Friday, February 25, 2011

A New Way For Analytics

A good friend, executive at a large analytics company, asked me how I would articulate the next product in the space. Here is my attempt:

New markets require new analytics products that begin in a new fashion.

A generic and current example would be: We can dissect the consumer population into demographics (have schema), and run campaigns on them (poor distributions, hitting same consumer with many, at times conflicting messages).

New approach... Let's define the demographic we seek, and develop the application around understanding it.

For example, there is no analytics application that comprehends the Muslim Consumer in-depth. The new application has to be incremental in learning, i.e. it never stops learning; it has to comprehend the unseen, i.e. it is schema free, hence, leading to intuitive outcomes; it has to be real-time i.e. no more running the models for days to get an answer; it has to be built upon existing infrastructure i.e. it does not throw away the clients current spending on IT; it has to provide the unification of statistics and semantics i.e. no arm twisting to "try" to make "sense".

There is no such application in the market currently. Such a proposition is where insights into products for families living below $2/day can come from. This is where the analytics engine begins to develop a “memory” of the entity, may it be a consumer or a automotive.

This is NOT aspirational, this IS the “Near Tomorrow” (Copyright 2011 The RBR Group).

The execution requires for example for CPG/FMCG: (1) Understanding a client's goal for new market entry, (2) developing the ideal consumer (no, not digging within the existing, looking back at the past data, it was yesterday), (3) searching for the ideal consumer, (4) begin to consume all data on the closest matches in the method defined above, i.e. schema free, incremental, with statistic and semantic unification.

This delivers the consumer’s unstated needs, where the value is the highest for a product that fulfills it. Example, the consumer had no clue they wanted an iPhone. Keypads, smaller and smaller were fine too!

For the 21st century, analytics is the when predictive goes to forecasting to deterministic.

For a thought provoking look at engine technology that delivers the above, see Saffron Technology.

No comments: