Ralph Kimball, known by millions as the father of data warehousing, is famous for an insight which underpins the reason that the business intelligence (BI) market exists: "Back in the early '70s, we had a vision that relational databases could deliver intelligence to business. Fortunately the moment passed, and it became business as usual."
What Kimball was referring to was the invention of the relational database management system, and its potential to establish relations, or causal links, between individual entities of data. However, the rows and columns nature of the relational database, or RDBMS as it came to be known, inhibited insight into the underlying data. Simply stated, the overhead on computing resources in establishing linkages was so high and cumbersome as to render unusable the very databases that had been seen as corporate saviours.
What irony. Management had been told that “data in equalled information out”, but here they were now, effectively hindered in being able to understand the relationships between data entities. The situation has got worse down the years, despite the investment of many billions in business intelligence systems.
There are two primary reasons for this:
* Companies continue to exist with their operations in silos. One leading local insurer made the decision a few years back to break its business into nine clearly delineated and separate silos. Today it is paying the price as it tries to become customer-centric. It is finding it impossible to gain a holistic view of its customers and is now considering re-integrating the silos.
* Data continues to be an ongoing challenge in almost all companies of substance, whether it's the accuracy, completeness, ownership or easy availability of data.
The fact, though, is that identification and management of relationships is of enormous value. For instance, one of the premises behind CRM is the ability to cross- and up-sell to existing customers. But unless a company can identify all of its customers correctly, and know precisely what products or services they already have, how can they do this?
(It is precisely because companies have a fragmented view of their customers that we all have the frustrating experience of being offered a product we already own, or we are targeted with inappropriate goods.)
No secrets
But CRM is only one aspect of relationship management; consider risk management. By way of example, Basel II requires that banks dig deep into companies, partnerships, trusts, foundations and other organisations and identify the principals, their partners and even family - this is to ascertain who has actual control over organisations and assets. This also applies locally, where Fica regulations stipulate as much transparency as possible.
The overhead on computing resources in establishing linkages was so high and cumbersome as to render unusable the very databases that had been seen as corporate saviours.
Mervyn Mooi is director of Knowledge Integration Dynamics.
As an example, in looking at a subsidiary, a bank would want to know who are the actual entities in charge: is it a GM or MD, or does ownership vest higher up, with the holding company or principal? And is there a clear line of sight between principal and subsidiary?
It gets even more complex when financial services companies target high net worth individuals, but find they are hidden behind a maze of trusts. How do you market your goods or services to someone when you can't even see them?
A number of tools and technologies have arisen to deal with this issue, but they tend to tinker on the front-end: data mining and data visualisation, to mention two. Such solutions will deliver variable, even indifferent results if they are not supported by a properly instituted and coherent data strategy.
Master data management
At the heart of such a strategy is the business function called master data management, or MDM. Unless MDM is addressed up-front and maintained on an ongoing basis, a company will find it difficult to move from an account- or product-centric way of running the business to a customer-centric approach.
MDM is a set of processes and tools, resourced by skilled staff, which consistently defines and manages non-transactional data entities that are commonly used (or generic) across an organisation's business units. It provides processes for collecting, aggregating, matching, consolidating, quality assuring, persisting and distributing data throughout a company, helping to ensure consistency and centralised control in the maintenance and use of this data.
Critically, MDM aims to prevent the use of multiple, usually inconsistent versions of the same master data in different divisions of the company, as mentioned above, thus preventing data duplication and overstating and enhancing a single view of customers.
* In part two of this series, I'll look at how to apply MDM in relationship management.
* Mervyn Mooi is director of Knowledge Integration Dynamics.
Share