It is worth noting that last month school closures made the front page of The Canberra Times. One of the arguments raised was that the data used to make the decision to close some schools was incorrect. It is not clear if the data was of sufficient quality to support this decision but it does appear that opponents have focused on this aspect of the decision.
Once this documentation starts to be produced, questions will arise about how people currently use this information in the organisation. This should lead to an improved understanding of which information is important and hence needs attention. This also applies to unstructured information and how people in the organisation find and use corporate knowledge. My guess is there are number of key people who have this information in their heads and become the unofficial gatekeepers.
Further issues have to be considered. The currency of this information - that is, how up to date it is - is important. Parents levelled this criticism at the data about schools. Then there is the question of who controls both the meaning and the data item itself. Traditionally this control role was taken by the technologists and database administrators. In reality it should be controlled by whoever is maintaining the metadata and the meaning of data items.
This level of control should prevent duplication, for example, multiple addresses being stored across multiple systems each with a slightly different meaning, depending on who collected them and for what reason. Some parts of an organisation will collect an address as a means of verifying a person's identity, while others will use it for mailing.
If the people who use the database are not aware of these subtle differences, the wrong address could be used for the wrong reason. Many people have multiple addresses. When the corporate data collection is searched which addresses are found? A more likely question is which legacy systems will allow more than one address to be stored?
There are other corporate information resources which must be managed. Culling and archiving are important - although disk space is cheap, keeping multiple copies of documents does not make sense.
The management and maintenance of corporate information such as phone lists, mailing lists, postcode look-ups and other common data sources is part of the information architecture. Setting rules for access to databases and controlling updates to corporate information is vital if the quality of the data is to be maintained.
Similarly, there must be rules and procedures for using data items that are liable to be duplicated across systems, for example, name and address information for customers. The chief information officer usually controls these rules because no other manager in the organisation is empowered to make and maintain the access mechanisms. This is clearly not the role of the CIO - although I expect some will argue with me.
In practice not all the above requirements for IA will be met immediately, but the organisation should be moving to an environment where information is managed. One suggestion is that only a few key data items are closely managed and monitored. These data items will be standardised and documented and controlled. Once the benefits become obvious other items will be added.
Organisations will gradually become more aware of the need to manage the information environment and the requirement to share data with another organisation is often a turning point. In an e-government environment or whole-of-government environment, duplicated data-collection processes will not be tolerated and the move to an IA will be mandated.
Interoperability will force system owners to document the data and metadata for items which are used by, or transferred to, other systems. After some interoperability failures and problems, policy and procedures will be developed and the organisation will begin to track information use and monitor the processes by which information is changed. This will flow to the relevant business practices, which in turn will be updated and maintained as part of the IA.
So what would the information environment begin to look like? Accountability for provision of data to other parts of the organisation or to another agency will be increased because the recipients will demand that the data used for decisions is "fit for purpose".
A global high-level enterprise model and supporting metadata will be established to support this new IA environment. Business functions will be included in the enterprise model to ensure that there is corporate alignment and data interoperability.
Publication of data-quality metrics will become common within organisations. This will happen when errors are discovered in the main business metrics.
Key data items will be documented as part of an information chain and all supporting processes will also be documented. Some time after this, managers will become aware of the value of a documented information chain. After the preparation of information chains, they will begin to exploit this new information.
This recognition will foster a climate in which navigation of the information chains and supporting metadata will be routine and a set of enterprise standards will be generated to ensure that future work in this area generates commensurate benefits.
At some stage the accountants will recognise the value of information and assign a dollar value to it and it will become part of the balance sheet in the same way that organisations account for goodwill.
Finally, the term information manager will disappear because all managers will be accountable for their information assets.
Autor(en)/Author(s): Steve Neilsen
Quelle/Source: The Canberra Times, 04.07.2006
