Killing Productivity by Measurement
blogaborty.blogspot.in
blogaborty.blogspot.in
If there's a community where this can produce nice results, it's HN.
In work, if I am given what is considered the next most important thing, what will those measurement tools get us? Maybe, at best, an idea if we want more people on the project but those projections can be rudimentary at best. Couldn't we just eyeball it and come to similar conclusions?
Productivity measurement reminds me of stock market analysis. Previous performance isn't indicative of future success. Projects are generally different enough that each one will have new challenges to figure out and learn.
This has been the holy grail for a while - an app that runs silently and automatically records everything you do while working at your computer, categorizies it if it knows how, then lets you run reports on it later to see how you spend your time.
No manual data entry, no clicking an app button every time you change tasks, etc. Just everything automatic in the background until you want to see the results.
Afaik there's only one that's achieved anything like this, a Mac-only app built by an Oxford Philosophy grad student iirc, which took him several years of hacking OS X to figure out. Don't recall the app's name at the moment though.
I looked into doing this on Linux a while back, and figured an MVP version would be one in which you recorded each window that gains focus, its title, and the focus date/time. I would know just from the window title whether it's work or play, and could categorize/tag each as such, or implement some Bayesian auto-tagging method.
But the problem, on Gnome and KDE at least, was that there's no way for a third-party app to know which windows gain focus and what their title is. They're completely sandboxed from each other and blind to that, and no window manager API they can listen in on either.
The Mac guy had the same problem and figured out some intricate hacks to overcome it, to the extent Wired or someone did an article about (a couple years ago, details hazy). I didn't get that far, but haven't been paying attention recently, maybe others have?
You can't get feedback on what you don't measure. It doesn't mean you can't manage however. And the cost of getting that feedback can be lower employee retention, lower productivity (which is almost completely hidden from you near term), and false early returns (people tend to preform better briefly when you start to measure them, then go back to normal).
If the staff doesn't have the knowledge/tools required and/or the work morale that is needed to be productive and get the job done I really don't think that micromanagement is the proper solution.
It's almost as if it's subject to the Heisenberg Uncertainty Principle. In this case with the tools you're using you're not getting an accurate reading of productivity because typically the burden is on the subject to make the readings and they will get annoyed at this intrusion into their day job.
So how do you measure tasks without being intrusive? Is it really possible?
At https://starthq.com, we are considering introducing a feature that would let people see how much each web app is used within their team in an attempt to increase their productivity, but much like the comic strip I wonder whether that will do more harm than good.