There's Nothing Funny about "Funny" Data

Data Integration Advisor: There’s Nothing Funny about “Funny” Data

Many business intelligence projects fail simply because the project team was surprised by the data – specifically, they were surprised by data anomalies lurking in the source systems. Sometimes the problems can be blamed on data quality, but often there is no problem with the correctness of data. Rather, the problems lie with the data’s completeness or the way specific fields are being used. A field might be used for multiple purposes depending on the business process or division that updated it, but those multiple uses are not documented, so it ends up looking wrong.