Process and Information Architectures

Last day of the Building Business Capability conference, and I attended Louise Harris’ session on process and information architectures as the missing link to improving enterprise performance. She was on the panel on business versus IT architecture that I moderated yesterday, and had a lot of great insight into business architecture and enterprise architecture.

Today’s session highlighted how business processes and information are tightly interconnected – business processes create and maintain information, and information informs and guides business processes – but that different types of processes use information differently. This is a good distinction: looking at what she called “transactional” (structured)  versus “creative” (case management) versus “social” (ad hoc), where transactional processes required exact data, but the creative and social processes may require interpretation of a variety of information sources that may not be known at design time. She showed the Burlton Hexagon to illustrate how information is not just input to be processed into output, but also used to guide processes, inform desisions and measure process results.

This led to Harris’ definition of a business process architecture as “defining the business processes delivering results to stakeholders and supported by the organization/enterprise showing how they are related to each other and to the strategic goals of the organization/enterprise”. (whew) This includes four levels of process models:

  • Business capability models, also called business service models or end-to-end business process models, which is the top level of the work hierarchy that defined what business processes are, but not how they are performed. Louise referenced this to a classic EA standpoint as being row 1 of Zachman (in column 2).
  • Business process models, which provide deeper decomposition of the end-to-end models that tie them to the KPIs/goals. This has the effect of building process governance into the architecture directly.
  • Business process flow models, showing the flow of business processes at the level of logistical flow, such as value chains or asset lifecycles, depending on the type of process.
  • Business process scope models (IGOEs, that is, Inputs, Guides, Outputs, Enablers), identifying the resources involved in the process, including information, people and systems.

She moved on to discuss information architecture, and its value in defining information assets as well as content and usage standards. This includes three models:

  • Information concept model with the top level of the information related to the business, often organized into domains such as finance or HR. For example, in the information domain of finance, we might have information subject areas (concepts) of Invoicing, capital assets, budget, etc.
  • Information relationship model defines the relationships between the concepts identified in the information concept model, which can span different subject areas. This can look like an ERD, but the objects being connected are higher-level business objects rather than database objects: this makes it fairly tightly tied to the processes that those business objects undergo.
  • Information governance model, which defines that has to be done to maintain information integrity: governance structure, roles responsible, and policy and business standards.

Next was bringing together the process and information architectures, which is where IGOE (business process scope models) come into play, since they align information subject areas with top level business processes or business capabilities, allowing identification of gaps between process and information. This creates a framework for ensuring alignment at the design and operational levels, but does not map information subject areas to business functions since that is too dependent on the organizational structure.

Harris presented these models as being the business architecture, corresponding to rows 1 and 2 of Zachman (for example), which can then be used to provide context for the remainder of the enterprise architecture and into design. For example, once these models are established, the detailed process design can be integrated with logical data models.

She finished up by looking at how process and information architectures need to be developed in lock step, since business process ensures information quality, while information ensures process effectiveness.

3 thoughts on “Process and Information Architectures”

  1. Hi Sandy,

    Thanks for posting from the conference.

    It’s great to see this subject get attention. When you think about it, it’s amazing people ever separated process from information, you need them both in order to work in context.

    Unfortunately you see this divide of code from data all across IT and it is the underlying cause of much of today’s enterprise technology problems.

    We all accept information out of context is simply data so what does that make process without context…


  2. Dave, I completely agree. One of the areas that I cover in my presentation on BPM in an enterprise architecture context is the linkage that need to be made between different model types, especially data and process models. BPM is not just about process models: it’s also about data, organization, rules, security and more.

  3. Agreed! Technology-aside, we’re really talking about integrating work in its real-time business context to best meet business objectives.

    To implement these ideas, I think its fair to question whether BPMx is well suited for this purpose – at what point does BPMN become overloaded as CRUD system orchestration?

    To achieve more interactive BPMN you need to extend with SOA, then decompose BPMN with external events, rules, data, etc. This degree of indirection is ‘expensive’ from a hardware/software and performance perspective, difficult to govern, introduces coupling which makes it brittle and hard to change, and for all that additional complexity – only delivers an incremental gain in system expression.

    If we agree work ranges from structured to unstructured, and real-time enterprise context is key to good decisions and effective processes, then I think BPMN starts to look like the wrong tool for the job.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.