EXCELlent Ones Consulting Skip to Content Skip to Search

Print | | Download PDF

The Dark Side of Data

 

Data driven.  Our society has increasingly become built upon the knowledge that is derived from data.  In government, policy decisions yield to a poll of the people; in corporate America, analyses are done to determine the most profitable products, in non-profits reports are generated on outcomes and effectiveness.  Dark data are things that cloud the truth in data to be unleashed.  The result is inappropriate decisions, wasted time and resources which can prove fatal for organizations.  In many cases dark data are the result of the human agendas.  Yes, data are typically derived from systems but systems, reports, analysis are all built by humans. In the following sections, we will shed some lights on how lies with data can lead to shadows around data.

 

Sample Size

Data can be deceptive at times.  Often the truth of the data can be clouded or misinterpreted.  It can be done in a few ways.  One approach is sampling.  Sampling is the process of selecting a small group of numbers with the assumption that it is representative of the whole.  For example, if an organization records one million transactions and an auditor wants to review the accuracy of the transaction, checking each one would prove time consuming.  A small slice would be selected.  One always needs to remember that the smaller the sample size the less likely it is representative of the whole.

 

Question Framing

Another way, data results can be misleading is from question framing.  For example, an object may look different depending on the level or kind of light that shines upon it.  In the same way how a question is framed or answered can lead to wrong conclusions.  The first form of question framing is “answering the right question with the wrong data”.  It happens all the time with external and internal information requests.  Always keep in mind people have agendas.  Those agendas can shape and color the information that is provided.  In short, typical people want themselves, departments, initiatives or organizations to look good.  The desire can motivate them to distort information.  One approach is what we call wrong data, right question. For example, if someone asks, show me your financial performance trend?  That question leaves a lot of room for interpretation.  Are we talking about net income (i.e. income after expenses) or gross income? Is the period over 3 years or 5? Will an average or median suffice?  The question is vague enough to leave the one supplying the information with the ability to lean to the measure(s) that will give the best story.   Another approach connected to that request is called the group fallacy.  Many times data like that can be summarized or doctored in slight ways to eliminate discussion or deep scrutiny.  Data are not always summarized for good reasons, but to conceal issues, divert attention or simplify discussions.    

 

Misunderstanding Meta Data

My father use to tell me that if all calculators broke many people wouldn’t know how to add.  There is truth to this statement.  Many people don’t understand the rules and procedures of basic arithmetic.  They have been conditioned to press a button and receive a result without scrutinizing the data.  One of the reasons that data appears dark is due to a lack of understanding.  The byproducts of the wonderful analytical tools that we have at our disposal today is that no one understands them.  People simply press a button and hand a report to their manager.  Analysts, Managers and Directors alike do not understand the meaning of the data they are looking at.  For example, if you are asked to forecast sales data, and you simply rely on the growth rate from one year to the next, the results would be wrong.  Deep dive into the data is needed.  What are the relationships within the data? Were there any large one-time occurrences that need to be accounted for? What does this mean?

If human agendas are not monitored and checked, data will always be dark.  I share these insights to sharpen your analysis of data, that you may be able to shine a light on dark data.

Footer

Email: info@excellentones.com | P.O. Box 160493 | Brooklyn, NY 11216 | Phone: (347) 763-1035 | Fax: (718) 228-7859

All rights reserved. Copyright 2014 Excellent Ones Consulting LLC.