Human behavior studies show that if you are offered a pizza menu with a number of combination offerings as well as “build it yourself” options, you will order more toppings than if presented with a menu that only presents build it yourself. The same thing for medical tests – if a doctor is presented with a pre-filled in menu of recommended clinical tests for specific diseases (and the option to strike out tests that are not needed) they order more tests than if presented with the same order form but with nothing checked in advance. So what is the relevance for Master Data Management (MDM)?
The lesson for MDM is that if you start with a large complex model, like a financial services data model or a supply chain model (which any number of vendors would be happy to sell to you), you will end up with a more heavyweight model than if you start with a simple skeleton model and add entities and elements as you need them. Your data is unique; even though you share common practices with your industry peers, your data and how you organize and act on it is quite often a differentiator. Investing time to appropriately structure, plan and scale your underlying data model will pay big dividends along your MDM roadmap and provide more flexibility to maximize return on investment.
The enterprise data models that you can purchase look very robust and well thought out – and they are. If you decide to acquire one you may have the intention of making it lean by trimming out the unneeded elements, but somehow it seems hard to remove anything. Maybe it’s due to the same hoarding mentality that results in your closet being stuffed with clothes that you haven’t worn for years, but you keep them anyway because they are still good.
In any event, if you buy one you will have an MDM data model of which you are using only a small fraction. The size and complexity of the model makes maintenance difficult, upgrades harder, has a long learning curve for new staff, and generally makes changes and project work sluggish.
Lean manufacturing introduced the concept of “pull” and Just-in-time delivery; the basic idea is to not build anything until there is a demand, and then build it fast. Sometimes it makes sense to anticipate needs and build things in advance so that we can deliver faster, but building solutions in advance ties up resources and our track record in IT for guessing the requirements just right is not something to be proud of. The old adage of ‘build it and they will come’ has proven to be generally wrong, and very costly, in the MDM arena. So too is the reality of the underlying data model. The strategy of starting simple with MDM and addressing an acute immediate business need naturally drives a focused, appropriate data model that leverages best practices while avoiding the ‘more is better’ trap of a one size fits all industry approach. MDM is a journey, so don’t start with a data model that assumes a final destination. Start simple and invest in a supporting technology solution that provides the flexibility to right size the data model along the way as your organization matures along the MDM continuum.
In summary, my advice is, have the enterprise master data big picture in mind and spend some time laying out the master plan, but then build the master data model incrementally. The net result you can expect is faster delivery of tactical solutions that demonstrate business value and build momentum towards the vision.
To learn more about Lean MDM visit www.integrationfactory.com.