables have been carried through to the SAARF Universal LSM®. In addition, the total number of variables has been increased to 29 to give finer definition to the scale.
Key Issues to Remember