Monday, April 16, 2012

Data Snitching

Things at work got surprisingly busy the middle of last week, as one of the senior government employees approached me to see if I could possibly extract and analyze some of the raw data in one of our proprietary process handlers, because the pre-programmed analysis reports weren’t quite looking at things in the way this employee wanted. And in fact I was able to do so and provide what the guy was looking for, and in further fact I was so excited to have something remotely technical and solution-oriented to do that I nearly forgot to pad out my estimates of how long the job would take. But then once I had turned the request around, the results immediately prompted further requests for similar extractions and analysis in the same vein, and requests from different sources, and it got to be a bit of legitimate effort staying on top of it all.

The problem was that the process in question was taking too long on average, as revealed by an end-of-quarter standard analysis, and thus began a round of trying to figure out how the current metrics compared with the same period a year ago, or a quarter ago, and once there was no denying that things had clearly slowed down from previous levels, it turned into a matter of trying to figure out why. And then after that, it slowly dawned on me that it was turning into a matter of trying to figure out who was to blame.

I just need a small stack of these.
The answer was never going to be “me” – I have no responsibilities at all in the process, I just handle the technical upkeep on the tool that other people use to facilitate the process. (Sorry for the overwhelming vagueness but as always I don’t want to get too bogged down in details which are both technical and also potentially military secrets.) But several of my fellow contractors (in the sense of being employed by the same company as me, as opposed to the other quasi-rival contractors who happen to have tasks in the same agency office as me) do work on the process, so at the same time the senior government employee is asking for this additional data analysis, my contracting boss’s right-hand man is asking for some of the same analysis and some further hacks at the data as well, because if the government decides the process is not being executed well, it’s our contract that could be on the line. And that of course is more than enough incentive for me to answer every question I’m asked and turn around every request for new numbers as fast as I can. But in talking to my contracting colleague I realized he already had some idea of where the bottlenecks in the process were recurring, and when I ran the numbers he asked or they essentially confirmed his suspicions. At that point I realized that I was helping to build a case for laying the blame squarely on one of my co-workers, which for all I knew might get them fired.

Which, frankly, takes all the nerdy fun out of scripting database queries. It’s not even one of those ethical workplace quandaries where you have a very collegial relationship with your manager and then one day he says, “So I’m thinking about firing Joe, does that strike you as a good idea?” and you don’t want to lie and cover for Joe but you don’t want to throw him under the bus either and on the third hand you don’t want to hazard being frozen out of your manager’s confidence by not playing along in the right way. It’s not that. “Tell me how long X part of Process Z takes on average and what the variance factors are” is pretty straightforward and there’s no way to either fudge or finesse it. And if it turns out that the data reveals someone isn’t fulfilling their job duties and that leads to negative consequences for said employee, it seems fairly cut and dry. But it’s still a drag. I kept expecting my boss’s second to ask me for an employee-by-employee breakdown of process completion times, and I even started preparing them in advance, but by the end of last week and through today that specific request hasn’t come. Maybe the storm has blown over, maybe the government side is rethinking how strict their timeline expectations need to be (once it moves away from empirical data and into theoretical policy-making, I’m long since cut out of the loop, anyway). I’m hoping so.

No comments:

Post a Comment