Manual Process and MDM

Written by Michael Zuckerman Friday, January 28, 2011 06:42 AM

It is interesting to review all of the market opportunity data relating to master data management. Visible almost immediately is the huge chunk of professional services required for implementation. Upon closer inspections this seems to be a very large percentage of the market. Vendors will tell you, and rightfully so, that all of this is paid for by the cost of a successful project. The return on investment is there. Of course, but this requires a successful and timely completion. However, it is still a large chunk of your cost and manual process also reads risk.

So look closer. There is a closely related but sometimes separately tracked market called (master data management) integration services. That market seems to include an even larger swath of professional services packaged with a few ETL links, extra data storage, and a data mart (or two). In total, when you add this all up, you can see that master data management is mostly about manual processes and human interaction. The software expense feels like, on a very good day, something approximately over 70% of the total expense for initial implementation and likely approaching 80%. The software in that bucket is a small slice of the overall expense. It is all about manual process and people.

If you don't see this expense you are not looking closely enough, IMHO.

Perhaps the costs are, in fact, even higher. My data about expense is all anecdotal. I ask the question and get answers. But most people are not completely sure. The vendor information is somewhat optimistic about implementation time. The industry analysts are a bit more astute, but they tend to leave the integration expense and consulting in another bucket. It is very hard to tell. When all is said and done, no one has likely been able to capture all of the email, conference calls, political wrangling and hallway meetings required to align your data models on a global basis. It is likely far more expensive than you know.

Manual process and human interaction does not scale well. It is not more efficient. It gets less efficient as the number of intermediaries to measure progress and “manage” grow. These resources cost more, raise the overhead and generally slow things down. You have alot of politics to deal with. Politics is the art of getting people to do what you want them to do as a function of influence, not authority. The results are not easily predicted, cooperation is not assured and the goals of the different parties are not in alignment. It should be very clear, especially in the wake of the economic debacle of Q4 2008, that the needs and interests of these groups are focused first on their top goals and objectives. Master data management may not be one of them.

Of course, to make this more difficult, all of this manual process and human interaction involved is distributed far and wide. You are necessarily dealing with people from diverse working groups and these working groups are attached to every line of business, IT team, geographical entity and P&L in your company. The larger you are, the better the return on investment should work for you, the more diverse the interaction you must sponsor. Cooperation is always a function of relationship, and you cannot always build close working relationships over a videoconference line or a conference call. Cooperation does not work well by command, except perhaps in the military. So you need to get on planes and visit your various work groups, at least once or twice. More process and time tied to hours of your time.

Lack of automation is a big problem. Right now everything is wrestled to the ground by committee. Emails and policy statements then flow out regarding data governance decisions. It is all manual process – not much is automated. There is no computer software enforcing and implementing the policies. You may need to scrub the duplicates out and if you do not… nothing happens.

Data governance teams need the same sorts of tools the information security teams have for implementation and enforcement. For example, your password expires and you must change it or you lose access to the system; you cannot install a 3rd party piece of software that is not approved; your email storage is limited in size and in some financial institutions must be auto archived and/or expired depending on the date and regulatory oversight; and so on. Automation needs to support authority but there are no tools, dashboards or automated workflow to tie this to legacy MDM architectures in a meaningful way.