System architects like to create beautiful models of operations. The models are based on information that moves between components. The model runs like a clockwork, but the problem is that the data entry is manual. It is quite easy to make mistakes while entering the data and there is no mechanism that corrects the mistakes. Soon the system becomes tainted. As Mark put it, it is like mixing wine and dishwater. Adding a little wine to dishwater doesn’t change the nature of dishwater, but adding a little dishwater to wine certainly does.
There are two major activities that include a lot of manual data entry in ITSM: incident and configuration management. Both suffer from this data quality problem.
In incident management the staff typically add service and configuration information to the ticket. The problem is that in many cases they do not have the required information and therefore have to guess. The result is like dishwater in wine. Nobody trusts the incident data and the reports based on it are therefore generally worthless.
In configuration management all changes must be recorded in the CMDB. It takes a lot of effort to build and maintain the CMDB. Unfortunately, it takes very little effort to ruin the system. Imagine a person making an emergency change at 5 AM to solve a major system outage. After a successful operation, he goes home to sleep. The next day he updates the CMDB but makes a mistake or forgets something. Then people stop trusting the CMDB data, they realize that they need to check the actual situation to be sure. After that it becomes less important to record the changes.