We have yet another off-the-shelf new software modified for use by our company that doesn't work as advertised.

We have been told that we can modify the data if we come across any that are not accurate.

How nice of them. But what they don't understand is that once you know that the data are not reliable, every single item accessed has to be checked manually against original documents for accuracy, because you cannot know if something correlates properly until you run the numbers yourself.

This is automation?

My frustration hit maximum warp when the administrative analyst commented that only a few items were "off" so corrections wouldn't be too time-consuming. He doesn't understand (because he's a glorified secretary) that when some data (even just a handful) are corrupt, ALL data are suspect and have to be double-checked.

In my case, between 750 and 1,000 individual items.

Do you see why I'm frustrated? Not only with the program, but with the analyst who doesn't speak "Data".

I tried explaining it to him using balancing his checkbook as an example: if you've written 100 checks and discover that 2 or 3 were recorded incorrectly, you have to check every.single.one. for accuracy before you can trust the account balance.

Don't know if he gets it.