Page 86 - 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

16

Moving Up the Analytics

Maturity Ladder

Let us say you work for a bank, car manufacturer, or cellular operator. Your company completes
many tasks and transactions designed to meet the needs of its customers and accumulates
written reports, call center transcripts and recordings, engineering specs, financial data,
inventory levels, and many other types of data. What you and the company do with this data
reflects the maturity level of your company’s management approach and analytic capabilities.
Understanding this maturity level helps you do a better job leading data science efforts,
both short and long term. We distinguish between five maturity levels:

Level 1. Firefighting: random reports to be delivered yesterday.
Level 2. Inspection: a focus on descriptive statistics.
Level 3. Process view: modeling variability with statistical distributions.
Level 4. Quality by design: planning interventions and experiments for data gathering.
Level 5. Learning and discovery: a holistic view of data science.

Here we expand on the characteristics of these levels and their impact on data science.
Going up the maturity ladder provides deeper and wider benefits from data and data science.
This parallels the effect known as the “statistical efficiency conjecture.” 1
Let’s start with firefighting. Firefighting reflects a heroic level of maturity of organizations,
so chaotic that people can’t think beyond the short term. Most firefights don’t require much
data analysis, as fires are visible and generate flames, heat, and smells. Firefighters need to
figure out the problem immediately and provide quick‐fix solutions. The work is frantic! Most
companies cannot stay in firefighting mode long. Products and services don’t measure up, and
both customers and employees leave. The data scientist, in such environments, is asked for
reports to be produced yesterday. These are typically shallow and lack insight. They are

1 The statistical efficiency conjecture states that as organizations move up these levels, they become more efficient at
solving problems, offering better products and services at lower costs. This conjecture has been tested with 21 case
studies (Kenett et al. 2008).

The Real Work of Data Science: Turning Data into Information, Better Decisions, and Stronger Organizations,
First Edition. Ron S. Kenett and Thomas C. Redman.
© 2019 Ron S. Kenett and Thomas C. Redman. Published 2019 by John Wiley & Sons Ltd.
Companion website: www.wiley.com/go/kenett-redman/datascience