3

I'm a software developer who (because of historical reasons that probably aren't relevant) is forced spend at least 30 minutes every day reviewing files with financial data in them. I'm not an actuary so I'm not qualified to review these files. This work is extremely tedious and error prone and I hate doing it. I've told my manager and my manager's manager as much and they said it's not a priority right now to automate it. This is frustrating because I've already written a program to do the review automatically but I can't use the code because it hasn't been reviewed yet. My coworkers and I tracked our time spent doing this work for a month and showed it to our manager's manager and he says he wants the data analyzed more thoroughly (broken out by task, etc). Any advice on how to convince management to either let me use the code I wrote or prioritize this would be appreciated!

EDIT (responding to comments/questions)

There has been a mandate from an executive at my company that too many small mistakes have been made lately so they're cracking down on manual processes, hence the review. After reviewing these files I have to send out emails to the consumers saying that the data is good. Since I'm not an actuary, I just say the data is good if it follows the business rules the actuaries have given us. Management does not want any analysis done on the data unless it fails one/more of the business rules. My main question is how to convince management to prioritize having someone else do this review or prioritize saying the code I wrote is good to go so I don't have to do it manually.

reggaeguitar
  • 149
  • 1
  • 6
  • 9
    Can you explain how you are "not qualified" to do this review, given that you seem to be able to do it (and even wrote code that does it) ? – Erik Mar 12 '18 at 21:03
  • The actuaries give us rules like column Foo must be greater than 0 that I turned into code. I don't understand the data I'm actually looking at though so reviewing it manually is a complete waste of time in my opinion. – reggaeguitar Mar 12 '18 at 21:08
  • Reminds me of our badge of shame - Is it unethical for me to not tell my employer I’ve automated my job? – Bernhard Barker Mar 12 '18 at 21:32
  • Can you identify the structure and the critical areas of the file? – Isaiah3015 Mar 12 '18 at 22:39
  • "The actuaries give us rules like column Foo must be greater than 0." That sounds like it can be done in one cell in an Excel spreadsheet. – Simon B Mar 12 '18 at 23:10
  • 3
    Why not use the code without telling them ("better to ask forgiveness than permission")? Or is there a good reason why code has to be "reviewed"? – cag51 Mar 13 '18 at 01:29
  • 2
    What does "reviewed" even mean here? Are you supposed to output something? Stamp it as reviewed? File it to a folder on a network share? – Jack Mar 13 '18 at 07:55
  • "The actuaries give us rules like column Foo must be greater than 0." What reason do you have to believe that management wants anything more than this done in manual review? You say you don't understand the data beyond these simple rules, but do they actually expect analysis beyond these simple rules? – Steve-O Mar 13 '18 at 13:53
  • How can I make this "on-topic"? I read the meta post and it seems like my question has a goal to address: how to get management to prioritize getting this off of our plate or signing off on the code I've written. – reggaeguitar Mar 13 '18 at 15:30
  • I voted to reopen as I think it's sort of on-topic. There are two benefits to your automation - 1: Vastly reduced time to complete the reviews. 2: Elimination of human error (you could miss something looking over it with your eyes, a computer program won't unless there's a bug). The rest is up to your powers of persuasion (which may be why it was closed since we can't help you with that!) – komodosp Mar 13 '18 at 15:36

1 Answers1

6

To expand on cag51's comment

Use the automation code to greatly improve the efficiency of your review! The program will be a lot quicker and a lot less prone to human error (e.g. overlooking some rows, etc.), though admittedly could contain bugs.

You could start doing the human/eyeball review and the programmatic one side by side for the first few days until you can tell for sure that at least some of the rules are working perfectly (e.g. the simpler ones) while at the same time using it as a review of your program to debug the more complex ones.

This way, after a few goes at it, you can at least reduce the time it takes to do the review, hopefully significantly.

komodosp
  • 5,720
  • 17
  • 26
  • I'd change first few days to first few weeks but otherwise +1 – rath Mar 13 '18 at 10:39
  • Agree with this approach. You may not be able to merge the automation code into the general codebase without approval, but surely you can keep it on your local machine and run it yourself? After a suitable "testing period" of doing it manually as well to check for bugs, this could be a real time-saver for you. Furthermore, if you document the usage and results of the program somewhere, that could be used as further leverage to convince management of the benefits of automation. – Steve-O Mar 13 '18 at 13:50
  • I've been running the code and it works great, it's not that complicated of code. I still have to do the manual review though because of mandates from an executive level and there's no priority to get my code reviewed which is my main problem. – reggaeguitar Mar 13 '18 at 15:21
  • @reggaeguitar - But see the question/comment by Jack in response to your original question. What is the output of your review? – komodosp Mar 13 '18 at 15:29
  • @colmde read the Edit of my question above – reggaeguitar Mar 13 '18 at 15:31
  • Will anyone notice if you don't do the manual review? – ThiefMaster Mar 13 '18 at 16:16