Analytics & the Individual

I know lots of people have seen Stephen Wolfram’s “personal analytics” – look here if you haven’t read about it.  Essentially he’s been tracking how many emails he sends, when he’s on the phone, how many keypresses he makes, and tons of other metrics.  He then visualizes the data in ways that bring out patterns. Looking at his plots of email send-times, or file modification times you can see both day to day consistency, and also big events in his life, like working on a new book or traveling to different timezones.

I think most people would find this instantly useful – geeky, yes, but useful.  It’s like if you want to lose weight, weighing yourself every day helps you stay on track … without data we drift from our plan and our goals.  I agree with Wolfram when he says:

As personal analytics develops, it’s going to give us a whole new dimension to experiencing our lives. At first it all may seem quite nerdy (and certainly as I glance back at this blog post there’s a risk of that). But it won’t be long before it’s clear how incredibly useful it all is—and everyone will be doing it, and wondering how they could have ever gotten by before. And wishing they had started sooner, and hadn’t “lost” their earlier years.

All well and good, and I can’t wait to get my golf game, my work productivity and, yes, my weight, into a personal analytics system.

In my work we use analytics for lots of things, and we are also trying to think up new ways where analytics adds value for business.  Recently a number of individually focused ideas have come up in conversation.  For example, companies track lots of data about each employee: education, years of service, performance ratings history, IP contributions (eg, patents) history, documents posted, defects fixed, age, family status and more.  It is conceivable to me that with this mass of data analytics could be devised to determine who, for example, might be a retention risk.

I don’t know what to think about that.  Certainly these assessments are already being made today; they are just made by people.  The data belongs to the business, it certainly can use it in an automated way if it wants to.  And finally, if the data was used in the aggregate – meaning to assess in a population what proportion of people are retention risks, vs. assessing what individuals are – then I wouldn’t mind that.  But using the data to highlight individuals automatically invokes my Minority Report reflex.

Why is it ok for me to do analytics on myself, but not to have someone else do it on me?  The thing I come up with is that I “know me” and the system doesn’t.  For example, maybe someone really likes their employer, or really likes their co-workers.  How are we going to capture that in the retention-risk analytics?  Don’t see how we can.  Bear in mind that the retention-risk example is mostly benign – who knows, if Deep Thought decides you might quit, maybe you’ll magically get extra perks at work (not likely, but I’m trying to stay positive here).

Hopefully as these kinds of analytics become real we’ll remember to keep “wetware” as part of the system.