What's up with roles in data management?

Many requests reach my employer for help in the area of Data Management, Information Management, Business Intelligence (BI), Business Information Management (BIM), Information Architecture, Data Architecture and related roles. Also, there are a good amount of jobs available in these area’s. But when you ask twelve people to explain these specialisms, unfortunately, at least thirteen different answers come back.

There is no unambiguous understanding about the parts that have to be acted in roles like Information Manager, Data Architect, Data Manager, (business) Data Steward, Architect, Data Warehouse Architect, Data Architect and even Chief Information Officer. I think it can come across as quite immature to non-IT people (as well as to some IT people) that, apparently, we are not able to build and communicate a simple function matrix.

From my own experience, here are some examples of big differences of opinion. Take for example the “Information Manager”. While one team manager has the notion that this is a role almost equivalent to Chief Information Officer (CIO), or at least the role that fulfills the demands and requirements, that have been put forward by the ‘Data Governance’ function, with regards to enterprise wide data management. For another potential employer this role is the guy that prepares data marts and, based on user requirements, loads them with data, in order for the “BI Manager” to use it for reporting.

Another example is the “Information Architect”. One opinion is this is the designer of datamodels in the data warehouse, another opinion is this is somebody who makes a blueprint, spanning the entire enterprise, of the information and data housekeeping. That means to design who can see, maintain or manage, when and what information, enrich it, or delete it. In this process constantly upholding the strategy and essential processes of the enterprise. It is not sufficient to put a prefix like “Principal”, “Senior”, “” (sic), or “Junior” to indicate the precise content of the role. It’s just as pointless as indicating a scale for this job because the two are fundamentally different roles and matching education.

I’ll mention two more examples: The “Data Warehouse Architect” and the “CIO”. To start with the latter, my understanding was that the role of Chief Information Officer meant being in charge of the Information Managers, and ultimately responsible for the availability and quality of the information the organization has to work with. In many cases, the CIO role declined to that of the IT Manager. Often this is the geezer that purchases computers, software and other IT resources. The same goes for the role “Data Warehouse Architect”. I would plea this role is responsible for the information model, the layer model, the technical foundation (infrastructure) on which information production takes place, design of historical storage of data and backup and recovery plans.
This role is sometimes confused with Data Architect, one of the roles that has the most discussion (just like the “Business Analist” is such a role). Of the Data Architect (I’ve seen the name at a governmental institution) we can imagine the same sort of responsibilities as the Data Warehouse Architect has. The difference is in the domain. The Data Architect does what the DW Architect does for the domains transaction, process support and master and reference data, preferably on an enterprise level.

Standards within Data Management, BI & IM or BIM?
Are there no standards? Well... No, not really. But lo and behold the classification of the ‘Data Management Association’ - DAMA. DAMA did research with about 60 ´data professionals´ from all over the world, into how modern Data Management is and should be organized and executed. DAMA presents a classification of functions involved in making information and managing data. This is published in the “DAMA Guide to the Data Management Body of Knowledge” (DAMA-DMBOK Guide), an elaborate work in which all aspects of Data Management have been attended to. For this book DAMA have set the first goal to be “To build a consensus” and as second (of seven goals) “To provide standard definitions”. Below is a depiction of the model of business functions as DAMA sees it.
Source: DAMA DMBOK Guide, p. 10

In the DMBOK Guide Information- and data governance is put under the heading ‘Data Governance’, because information is seen as a type of metadata rich data with a lot of information value. Activities and areas for attention that are in the pie wedges helps us imagine what roles are necessary to provide for the functions mentioned in the picture above.

This pie shows that managing and controlling of business intelligence and data warehouse is part of the business function ‘Data Management’. That doesn’t mean ‘BI is part of data management’. A BI program can, shall, must and will have data governance and -management issues dealt with. To be able to create a valuable and complete BI implementation we’ll have to take bites of almost each wedge of the cake.

In a report of the beginning of 2010, Gartner states that, if we want to be able to manage our enterprise data, we have to put 4 key data management roles in place. One role on data management and legal knowledge, one digital archive management role, one business information manager role and an enterprise information architect. Although it is a compelling statement from Gartner, only a small part of the data management gamut is covered by it. Furthermore the roles aren’t as thoroughly defined and filled in as those of DAMA.

It is interesting to see that so little of the, hard to digest, DAMA DMBOK Guide is finding its way to the world of standards and publications. Some of the ideas are adopted and are popping up in notions, blogs and even methods. If these will connect to the original ideas of DAMA remains to be seen.

I think we should indeed take a good look into the DAMA classification. It seems to be a very usable and useful concept for a beginning of order in our profession.


Populaire posts van deze blog

OTAP reloaded

Boos om "BI betaalt zich bijna nooit terug" kop in Computable