The pace at which tastes, preferences and patterns of consumption are changing is quite bewildering to everyone, so much so that marketers are having a hard time understanding the underlying dynamics of the ever-evolving consumption landscape. To be fair to them, they have over the years experimented with various innovative approaches, albeit with limited success, to effectively segment consumers and make sense of the multi-headed Indian consumer. However the question remains: Can one, in the land mass that is India, home to countless cultures, subcultures, traditions and tastes, really rely on a single segmentation model or matrix If not, how does one then effectively segment consumers
However, before addressing this question, one must first understand the underlying rationale for segmentation. Companies need to understand several features about their consumers. Who are their consumers and where do they reside What are their tastes and preferences
How much are they willing to pay Such questions can be effectively answered only if the eventual segmentation results in successfully categorising consumers with similar purchasing power, almost homogeneous tastes, preferences, lifestyle habits and consumption patterns.
A central problem facing us is the absence of reliable data on various facets that would aid in a reliable and effective segmentation process. Income or purchasing power, which is of foremost importance, is extremely difficult to measure. Lack of reliable datasets projecting income at the national level complicates matters. In order to circumvent this problem, marketers have relied on various other parameters such as education and occupation to serve as proxies of income.
This led to the construction of the socio-economic classification (SEC), which has been widely used by industry over the years. This method of classification essentially relies on variables such as education, occupation, type of dwelling and (more recently) assets of segment households. Although these variables are relatively easier to collect as compared to income, needless to say the designed matrix too has its drawbacks. First, although the idea of using assets is interesting, ideally a more advanced weighted index should be utilised as opposed to a simple summation of assets owned. Secondly, the matrix relies simply on ownership of the asset, failing to account for price differentials in each asset category. Assets such as ACs, TVs and cars have extremely steep price ranges. So clustering households only on the basis of ownership of a said asset is too simplistic. Thirdly, different combinations of different assets can place a household in the same segment, which may not be ideal. Last but not the least, there could be serious stability issues in this definition going forward. Mix of durables will keep changing continuously; fast-changing penetration in durable ownership can render the system very unstable. These critiques should not be misunderstood as an attempt to nullify the usefulness of the classification; they just point out its limitations and prevent it from being viewed as a panacea for our troubles. This segmentation at best presents a stopgap arrangement, till a more comprehensive, reliable and robust system can be put in place.
While not questioning the validity of using education, occupation and assets, one could ask whether a behaviour-stage-based segmentation will do a better job of predicting consumption and provide more intuitive segments for marketing and sales organisations. Rather than forcing behaviour to fit into predefined demographic segments, the behaviour-stage segments will potentially accurately and intuitively mirror the ways in which household consumption shifts over the members lifetimes. Additionally, variables measuring tastes, preferences could also be added for finer profiling. A segmentation scheme combining both psychographic and socio-economic attributes could identify, personify and depict the typical lifestyles of consumers more accurately and effectively than just using any one of these segmentations in isolation. By no stretch of imagination are we suggesting here an unmanageably complex design architecture to construct the desired classificatory system. But suggested additional dimensions can be incorporated rather creatively through fusion-like techniques without the unnecessary burden of additional data collection.
While psychological characteristics explain consumer behaviour in microcosmic detail, an effective social grade can explain consumers from a macroscopic view. These two separate segmentation concepts once put together in an innovative fashion have the potential to differentiate interests, opinions, attitudes, choice of media, decision-making processes and purchase behaviour of consumers incredibly effectively. So, while we may first like to prioritise the creation of an improvised version of the new socio-economic classification, we want to make sure that we are all aligned in our end goal, which in our opinion should be to carve out this dream segmentation by undertaking a more inclusive value-added analysis, using all possible indicators. The onus is on all stakeholders to actively participate in this endeavour and collaborate passionately and effectively to fulfil this elusive dream. Given the proverbial diversity in India, we do not hesitate to ask whether a one size fits all approach is the right way to go. Admittedly, adopting multiple segmentation matrices is confusing but the question nonetheless remains.
Rajesh Shukla is an independent consumer economy expert and Bikramjit Chaudhuri is senior VP, MSci, Nielsen. Views are personal