We Like Bad Data
Ian Bailey
It’s not the typo problem that is the main concern for this particular blog-rant, though. There’s a category of data quality problem that seems to be targeted above all others: putting the wrong type of information into a text field. This is considered particularly evil, as the user is knowingly entering the wrong information into the database – usually as text. A typo’s a typo, but surely this is sabotage ?
To understand this problem properly, we have to look at how the text field that’s being so callously violated came into being in the first place. Initially, a business analyst (BA) will have interviewed the users about what they do. They will almost all have lied. Either from fear that the big bad consultant has come to take away their job, or because they have an inflated sense of their own productivity. To avoid this problem, some BAs have adopted the disturbing approach of simply “observing the users”. This has the effect of altering their behaviour (a bit like quantum uncertainty, you have no idea what they’re up to unless you observe them, and when you do, it modifies their behaviour). Whichever approach is used though, the BA will develop a process model. Large organisations love process models. An unkind observer might suggest that these models allow managers to see what their business is doing without having to actually spend any time interacting with the staff. The usual claim though is that they aid decision making process in change programmes. Either way, they mean more time for golf. Lovely.
del.icio.us