Data plays different roles in organizations

25-09-2015 - Wouter van Aerle

As data practitioners and information professionals our day job is to manage and exploit data for the benefit of our stakeholders. Whether it has to do with improving data quality, integrating data to support analytics, ensuring data is protected or assigning responsibilities for data governance, you name it. Fortunately, there’s a truckload of best-practices, handbooks, frameworks and How To’s that provide us with valuable knowledge and experiences.

However, putting that knowledge into practice is not always straightforward and sometimes even a major challenge. If we jump over the most common pitfall of “Gimme the template!”, the more structural obstacles come into sight like management buy-in, business driven approach, managing the organizational change (about which Dylan Jones wrote an excellent post by the way) to name a new. These are all important and relevant conditions that are also often overlooked or managed poorly. Still, even if all the conditions are in place, applying data management practices can still be cumbersome. It’s like the practice or framework simply doesn’t seem to fit the situation at hand. Ever had that feeling? It’s a rather strange experience because there’s no doubt there’s a data problem at hand. And yet solving this problem with all the relevant knowledge around isn’t straightforward.  

After analysing several of these cases, my conclusion is that there are fundamental differences in the role data plays within an organization. And these differences explain the experience described above, because most best-practices and approaches are (implictly) written from the viewpoint of one and the same role (and ignoring different approaches that are required for the other roles).   So, what are we talking about? Depending on the role data plays within an organization, three types of organization can be distinghuished as depicted in the table below:

 

 

Type I

Type II

Type III

Role of data Product is data Product or service is expressed in data Product or service is supported by data
Examples Credit rating agency, market industry data providers (Nielsen, IRI), statistical bureau Commercial banks, Insurance companies, Tax agencies Retail, logistics, agriculture
Business process data processing & management = primary process  hybrid data processing & management = secondary process

The first thing that becomes clear is the role data plays with regard to the product or service an organization provides. This has a direct implication for the implementation of data management and data governance as this will differ between the types. This is visible in current discussions whether or not organizations need a CDO. If you’re a type I organization, the answer can (and maybe should!) be no as data processing is your primary process. In this situation, the COO or even the CEO is the CDO. The opposite is a type III organization where data supports primary processes. Here, it makes much more sense to appoint a CDO or some equivalent function.   Another example is data quality management. In a type III organization, data quality is managed to ensure that the data that is used in primary processes (e.g. order-to-cash) is fit for purpose (e.g. information with regard to prices, stock levels, delivery schedules, customer information etc.). Now, in a type I organization, data quality management actually is part of the primairy process as the process chain itself consist of receiving, processing and delivering data. Checking data quality is often a process step in such a chain. So typically, in a type III organization, data management-processes are supportive and operated by staff positions whereas in a type I organization it is operated by line positions. This distinction leads to different process designs, governance structures and organization.  

The type II organizations are a mix of both worlds. They often contain data logistical processes (type I), for example in financial transaction processing as well as customer facing processes (type III). This leads to a certain blend when implementing data governance and data management. For example, in financial organizations, accountability for data is often allocated to the CFO.  

There are more interesting analyses to be made, using this typology. For example, an organization can inhibit multiple types at the same time. Consider a BICC-department in a hospital. The hospital itself is a type III organization, whereas the BICC-department is a type I. What about Amazon? The Kindle-books or Amazon Cloud Services are type II, whereas their webshop qualifies as type III.   It’s necessary to emphasize that type III-organizations cover a very wide range, in terms of the level of support data provides for primary process. I’ve worked for an organization involved with asylum seekers. They do need data and its quality is important but only to a certain level. Other conditions to fulfill their primary service (of providing housing) are much more critical. This is the low-end of the type III-spectrum. The high-end covers those organizations that rely very heavily on data, like Amazon, big food retailers like Wal-Mart or Ahold and parcel service providers like UPS or PostNL.  

Another interesting feature is inheritence. By this I mean that a type I organization is also a type II which in turn is also a type III organization. Consider a type I organization like a credit rating agency. Their primary process is processing data to calculate and provide credit rates. These primary (data processing) processes are supported by…. data. One needs to know details about the data providers, schedules when what kind of data is deliverad, customer data, delivery agreements that automatically drive data processing etc. Managing this type of data is type III.   I hope this helps and speaking for myself, it has greatly enhanced my understanding of the role of data in organizations. By now, it should be clear that data plays different roles within organizations.   A final word of advice: in my introduction I referenced available knowledge and literature. Very much of what I’ve read doesn’t differentiate the role of data and implicitly assumes a type III organization when describing concepts, theories and practices. This is what caused the strange feeling I mentioned earlier. Be aware of this potential pitfull when putting the practices to work in your particular context.