On Tues 9th April 2013, the BCS-CMSG configuration management specialist group held a very useful presentation on maintaining CMDB (Configuration Management Database) accuracy, from Michel Delran of Unisys. This prompted several thoughts, including the thought that a high-quality CMDB needn't be one physical store (it could be a logical CMDB federated out of a lot of different configuration information stores; possibly, in part, even including a CMDB product). Another thought is that if moving from a state of poor CMDB accuracy, with considerable dissatisfaction from CMDB users, to a state of high accuracy, and a popular CMDB, simply results from understanding the CI (configuration item) life-cycle and exploiting automation and tools sensibly, why isn't everyone doing it? Especially as Delran has case studies showing that this can be done while more than halving the staff resources needed and increasing user satisfaction at the same time.
Well, so you can manage CMDB accuracy - but remember that this has to be justified in terms of business outcomes, not the number of CIs managed successfully. On the other hand, you don't achieve high user satisfaction levels with a large CMDB if no-one finds it useful or helpful - even if you do reduce cost by halving the number of people needed to maintain it. So I'm pretty confident that Delran does achieve top management buy-in for his CMDBs from delivering business outcomes; and also produces something that workers at the coal-face find useful. You can see his presentation here - but it's in 6 parts and it may take a little time for it all to be processed.
His presentation also helps to promote understanding of why some CMDB projects fail. But what about the majority of organisations which do basic 'change management' instead of 'configuration management' (you can find definitions of both in the glossaries here) and have no intention of building a CMDB? One common characteristic, I suspect, is that their top managers have no knowledge of Configuration Management and/or no interest in it.
Nevertheless, that really would be an interesting BCS-CMSG workshop - a practical guide on how to deliver resilient Business Outcomes in an evolving business without using Configuration Management and, certainly, without implementing a CMDB (federated or otherwise). It's apparently a popular approach, so there should be a lot of evangelists for it and if I get any offers, I'll pass them along to the CMSG committee for consideration. Somehow, however, I don't expect to get many takers.