- “This mirror will give us neither knowledge or truth.”
So says Dumbledore in J.K. Rowling’s book, Harry Potter and the Sorcerer’s Stone, commenting on a mirror that shows us what our most desperate desires want us to see.
This is an apt analogy when describing the analytics available in big data solutions. When you suddenly have all the data you could want and can quickly analyze it anyway you like, unencumbered by extraneous effort that we have historically had to endure, what happens? Being human beings with a tendency to confirm what we so want to have happen or to relive what felt so good in the past, managers often drift into self-sealing and circular analysis that at first doesn’t seem so wrong. Big data has to poke through the subtle and instinctual responses of data denial.
When presented with uncomfortable data, I have seen many leaders (including me), instinctually grimace and say the following to those who bring the message:
- “That can’t be right!”
- “There is an error in the data throwing this off.”
- “You can’t look at the data that way, that’s unfair. You have to look at it this way.”
- “Let me show you the right way to analyze this.”
- “You shouldn’t share this. This analysis can be used for the wrong purpose, hurting others.”
- “I need to analyze this myself, with my team, not yours.”
- “Why are you looking at this data? Look at the other units who are having bigger problems.”
While many things relegated to the category of “politics” give rise to these sorts of questions, this type of behavior is natural and is not necessarily a bad sign. All people want to believe and be perceived as effective individuals in control of their environment and circumstances. We all will incline to “control the message” to ensure this perception is upheld. This behavior is part of our human DNA. The types of aspirations behind these responses include:
- “But we are doing good things in our unit.”
- “If you were to spend time with us you would see that the way we are doing this is correct.”
- “I want to return to a state of control and performance I had in the past, which was much better.”
- “We are working hard and are highly capable people.”
- “If only I could get things done my way, all this would improve.”
Unfettered access to data and effortless analysis can lead to bringing data together that confirms our desires. With big data systems we can have big analysis units spending big amounts of time on self-congratulatory analysis forever. If not that, it can enable ongoing analysis that might be more of a “data for data’s sake” or analysis addiction that looks like analysis paralysis. When you can freely exercise the analysis behavior and the environment makes it pleasurable, you can be staring into the Mirror of Erised.
I am not so deluded to believe that access to rich, fast, and big data alone will suddenly improve things. On the contrary, it is likely to bring out more of our human nature, for better or for worse. The good news is that most organizations still have enough frustrations with accessing and analyzing data that most would prefer other activities. But for those experiencing the benefits of better analysis, it might be time to consider some healthy analysis behaviors. These include:
Ensure cognitive diversity. Make sure the teams doing the analysis bring different perspectives, different backgrounds, and different ways of analyzing data. Make the teams aware of their diversity and work with the teams so they can appreciate this diversity. From this diversity comes closer scrutiny of data and assumptions that, if left unchecked, can lead to dysfunctional results.
Make conflict productive. Rather than trying to douse out conflict or negative emotions surrounding uncomfortable insights, structure the conflict so that a community of concerned individuals can collectively interpret data, reconcile differences, and apply the insights. Make sure some of the team members are “conflict scouts” and can identify where conflict is likely to arise and then invest the face-to-face time with individuals to start examining data.
Accept the fallibility of human beings. To improve unit performance, you have to meet the unit where it is, accepting the imperfections. Certainly you need the right players on the team and you need the wrong players off the team, but once you have the team assembled, team members and management must agree that the path to improvement wanders through the garden of error. If you punish error, you encourage self-congratulatory analysis. We don’t pay people to be perfect; we pay them to become better human beings.
Make looking for disconfirming data a standard process. Individuals and teams need to continually seek out data and evidence that disconfirms what they so fondly wish to be true. Don’t mistake having a “God’s eye” view of things with continual self-doubt. The greatest performers are always scouring about looking for flaws in their performance and errors in their thinking. The difference between experts and nonexperts is that experts are usually dispassionate and incessant about this. To a casual observer the constant monitoring looks obsessive. For the expert performer, it feels like brushing your teeth — routinized daily maintenance that enhances health.
Keep the process open. Bad analysis that we read about in the papers most likely arises because the analysis agenda and approach are not open enough, providing comments and guidance from a diverse group of analysts and stakeholders. Making the process open gives people the opportunity to identify when an analysis program is jumping off the rails.
Keep focused on mission, strategy, and objectives. Above all else, and when in doubt, always remember the mission of the organization, its overall strategy, and the objective of the initiative at hand. Periodically review analysis work to ensure alignment with mission, strategy, and objectives. Prioritize analytic efforts consistent with the mission, strategy, and objectives. Any analysis not contributing to the mission and objective, regardless of how easily done, can be problematic.
Ideally, you want a diverse environment where the team players can freely come to their own independent conclusions from the data using their own logic and means of analysis, but doing so in a shared and open manner. Most of the time, when people are confronted with disconfirming data and given the opportunity to openly understand, review, and synthesize the data, they come to solutions that can integrate with their peers and the organization. Out of this process emerges a shared understanding of the nature of things that isn’t led astray by delusional, self-fulfilling analysis.
But as often occurs in life, much stands in the way between data and a good decision, most of that being human nature. If organizational leaders can develop a collective, emerging, and self-guiding process for analyzing data, anything is possible.
Photo Credit: Cleavers in Canada