Database Reference
In-Depth Information
creating a Data Science initiative that is not only efficient but also has a significant voice
in the decisions derived from fresh-breaking BI.
To be truly data driven, the enterprise must embrace a culture wherein the very first
question asked before decisions are made is “What does the data say?” (The shrewd exec-
utive will also be asking key follow-up questions such as “From where was this data de-
rived?” and “What kinds of analyses have been applied?” and “How confident are we in
these results?”) As well, to be truly data driven, the enterprise must include in its culture an
environment wherein top decision-makers are not only willing to be overruled by data, but
embrace fact-based changes in business plans.
Top management of course still remains at the helm of the enterprise. They chart the
course, steer the vessel, and thus define in what waters the enterprise shall be making
its way. Thereby they also define in what areas of inquiry Data Scientists must focus
their attention. There is no substitute for this domain expertise, this knowledge of where
the biggest profit opportunities (and often the most treacherous shoals and currents) lie.
Thus the value of the HiPPO kingpins will morph from being seat-of-the-pants dictators to
simply knowing what questions to ask. They'll be the proofs of Pablo Picasso's old admon-
ition: “Computers are useless. They can only give you answers.”
As in all things, leadership is critical when it comes to the practice of effective Data
Science. Clear goals must be set. And they must be set by domain experts who understand
how a market, discipline, or environment is developing, who can think in new ways and
ponder novel solutions, who can embrace compelling new ideas - and can do all these
things while at the same time balancing the needs of all stakeholders, including customers,
stockholders, and employees.
Data Science demands practitioners with skill-sets not native to most enterprises - not
even most IT departments. In particular we are talking about integrating statisticians (most
of all statisticians skilled in the social sciences rather than business applications), program-
mers skilled in new non-traditional software such as Hadoop for cleaning and modeling
unstructured data, and professionals skilled in predictive analysis and data visualization.
These professionals - whose techniques along with their data are largely unstructured -
occupy a previously undefined place on the map of the enterprise, one which lies in a one-
time no-man's-land somewhere midway between the IT and marketing departments.
With these new professionals must come new chains of command, work-flow-prac-
tices, and procedures. Some toes will be stepped on. IT professionals and marketing profes-
sionals will lose some measure of autonomy. A new and expanded atmosphere of collabor-
ation - both in practice and attitude - must be made to prevail. Most importantly, decision-
making-rights must be adjusted and refined to reflect a shared collaborative environment
