Being an analytics professional carries with it several occupational hazards.
I tend to find that one of the main ones is the experience analysts have of presenting their findings to leaders, only to be greeted by a cacophony yawns.
It’s tempting to excuse this in one of two ways – either by blaming the leaders or thinking that you yourself need to tell a better story.
But the reality is that this that may not be it.
The problem may also be that the information itself has no intrinsic value.
The idea that information may or may not be useful is straightforward.
It has also been studied formally and you may come across a definition of the value of information analysis – such as “quantifying the expected value of research in reducing decision uncertainty to inform whether a decision can be made based on existing evidence or if additional evidence is required and worthwhile.”
It’s useful for analytics pros to have a good feel for this basic idea, however, it’s rarely useful to try to rigorously quantify the expected value of research.
It’s just too hard to do with any precision to justify the effort.
Instead, just pay attention to these general principles:
- Information is valuable if it corrects what would have been an incorrect decision
- Information is valuable if it reduces uncertainty
- Information is valuable if it prompts a change in behavior
- Information is valuable if it reveals more solution options
Let’s consider some examples of why people analytics results may have little value to managers:
|Analytics Finding||Why Managers Yawn|
|Morale is low in region A||We already knew that.|
|We can speed up hiring with this process||Right now, we have more urgent priorities than speeding up hiring.|
|Seventy percent of our engineers are extroverts||Who cares? Why does that matter?|
|Structured interviews have good predictive validity||We are already using those.|
On the other hand, these analytics findings could be helpful
|Analytics Finding||Why Managers Pay Attention|
|The cost of terminations if we close down that plant could be $400k-$1 million||We had no idea what the number might be…you have reduced our uncertainty.|
|We had a chance to interview four employees at the plant and they think it’s unlikely the proposed automation will lead to a strike||Thanks, it’s not a lot of data but this is a top priority, and anything helps.|
|The groups that had safety training do not have fewer accidents||Okay, we better stop that training program and do something different.|
|Instead of trying to improve training, we could make changes to the machinery, stop making that one troublesome product, or outsource the work to a specialist firm.||We were stuck too narrowly on training options; these other options may be better.|
One example John Boudreau, professor Emeritus of management and organization and a senior research scientist with the Center for Effective Organizations, University of Southern California, recently shared with me revolved around assessing the productivity of teams.
Our baseline assumption is that most teams have acceptable productivity.
If our study shows that a team has acceptable productivity, then we take no action – hence that information has no value.
Information that a team’s productivity is unacceptable will lead to action so that information is of value.
The implication is that rather than trying to accurately measure the productivity of every team, we might take on the simpler question of looking for hints or indications that a team might have unacceptable productivity and only do a deeper analysis there.
The concept of the value of information leads us back to the first principle of analytics: start with a decision that needs to be made or a question that needs to be answered.
If you start there, then any relevant information you glean ‘will’ grab managers’ attention.