The final installment on the ingredients for successful BI is about data quality. For some strange reason, if users do not trust the quality of data in a BI system the system will fall into disrepute and be seen as a failure. Now I said "strange reason" and in truth there is a degree of strangeness about this; the data in the BI system can be as good as the source (ERP) system but still be too flawed for a successful BI system; quirks that are tolerated on the transactional side of the business are abhorred on the reporting side.
You'd think that a BI system that had a single data source would be relatively immune from data quality problems. But data is entered by humans, business processes are defined by humans, and humans have the power to bend rules for short-term expediencies; common failings are incomplete reference data - people are just too busy to add all of the attributes to products - and obsolete data not being deleted, after all it's a struggle to keep up with new stuff so why make work dealing with the old.
But when you take information from a multitude of incompatible sources and try to impose a 'common view of the world' on it the big problems arise; is that customer on CRM system the same as the one in the sales ledger? Is this address the same as that one? Then add in all of those human factors around data entry and you will get an idea the nature of the problem