One of my favorite memes of the modern technology era are those funny texting autocorrect disasters, where people mean to say one thing, but the ever-helpful device replaces a few choice words to make the actual reply, once you've committed and hit 'Send,' something comical. It's entertaining stuff. I often wonder if Apple, Google, and Microsoft intentionally tweak their software to cause some of these errors for the purpose of generating consumer buzz -- but that would be a very cynical view of their intentions.
Spell check has made us lazy. The fact that so many of these errors get out -- enough of them for this epidemic to be called a meme (definition: "an idea, behavior, or style that spreads from person to person within a culture") -- proves my point. We're not even paying attention any more, relying on automation to clean up after ourselves. In the April 2013 edition of Wired, one of my favorite columnists, Clive Thompson, talks about over-reliance on machines (Leave the Driving to Us), and how "tools that make hard things easy can make us less likely to tolerate things that are hard."
Of course, Thompson goes into the dark side of losing our grip on technology: "Things get even dicier when society at large outsources its biggest moral decisions to technology." His article shares ideas from Evgeny Morozov's book 'To Save Everything, Click Here' in which Morozov points out that over-reliance of data automation in law enforcement, for example, has led to problems. "The algorithm would wind up amplifying flaws in existing law enforcement. Remove the deliberation of what police focus on and you wind up deforming policy."
From time to time I have written about the skewed world-view of managers who become over-reliant on specific analytics, losing perspective on what is really happening on the ground -- managers who spend so much time looking at spreadsheets and performance indicators that they lose sight on the human aspect of their jobs. It is, in my opinion, much the same as what Morozov and Thompson describe here. As Thompson further points out, "Efficiency isn't always a good thing. Technology lets us do things more easily. But this can mean doing things less effectively, too."
Case in point: a previous team that spent an inordinate amount of time watching a handful of core metrics, constantly looking for ways to further automate those data points, creating rules, policies, and even a performance reward system almost solely around those metrics -- but they had a difficult time recognizing any flaws of their system when it came to human examples. Employees who were high-performers but who tackled problems and customer issues that were not sufficiently weighted and measured within these automated statistics saw their performance numbers decline -- even though they had some of the highest customer satisfaction scores, and were viewed by customers, their peers, and partner organizations as some of the best employees in the organization. Management became myopic with regard to their metrics, and the performance records of good employees were mischaracterized. The management team had allowed automation to blind them to what was really happening.
Not all stats are bad -- but that's not my point. All stats (like good electronic music) require some degree of human interaction, providing adjustment, optimization, and sometimes a little common sense. A single KPI provides perspective -- a slice of data to help you understand the big picture. Morozov suggests that automation and machine-generated statistics are sometimes used best to insert different, or contrary data to help you question the status quo. I like this idea of stats as a disrupting factor in your thinking process. Thompson says "We're not going to throw out all technology, no should we. Efficiency isn't always bad. But Morozov suggests that sometimes tools should do the opposite - they should introduce friction."
I like this idea of friction in our analytical pursuits. Don't rely on automation to replace your management duties -- i.e. your need to make decisions -- but instead use it to highlight inconsistencies or exceptions to your model, allowing you to make more complete, tempered decisions. Especially important when managing people.
#reporting #Management #analytics #statistics #BusinessIntelligence