Bill Inmon, often referred to as the father of data warehousing, coined the phase in 1990 stating “A warehouse is a subject-oriented, integrated, time-variant and non-volatile collection of data in support of management’s decision making process”. By adding Ralph Kimball’s concept of Dimensional Modelling, specifically ‘Conformed Dimensions’ you come up with the three most important capabilities of a Data Warehouse, Subject-Oriented, Integrated, Conformed data entities that bring together all key data assets at an enterprise level, into a Single Version of the Truth.
It is true that in business today you will have many different definitions of the ‘same’ data. But that is the joy of data warehousing, during the data discovery stage you will work with business representatives to define each and every entity in the warehouse, gaining clear agreement on the definitions. The key action here is that you get down to the lowest level of granularity with your data, a level that is relevant to the context of the subject-area you are discussing. If you can still ask the question “Is there any ambiguity in this definition” and get an answer, you need to dig deeper, you need to discover those single truths to ensure that they are understood.
Having agreed, preferably signed off, Data Definitions, readily available in a Data Dictionary is a powerful tool, it is a foundation of fact that can not be disputed once agreed. Terms like ‘Customer’, ‘Prospect’, ‘Revenue’, ‘Net Revenue’ etc will all be defined at an anatomic level, it is the definition that drives the understanding. Clearly defined data entities, relevant to the context within which they are used, form the Single Version of the Truth.
There is a major player in the Single Version of the Truth game, and it is a clear heavyweight contender. It goes by the name of Master Data Management or MDM.
The requirement for a single, accurate, and timely source of key data entities has evolved. No longer is it just the domain of decision support that requires this information, the business is craving it at all levels.
We know that data warehousing has been providing Single Versions of the Truth to the decision support and analytic environments for nearly two decades, however, by its very nature, it is not designed to act as a source of data for transactional and operational applications. That’s not its game. Enter MDM, the ultimate provider of the Single Version of the Truth.
By definition, MDM is a comprehensive approach to linking all an enterprise’s data elements. It provides a single, consistent view of critical data by using both technology and data governance techniques. MDM as part of your business and technical DNA is critical to business success today.
Data Governance is a key aspect in MDM. With the correct processes and technology MDM will provide the ‘golden record’ that the business strives for across multiple subject areas, driving CRM initiatives, improving product management, and driving down cost in manufacture.
Having a Single Version of the Truth woven into your companies application architecture and business processes is essential to success, it drives Data Governance to the front line, improves application efficiency and enhances your decision support environment.
MDM is the Holy Grail when it comes to the Single Version of the Truth!
Geeesh, did you actually say anything there? What a load of hypothetical hogwash! How can MDM, and more specifically ‘the golden record’ be the Holy Grail? I agree that having common shared business entities is beneficial, but they can only be common within relevant business processes. Once again you can not get away from the fact that business divisions are disparate, they will have AND need their own definitions of data that at an enterprise level are essentially the same. Shared is the way to go, Single Versions of the Truth are a pipe dream, it’s not possible!
Those For – In Response
You need to get away from viewing this at such a high level. As with the context specific conformed data in a data warehouse, a golden record is contextually relevant to its intended use. Therefore the definition of that record needs to be as granular as required. It is the same argument as in data warehousing, taken to a different level, you are no longer just talking about data for slicing and dicing, you are talking about data that is at the front line of your business, in CRM and ERP applications, used by Sales, Service, Finance and Manufacturing at an operational level.
For example, the term Customer can be broken down into:
- Web Customer – someone who only interacts over the web
- Retail Customer – someone who only interacts in person in the store
- Business Partner Customer – a prospect to an agency that is a customer of its business partner
- Customer – the derived total of all lower level customer entities
As you can see it is possible to have single definitions that are contextually correct. That is where Data Governance comes in, ensuring that these definitions are correct, up to date and readily available in the Data Dictionary.
Data Governance is the key to this, through discovery, agreement and communication these single truths will be understood. Having the business understand these truths is the key to the Holy Grail.
Please take the time to read all three posts and then vote for who you think has won the debate. A link to the same poll is provided on all three blogs. Therefore, wherever you choose to cast your vote, you will be able to view an accurate tally of the current totals.
The poll will remain open for one week, closing at midnight on November 19 so that the “medal ceremony” can be conducted via Twitter on Friday, November 20. Additionally, please share your thoughts and perspectives on this debate by posting a comment below. Your comment may be copied (with full attribution) into the comments section of all of the blogs involved in this debate.