Page 29 - The Real Work Of Data Science Turning Data Into Information, Better Decisions, And Stronger Organizations by Ron S. Kenett, Thomas C. Redman (z-lib.org)_Neat

14 The Real Work of Data Science

Strategic analysis, beyond SWOTs, usually focuses on potential new projects or business
propositions. This can cover a wide scope, such as:
• fit with business or corporate strategy
• inventive merit
• strategic importance to the business
• durability of the competitive advantage
• reward based on financial expectations
• competitive impact of technologies
• probability of success
• R&D cost to completion
• time to completion
• capital on hand and marketing investment required to exploit new opportunities
• effect on market segments
• implications to product categories or product lines.

Data scientists should ask to see these
Sort Out the Value Structure studies (even participate in their creation)
when they are asked to join new initiatives.
An organization’s values play key roles in
shaping its direction, choice of metrics, and
the decisions made by individuals and The Balanced Scorecard and Key
groups. Redman learned this in one of his Performance Indicators
first consulting assignments, with a large To translate vision and strategy into objec-
investment bank. His assignment involved tives and measurable goals, many companies
helping the company sort out its data quality use a balanced scorecard (Kaplan and Norton
program, and he pitched the program as 1996). This dashboard helps managers keep
saving money. But he got little traction.
their finger on the pulse of the business. The
A chance event involving a Super Bowl original balanced scorecard featured four
pool helped Redman see that, even though broad categories: (i) financial performance;
the bank carefully tracked expenses, saving (ii) customers (e.g. customer satisfaction);
money was not high on its list of priorities. (iii) internal processes (e.g. efficiency,
Rather, the bank prided itself on growing safety); and (iv) learning and growth (e.g.
revenue and managing risk. Recasting the morale), and aimed to balance (often short‐
data quality program along these lines term) financial and longer‐term nonfinancial
helped move it along. performance by providing a broad view of
the business. Typically, each category
The vignette illustrates a more general point: includes two to five key performance indica-
it takes more to understand a company than tors (KPIs), customized by each organiza-
studying its formal documents. Of particular tion, based on its strategy, and implemented
concern to data scientists is who makes the in its own dashboard.
important decisions (e.g. senior or more junior The goal is to derive a set of measures
people), how they are made (e.g. by consensus matched to the business so that performance
or by the most senior person), and under can be monitored and evaluated. If the
what criteria (e.g. driving revenue, increasing business strategy is to increase market share
shareholder value, improving customer satis- and reduce operating costs, the indicators
faction, regulatory concerns, innovation, etc.).
may well include market share and cost per