The ultimate PEBKAC?
I worked for some years at a company which ran their business using a bespoke piece of software created by their in-house programming team. Some years later, when the company decided to replace it with a bought-in system, the vendors of the new system were astounded at what it could do, as the software they were selling could not handle some of it at all (my ex-employers decided to change the way they worked to suit the software they bought...). Anyway, I digress...
The PEBKAC of which I speak came to light about halfway through my employment there. The order processing team of one section of the company (the largest one, as it happened) kept having problems, and we were mystified by why, as nothing could be found amiss with the system or the data therein. Usually not long after the issue had been looked at by us, and nothing found to be wrong, the response from order processing would be along the lines of "It's OK, we've found and sorted the problem" - but we couldn't get any coherent idea from them of what exactly they'd found or how they'd solved it.
These constant reports of problems with the software which the programmers hadn't been able to fix, but which had been "solved" by much work by the users themselves, naturally caused criticism of the system from the management. Which exasperated my boss (manager of the software development team) so eventually she spoke to the manager of the affected order processing team and asked to be shown exactly how they were using the system.
It transpired that at some point (I have no idea when) it had become entrenched in that departments collective mind that because "the system was so unreliable" (translation: they felt that they understood spreadsheets, and so decided to use them to check that the bespoke system was giving correct figures) , that they would enter all data twice - once into the bespoke system, and once into spreadsheets, so they could tally the value of the orders processed each day themselves in their spreadsheets, and later compare that to the figures that the bespoke system gave. This, of course, meant there was at least twice as much work to be done as there should have been at the best of times, when no problems arose. But when they found discrepancies, that, of course, took up more of their time and ours trying to find where this mysterious (to us) problem was.
Now I don't know exactly what my boss said to their boss, but I suspect it was along the lines of pointing out that entering data twice into two systems doubles the chances of errors in data input, thus creating false claims of problems with the bespoke system which were actually data input errors. Data input errors which we couldn't find, because, of course, we had been kept in the dark about their use of spreadsheets to check the results coming from the bespoke system.
Suffice to say, not long after that, the numbers employed in that order processing team rapidly dwindled to about half of what it had been before, the same volume of orders still got processed, and there were far fewer problems reported by that team!