A client of mine was telling me about a problem they have with fraudulent credit card charges, which require a bit of work on the behalf of accounting staff to fix. As we discussed the issue, it dawned on me to ask “How often does this happen?” The response: “Once every couple of months.”
The message here is simple: Because we are pursuing success and not perfection, we’re never going to have a system that will control for every possible exception. While the chargebacks are annoying, they occur so infrequently that it’s not worth our time and effort to try to address this problem. We just have to learn to manage these exceptions.
So whenever you get into a discussion about what the database can’t do but should be able to do, remember to ask how often this situation actually occurs. You may be surprised to learn that the problem is so exceptional, it’s not worth discussing.
Wes — Good point. Sometimes the urgent things in life can take our focus off of the important. Tony
Well put, Tony.
Wes, very good advice.
During our design meetings with customers implementing netFORUM software, we train our analysts to ask this question all the time. I myself have been drawn into deep discussions about a very arcane and convoluted “requirement”. At some point, I have the sense to ask, “What kind of volume are we talking about here?” If it happens a few times a day, then maybe we continue the conversation. But if they say, “Remember that time last year?” then we put it in parking lot.
Over time, I’ve found more and more that people are not strident about coding to the exceptions. ROI is on everyone’s mind and they’re not going to want to spend $5,000 to customize and complicate a process that costs them 10 hours of labor per year to manage.
I will often get out the calculator and try to compute a rough cost of what it takes to deal with the exception (quantity of occurrence * minute to process each exception * $50/hour * 7 years of running the software). If the exceptions costs you $300,000 and a software mod will cost around $10,000 then do the mod. But if the mod will cost you $25,000 and your labor cost is $15,000 then don’t do it.
Well stated, Darryl. Calculating the ROI is a really good way to go, especially if the user doesn’t believe their exception is that exceptional. ;-)