Chapter 7. Making Sensible Decisions Using Data:
A Responsibility of All Executives

Making decisions based on data is essential to an organization. The organization should define and collect objective, accurate, and current data, which will describe clearly what the current situation in fact is, instead of falling into the trap of what one thinks the current situation is. Defining and collecting high quality data There are two general types of measures: in-process and outcome. A clear distinction must be made which measures are used to analyze the context. In-process measures are those that help anticipate what or/and how much is necessary to manage the process, whereas outcome measures are more financial. Because it is the company’s objective to identify the actual situation by collecting accurate and relevant information, it is highly recommended that for each data element collected there will be a clear, useful written definition (called operational definition).

Analyzing quantitative data Nowadays most quantitative analytical techniques are highly automated and provide fast and accurate capabilities (various software). The following are the techniques used to visualize the data: * Simple graphs and measures * Simple statistics, probability, and uncertainty measures (confidence intervals, etc.) * Trends and patterns (Histograms, Pareto Charts) * Time series forecasting (moving averages, exponential smoothing)

The Kloppenborg and Laning (2012) make an example of decision analysis technique, where decision maker is faced with several alternatives and future uncertainties or risk associated with them. The logic behind this is that it is very helpful to decompose problems into sequential steps (decision alternatives) using a decision tree form. Then based on analysis we choose the alternative that gives a largest expected value. The key part in this process is the ability to identify the possible alternatives that describe future uncertainties when formulating the problem.

Analyzing qualitative data Qualitative data comes in many flavors (ideas, observations, textual responses to questions, etc.), and the value is not a number that has a physical meaning. There are several techniques to attain such information for further analysis. Some of the techniques are * Surveys * Interviews * Affinity Diagrams * Interrelationship Diagraphs * Delphi Technique * Multi-voting

Surveys
Surveys are used to create useful sample information about the larger population observed. Surveys need to be representative and have a large enough sample size to be useful.

Interviews
Interviews can be a useful tool in talking with key decision-makers to understand what key issues they believe need to be looked at when examining business situation.

Affinity Diagrams
This technique is used when large numbers of complex ideas, opinions are in the table. Both logical and creative forms of thinking may give an opportunity to an effective team to overcome the obstacles of unrelated issues, which appeared unrelated at first.

Interrelationship Diagraphs
This is a graphical technique that analyzes the cause and effect relationships between a problem and the factors that may be causing it. The technique helps to identify the root causes/key drivers relating to an issue.
Project organization structure
Decisions & Committees
Project MGMT & planning
Project data analysis
Portfolio MGMT
Project organization structure
Decisions & Committees
Project MGMT & planning
Project data analysis
Portfolio MGMT
Standards
Standards

The disciplines that have most number of arrows out are the key drivers (Decisions & Committees), and the ones that have most number of arrows in are the effects (Portfolio Mgmt in this case).

Delphi Technique This technique is a structured and interactive communication process with a