Analytics: how not to do it

In the last few weeks, a couple of articles came out with findings that I found interesting, but I did wonder about the approach. One of the articles did serve as a warning to me on what not to do if you want to ethically maintain privacy.

The first article (“Microsoft Finds that Remote Staff Work More”) left me with a feeling that the author was on the side of doing away with remote work because it was not good for employees’ mental health. Intuitively, the article made sense: when you work from home, you tend to work more hours, you are more isolated, and you don’t apply boundaries on your time to prevent burnout. But, to me, the major flaw is that the study of remote work took place during the pandemic and the pandemic itself is a variable: it was a stress inducing event with life and death consequences and so behaviors during this period may not be normal.

Now, we need to go through a period of studying remote work under conditions without the pandemic. (Hopefully, we are getting closer to that period although recently cases appear to be rising in many states, according to the news.)

Just like in scientific experiments, we need to consider all of the variables that may be impacting the results of the study.

The next article (“Bossware is coming for almost every worker”) offers up a very scary situation: spying on remote workers. One software appears to rank employees’ productivity through the use of emails, phones, keystrokes, databases, etc. The question is: what are the real metrics that measures productivity? Keystrokes and emails may just be a sign of busyness rather than productive work. As one creative told me: “some of the best ideas appear when one is away from one’s computer”.

I know for myself, when I am creating a solution, I tend to push aside emails and phones to reserve time for the development. Tending to emails doesn’t help the development process along. As a matter of fact, one of my creative tools helped reduce the number of emails and phones by tackling with the preeminent problem of the day. All of those emails and phone calls? They were about the issue, namely, getting a report done.

So, if I’m understanding between the lines correctly in the article, a lot of the analytics approaches in the “spying” software will not address the productivity issue and actually may worsens it. You know that saying “what gets measured gets done”? Well, if you are measuring busyness, then you will be getting a lot of busywork done but not getting the real results you are looking for.

In addition, turning on the webcam should be a no-no: that’s Peeping Tom behavior.