There have been many articles, reports and features over the last few years about the death of demographics, particularly in the media and entertainment industry. It has become the fundamental issue; no longer just a ‘digital issue’.
For years, we have used simple demographic information—age, gender, income level, occupation, education, married with children, singles, retired—to segment and understand consumers.
In the statistical environment, sampling is concerned with the selection of a subset of individuals from within a statistical population to estimate characteristics of the whole population. There are at least nine different methods of sampling: simple random, systematic, stratified, PPS (probability proportional to size), cluster, quota, accidental, line-intercept and panel.
In the media and entertainment industry, sampling is being widely used for gathering information about a population and has remained the single largest source to calculate all efficiency- and effectiveness-related currencies.
At a fundamental level demographic, variables are now slowly losing relevance/utility as a sampling source to estimate media consumption. The reality is that the explosion of media and entertainment choices available to consumers is so overwhelming that it makes any of these methods of sampling vague or blurred. These choices far outstrip the homogeneity these samples can deliver, as each consumer emerges as a unique individual sample.
The volume escalation of mobile devices (from text to feature to smart) globally has given every individual the power to express himself/herself and be constantly connected. These expressions manifest in their online and offline behaviours as an ‘interest’. Consumer’s interests are thus now becoming a major aspect to recognising or identifying who they are and gradually forming part of their personal identity—in both the online and offline world.
If we go back to revisit those nine different methodologies in sampling, evidence indicates that interests are now becoming a stronger variable than demographics in estimating the behaviour of a larger or wider population. Especially true when the attention is towards understanding what they might want to do or buy, where they might want to go, or who they might want to meet or follow.
Each one of us can have an interest graph, which is a representation of the set of specific interests that each of us have. When we take into account our explicitly declared interests, for example ‘Likes’ on Facebook or ‘Interests’ in a LinkedIn profile, as well as implicit interest inferred from our activities such as clicks, comments, tagged photos and check-ins, the interest graph becomes more accurate and expressive.
There are a number of uses for interest graphs in the media and entertainment industry, especially when applied in conjunction with social graphs and the media consumption graphs. There is huge potential to use interest graphs as another form of behavioural profiling and targeting in audience analytics, audience-based buying, sentiment analysis and advertising.
Interest graphs can also be applied to product development, to help determine which new features or capabilities to provide in future versions of a product. They also have enormous potential in content discovery, providing inputs to recommendation engines for films, books, music or anything else, as well as applications for education.
Whilst there can be many configurations of how consumer interests can be tracked and represented, in my experience, tastes, preferences, influences and behaviour online give a unique configuration to link individual consumer interests, (anonymously / permission based) with the rest of the online community.
This approach is providing good results so far in our business and with our clients.
When millions of bits of data across tastes, preferences, influences and behaviour are captured, we can then combine a threshold level of affinity with enterprise data from individual businesses. And when mixed with media consumption data, magic happens.
It is not the 'holy grail', but a reality.
While online social networks are often a source for this data, the future belongs to running offline research as well to understand interests and demographics (or both) and integrating them into a single source.
Understanding the target consumer's interests has always been the primary area of focus for marketers. It’s just that in the past we didn’t have the relevant data available and so have been using demographics as the surrogate to represent individual consumer interests.
The speed at which data is now available is quickly rendering demographics irrelevant as this surrogate. We no longer have to wait for annual surveys to capture consumer interest, and this coupled with the multitude of interest variables available means it is far more relevant and applicable than the actual demographic variable itself.
In my view, interest based targeting will eventually replace demographic targeting as it will provide a far more powerful profile of the target consumer in real time across locations and occasions. Communicating with the ‘most valued consumer’ at the moment of truth and capping the frequency of our communication at the appropriate level is no longer just a dream, but a distinct possibility.
An approach that not only enhances the ROI of the marketing investments manifold but also forces the marketing ecosystem to look at multiple formats of communication depending upon time of day across different screens.
The future is not just interesting—it’s all about interest.
Gowthaman Ragothaman is Asia-Pacific COO with Mindshare