Thursday bi-weekly posts are to paid subscribers, and are shorter, off-the-cuff bits about what’s on Randy’s mind. This week has been crazy busy with Other Stuff.
Know how you look at the fields of a database table or log entry, plus a few rows of example data, and you can often make an educated guess at what generated that data? Normally I call that one of the super skills we pick up on the job while learning to work with data all the time. Well, I’ve recently found a surprising downside (sorta) to the skill.
The past two months I’ve been staring at a handful of rows of data within the NYC Dept of Buildings while trying to get some work done at the home inspected. We’ve had a bunch of issues so we’re up to inspection #5 soon. What’s interesting about the system is that the DoB has a fancy Tableau powered dashboard that shows the typical service levels for various things, including how long it takes for DoB to do an inspection when a request has been made by a contractor.
The whole inspection request system is a web front end where a contractor puts in a request, filling out the forms specifying what is to be inspected. This goes in and eventually a confirmation is sent, an inspection is scheduled, and the inspector arrives at a given time. In theory, at least looking at the metrics on the dashboard, we see that there’s supposed to be a handful of days between a request and inspection — great turnaround!
Reality is much different. I’ve had to go through as many as multiple weeks to get a re-inspection on the calendar, a far cry from the handful of days in Tableau. The must frustrating part is that since we’ve done this dance so many times now, I can see the history of timestamps when requests “materialize” into the system… They happen at weird times, all within a typical workday between 8:30AM and 5PM. Always at very weird and uneven minutes. The only conclusion I can draw is that there’s a human entering the requests from a queue into the system when it suits them. No automated script would be so inconsistent in timing.
But I don’t know who that human is. Maybe my contractor is playing dumb and messing with me about sending requests when they’re stalling. Or maybe DoB has a human that is manually moving requests from one queue into the calendar. My guess is the latter, but I’ve no proof other than watching the wisps of data. But if it’s some random clerk at DoB manually moving entries around, I’d be willing to bet that this manual process helps juice the metrics for response times. If an inspection request doesn’t “count” until it’s moved off the incoming queue into the main work queue, then it’s trival to hold things in the queue until you can handle them.
Sadly, since I don’t seem to know anyone who works with the actual system in question, all I have is speculation. But this, and a couple of other wonky details about making requests in this system has driven me nuts. I’ve called in trying to get things moving or request clarification and the service line people have no helpful information they can give me on how to make things move faster. So I wind up sitting at my desk, utterly frustrated that there’s probably a human out there that needs to be reminded to move my requests into the correct place but hasn’t gotten around to me yet.
This probably would’ve been “easier” if I didn’t make weird guesses about their process. Then I wouldn’t have to make phone calls trying to see if I could reach a person who theoretically moves things along — all of which failed.
I just really hope we pass the next one…