%202.png)
.png)
The Living Standards Measure is South Africa's most widely used way of sorting consumers into tiers, ten groups built not from income, but from what a household owns and where it lives. For three decades it has shaped media plans, retail strategy, and research samples. This guide explains how it works, how to use it well, and why the industry is steadily moving on.
The Living Standards Measure (LSM) is a marketing and research segmentation tool used across South Africa to classify standard of living. It sorts the adult population into ten groups based on their relative means, but it never asks how much anyone earns. Instead it infers wealth from a basket of observable assets and living conditions: the appliances in the home, the type of dwelling, access to water and electricity, and degree of urbanisation. It is, in effect, an asset-based proxy for income in a country where direct income questions are unreliable and refusal rates are high.
What the LSM actually measures
The LSM was developed by the South African Audience Research Foundation (SAARF) in the late 1980s and first published in the SAARF 1989/90 reports, which distilled 13 key variables from an original pool of 71 household characteristics. The logic is straightforward: rather than ask a respondent to disclose income, a question many people answer inaccurately or skip entirely, you ask what they own and where they live, then rank households against each other.
Because it ranks people relative to one another, the LSM is best understood as a set of deciles. LSM 1 is the group with the least means; LSM 10 is the group with the most. Crucially, the measure deliberately excludes income as an input, even though it functions as an income-inequality metric in practice. Its variables reflect the reality of a society with one of the world's highest Gini coefficients, a wide spread of haves and have-nots with comparatively little in the middle.
The ten groups at a glance
The groups run from survival-level households with minimal infrastructure to affluent, fully serviced urban homes. The share of the population in each group shifts with every survey wave, and the single largest concentration has historically sat in the lower-middle of the ladder rather than the extremes. The profiles below are a working summary, not official cut-offs, always pull current figures from the latest establishment survey before sizing a market.
| LSM band | Typical profile | Where they live | What it signals |
|---|---|---|---|
| LSM 1–3 | Lowest means; few durables; often no on-site water or flush toilet | Rural areas and informal settlements | Survival spending; cash economy; basic-needs categories |
| LSM 4–5 | Lower to lower-middle; electricity and basic appliances; modest income | Townships, peri-urban, smaller towns | Emerging consumers; value and entry-tier brands |
| LSM 6 | The pivot group, historically the single largest band on the ladder | Urban and township; mixed | The contested "emerging middle"; high marketing interest |
| LSM 7–8 | Established middle class; vehicle, computer, broader appliance ownership | Urban, formal housing | Mainstream aspirational spend; financial services |
| LSM 9–10 | Highest means; full durable ownership, security, pay-TV, swimming pool | Affluent urban suburbs | Premium and luxury; the top of the income distribution |
A practical caution: the LSM's distribution has long resembled a near-perfect bell curve, which implies a large, comfortable middle class. That picture sits awkwardly against census data, which shows a far more bottom-heavy society. We return to this gap below, it is the single most important thing to understand before you trust an LSM sample.
How the score is built
The current basket runs to roughly 29 variables. They fall into a few intuitive families. None of them is income, the point is to triangulate wealth from things that are easier to observe and harder to misreport.
- 01Dwelling & geography. Metropolitan vs non-urban residence; house, cluster, or town house; whether there is a domestic worker or gardener.
- 02Services & infrastructure. Tap water in the home or on the plot, a flush toilet inside, hot running water, a built-in kitchen sink, a home security service.
- 03Major durables. Refrigerator, electric stove, microwave, deep freezer, washing machine, tumble dryer, dishwasher, vacuum cleaner, air conditioner.
- 04Lifestyle & leisure. Swimming pool, home theatre system, pay-TV (DStv) subscription, motor vehicle in the household.
- 05Technology & communications. Number of cellphones, a landline, a desktop or laptop, a DVD or Blu-ray player, radio ownership.
A household's answers are scored and the household is placed into one of the ten groups. Add a second cellphone or a pay-TV decoder and a household can shift up a rung, which, as critics point out, is part of the problem. A new gadget is not the same as a genuine change in living standards.
How researchers actually use it
For a researcher running studies in South African markets, the LSM is most useful as a sampling and weighting frame and as a shared vocabulary, rather than as a deep behavioural model. Four common applications:
Quota design and sample sizing
Set quotas by LSM band so your sample mirrors, or deliberately over-represents, the segments your client cares about. Because the bands tie back to the establishment survey, you can benchmark your achieved sample against a known population distribution.
Weighting and representativeness
If your fieldwork skews toward easier-to-reach urban, higher-LSM respondents, LSM weights let you correct back toward the true population shape. This is where the bell-curve problem bites: weight to a flawed frame and you bake the flaw into your results.
Retail and channel strategy
The LSM maps cleanly onto where people shop, which is why it underpins so much retail planning. It lets a brand reason about which banners, price tiers, and store formats reach which tier of consumer.
Media and reach planning
The measure originated inside media research, and LSM bands remain a default way to describe an audience for radio, TV, print, and increasingly digital reach, though the durable-based variables make it a blunt instrument for digital behaviour.
How major retailers map to LSM
A frequently cited illustration comes from trade analysis of South Africa's grocery sector, which lines up the big banners against the LSM ladder. It is a useful mental model for where a tier of consumer is likely to shop.
| Retailer | LSM focus | Position |
|---|---|---|
| Shoprite | LSM 1–6 | Middle to lower end of the market |
| Pick n Pay | LSM 1–8 | Broad range, varies by store format |
| Spar | LSM 6–10 | Middle to upper income |
| Woolworths | LSM 6–10 | Middle to upper income, premium |
| Massmart / Cambridge | Full range | Walmart-owned; Cambridge serves the low end |
Source: USDA Foreign Agricultural Service, Understanding the Living Standards Measure Segmentation in South Africa (2020). Retailer footprints and ownership shift over time, treat this as directional.
Where the LSM falls short
The LSM was close to perfect in the late 1980s, when only a minority of households owned an electric stove, a fridge, or a TV. Owning one genuinely separated you from your neighbours. Today the majority, in many cases more than two-thirds, own those same durables, so the variables no longer discriminate the way they once did. Three problems compound this:
It implies a middle class that may not exist
The near-symmetrical bell curve the LSM produces would suit a country like Canada or Australia. South Africa's income distribution is nothing like that. On one analysis, the LSM placed only about 6% of households, roughly one in seventeen, in its bottom three groups, whereas a structural measure put 44% of households, nearly one in two, in the equivalent struggling tier. If you size a low-income market off the LSM, you can badly underestimate it.
Durables became contra-indicators
Some signals have inverted. A DStv subscription once marked an affluent household; falling costs and the rise of streaming mean it no longer reliably does. Technology variables, DVD players, computers, cellphones, change faster than the survey updates, and the LSM has been refreshed only intermittently, leaving the model chasing a moving target.
A built-in age and racial skew
Because durables take fifteen to twenty years to accumulate, and because of South Africa's history, the measure carried a structural bias. By one industry estimate, LSM 9–10, the notional top 16%, was only 38% Black, while the real top 16% by household income was around 60% Black. A measure used to direct advertising spend therefore risked systematically skewing it.
The move to SEM
SAARF was restructured into the Marketing Research Foundation (MRF) in 2017, and the industry introduced the Socio-Economic Measure (SEM) as the LSM's successor. Developed by researcher Neil Higgs and his team with the Publisher Research Council, the SEM keeps the familiar shape, ten segments, but rebuilds the engine.
Where the LSM leans on roughly 29 durable- and technology-heavy variables, the SEM uses 14, and deliberately shifts the weight toward structural and community items that change slowly and meaningfully: floor and roof material, water source, type of toilet, number of sleeping rooms, and proximity to a post office or police station. These move when a household's circumstances genuinely improve, building an extra room or replacing a mud floor with tile, rather than when someone buys another gadget.
| Dimension | LSM | SEM |
|---|---|---|
| Variables | ~29 | 14 |
| Primary basis | Durables & technology | Household structure & community infrastructure |
| Stability | Shifts with each gadget bought | Changes only with real lifestyle change |
| Structure | 10 fixed deciles | 10 segments on a 100-point scale |
| View of poverty | Understates the bottom tier | Aligns more closely with census data |
The SEM also offers more flexibility: because it runs on a 100-point scale, a brand is not forced into "SEM 5–7" but can define a target in precise percentage terms. The two measures ran in parallel for a transition period, and many teams still calibrate between them. The direction of travel, however, is clear, away from counting appliances and toward describing how people live.
Where the data lives now: MAPS
As of 2026 the LSM has not disappeared, it is published alongside the SEM in the MRF's Marketing All Product Survey (MAPS®), the continuous study that replaced the old AMPS and Establishment Survey. MAPS launched in mid-2020 and now offers five unbroken years of consumer and media data; its latest release covers the full 2025 calendar year and a South African adult (15+) population of 45.69 million across all nine provinces. For a researcher, the practical takeaway is that both frames are live and current, you can still pull an up-to-date LSM profile, but you should reach for the SEM (or run the two together) wherever the bottom of the distribution matters.
Using the LSM well in 2026
The LSM is not obsolete. It remains a useful common currency, a sampling frame baked into the major surveys, and a quick way to describe a consumer tier. But it should be used with its limits in mind, and rarely on its own.
Calibrate against a structural measure
Before sizing a low-income or emerging-market opportunity, cross-check your LSM frame against the SEM or census data. If the two disagree sharply at the bottom of the ladder, trust the structural measure.
Use it to segment, not to explain
The LSM tells you roughly how affluent a group is. It does not tell you what they want, why they switch brands, or how they decide. Pair the tier with qualitative and behavioural evidence collected from real consumers.
Layer in first-party behaviour
The richest segmentation today combines a stable socio-economic frame with behavioural and attitudinal data gathered directly, what people buy, say, and feel, rather than inferred from a decade-old appliance list.
This is where a measure like the LSM and a tool like Yazi are complementary, not competing. Use the LSM to define who you are sampling; use conversational, WhatsApp-native research to learn what they think and do, across exactly the lower-LSM, harder-to-reach segments where asset-based measures are weakest.
Frequently asked questions
Does the LSM measure income?
No, and that is deliberate. It infers relative wealth from assets, dwelling features, and access to services. It functions as an income-inequality metric, but income is never an input, which is why it works in a context where direct income questions are often refused or misreported.
How many LSM groups are there?
Ten. LSM 1 is the group with the least means and LSM 10 the group with the most. Researchers often work in "supergroups" such as LSM 1–4, 5–7, and 8–10 for simplicity.
Is the LSM still used, or has SEM replaced it?
Both are in circulation and both are current. The MRF publishes the LSM and the SEM together in its MAPS survey, whose latest release covers all of 2025. The SEM is the more accurate successor at the bottom of the distribution, but the LSM remains widely referenced and embedded in legacy data, so many teams still calibrate between the two.
Why does the LSM understate poverty?
Its durable-based variables produce a bell-shaped distribution that implies a large middle class. Structural measures and census data show a far more bottom-heavy society, so an LSM frame can significantly undercount low-income households.
Can I use the LSM outside South Africa?
Not directly. The LSM is calibrated to South African ownership patterns and infrastructure. Other markets use their own systems, NRS social grades in the UK, NCCS in India, PRIZM in the US. For pan-African work, treat the LSM as a South-Africa-specific frame and build local equivalents elsewhere.
Segment with the LSM. Understand people with Yazi.
A socio-economic tier tells you who. It does not tell you why. Yazi runs conversational, WhatsApp-native research that reaches consumers across the full LSM range, including the lower bands traditional panels miss, and brings back what they really think, in their own words and voices.
Book a Demo →