Business managers occasionally complain that information they are presented with isn’t actionable. This usually pisses off the data analysts who spent a great deal of effort acquiring data and turning them into information.
Data analysts will tell you that actionable information is:
- Trusted — Based on credible data, generated in a way or obtained from a source that provided adequate quality checks.
- Representative — Based on enough of the data elements and observations/subjects that are relevant to the problem to establish accurate and precise findings.
- Time Sensitive — Based on data from an appropriate time period, whether it be long-term, current, or a specific period.
- Accurate — Neither the data nor the analysis are biased in any way.
- Precise — Consistent, having a known, controlled, and minimum variance.
- Qualified — Variance is reported as uncertainties and risks.
Data analysts believe they build these characteristics into all the analyses they do. They take good data processed through a good analysis to produce good information. Consequently, saying the information from their analysis is not actionable is calling their baby ugly.
Decision makers have a totally different perspective on actionable information. They are consumers not producers. They don’t even consider the characteristics that data analysts accept as the foundation of actionable information. They’ll tell you that actionable information is:
- Interpreted — Not just compiled, graphed, and reported.
- Relevant — Usable for making a specific business decision.
- Understandable — Presented in a clear and convincing format (i.e., pie charts).
- Indicative — Translatable into possible actions for consideration.
- Predictive — Forward-looking, able to set a new course.
Salesmen recognize this dichotomy as the difference between features and benefits. Manufacturers, like data analysts, build all kinds of features into their product. Ultimately though, consumers, like decision makers, are only interested in benefits. That’s why salesmen sell benefits, not features. Still, without the associated feature, there would be no benefit at all.
So, the definition of actionable information depends on whom you ask. All trusted, representative, time sensitive, accurate, precise, and qualified information is, in one sense, actionable. In another sense, it is actionable if it is also interpreted and presented in a relevant and understandable manner so that decision makers can translate the information into actions that will allow them to effect change. That puts the burden on both data analysts and decision makers. But, just as a certain benefit may not be of value to a particular consumer, information actionable to a data analyst may not be actionable to a decision maker.
The definition of what information is actionable is usually taken from the perspective of the decision maker, the consumer of the information. With that prevailing notion, though, also comes some responsibility. Decision makers must be clear on what their goal is. They can’t just grab a data analysis off the shelf and expect it to suit their need. They must confirm that they are the ones who should take the action, and if so, judge whether the action is even worth taking. Not agreeing with the information or not wanting to make a decision the information indicates, however, is not cause for declaring that the information isn’t actionable. The inability to use information to make a decision may rest at the feet of the decision maker, not the information itself.
Lawyers don’t understand any of this; anything actionable goes to Court.
Read more about using statistics at the Stats with Cats blog. Join other fans at the Stats with Cats Facebook group and the Stats with Cats Facebook page. Order Stats with Cats: The Domesticated Guide to Statistics, Models, Graphs, and Other Breeds of Data Analysis at Wheatmark, amazon.com, barnesandnoble.com, or other online booksellers.
I think this goes hand-in-hand with complaints about “not learning anything NEW” from an analysis. Sometimes all the data can do is quantify something the client already knows is a problem, but not suggest a remedy. Also, clients don’t always understand that a negative finding (a finding that says something didn’t change, or something isn’t correlated with something else, etc.) *IS* still a finding. It’s just not necessarily the one the client was hoping for.
Reblogged this on chrisbeeleyimh and commented:
A brilliant post on the conflict between the needs of analysts and decision makers… with cats! Genius post on a genius blog
For decades I observe very different relation to data from medical and demographic communities. Having some political experience (pro/anti-choice debate) I could say that for both sides info means near nothing.
Pingback: All Information is Actionable – chrisbeeley
Pingback: Ten Ways Statistical Models Can Break Your Heart | Stats With Cats Blog
Pingback: It’s hard to be a data-driven organization – Big Data Made Simple – One source. Many perspectives.
Pingback: It’s hard to be a data-driven organization – Big Data Made Simple