I’m not saying that Customer Master Data Management is easy. But if we compare the capabilities within most companies with handling domestic customer records they are often stellar compared to the capabilities of handling foreign customer records.
It’s not that the knowledge, services and tools doesn’t exist. If you for example are headquartered in the USA, you will typically use best practice and services available there for domestic records. If you are headquartered in France, you will use best practice and services available there for domestic records. Using the best practices and services for foreign (seen from where you are) records is more seldom and if done, it is often done outside enterprise wide data management.
This situation can’t, and will not, continue to exist. With globalization running at full speed and more and more enterprise wide data management programs being launched, we will need best practices and services embracing worldwide customer records.
Also new regulatory compliance will add to this trend. Being effective next year the US Foreign Account Tax Compliance Act (FATCA) will urge both US Companies and Foreign Financial Institutions to better know your foreign customers and other business partners.
In doing that, you have to know about addresses, business directories and consumer/citizen hubs for an often large range of countries as described in the post The Big ABC of Reference Data.
It may seem a daunting task for each enterprise to be able to embrace big reference data for all the countries where you have customers and other business partners.
My guess, well, actually plan, is, that there will be services, based in the cloud, helping with that as indicated in the post Partnerships for the Cloud.
Liliendahl on Data Quality 
This is so true – we recently ran into most of these challenges when delivering a compliance project for an international banking group.
Some of these challenges were discussed here http://dataqualitymatters.wordpress.com/2011/10/12/bad-data-when-men-are-dressed-as-women/
Your idea of international partnerships that combine country specific rules from a local knowledge perspective is definitely useful – there is nothing available to do this currently
Thanks Gary. Your link is a good read.
My definition of reference data is as you say lists that are defined outside a single organization. A country list is a good example of such a small list with a few hundred rows.
Big reference data are, in my homemade vocabulary, such lists with many, many rows. An extreme example will be the Dun & Bradstreet WorldBase with over 200 million rows reflecting worldwide business entities. Another example is the UK Electoral Roll with many hundred millions of rows if we include several years.
Some great points, Henrik.
Master Data Management is actually easier to achieve than MDM practitioners would have the world believe. The main barriers to achieving it are the archaic methods being used in this arena. These have actually regressed, rather than advanced, and many practitioners are using practices that were outmoded twenty years ago.
These ineffective techniques include:
o Looking in existing data for master data entities
o Doing normalisation on that existing data
However, the greatest regression of all is not building a Corporate Logical Data Model (LDM) in which the master data entities are clearly defined.
Master data entities must be defined, not derived. It is nonsensical to look in existing data and try derive what constitutes a Product for the enterprise. This has to be DEFINED by senior management and built into the corporate data structures.
The LDM would also make it very simple to build a multidimensional mapping matrix to cope with the differing data structures for all of the countries with which the enterprise does business.
If MDM practitioners are to be taken seriously they cannot, on the one hand claim to be leading the way in the data revolution, while on the other they are using stone age tools and techniques.
Regards
John
Thanks John for adding in and giving attention to the Logical Data Model and thanks Gary for writing a follow up blog post on this very important subject.