Should Analytics be driven by a dedicated Analytics Organization or should it be ‘Point of Need’ based?
Is Analytics a ‘Top-down mandate driven by executives’ or ‘We have analytics champions all across the organization and they use the tools and solutions that they need to accomplish their objectives’? This is a question we are asked often, and there is no clear answer that has emerged, yet.

We believe that analytics is inherently distributed, and distributed randomly. Some of the factors driving a non-centralized nature are:

  • Departmental needs are different: HR or Finance organization will be different from that of marketing organization.
  • The skill levels of teams will be different. Marketing may have a dedicated team of analytics professionals, while HR may have ‘accidental analysts’.
  • The access to data by these organizations may be different. Marketing data is open to collaboration while HR data, is largely confidential
  • The volume of data available varies (think of 5000 rows/month at ledger level VS millions of rows of click-thru data per week for marketing)
  • Need for IT support may vary
  • Need for tools may vary. It you are a retailer with 100s of stores and millions of SKUs, you perhaps need SAS, but the same retailer may supplement with Spotfire to balance the need for power/cost ratio. Finance may still need the traditional BI tool for pixel-perfect reporting and burst reporting.

Diversity of needs, capabilities and value, will drive a distributed nature of analytics. So you can expect islands of analytics minded people strewn around the company who will be busy improving the operations, cutting costs or finding the next big business opportunity. Culturally tho, the organization needs to drive analytics as a central theme. To build a data-driven organization, executives will have to be on-board and be driving the adoption centrally, much like what P&G does or does. So the effort initiation around analytics should be centralized, but the result could well be distributed.