Tuesday, April 19, 2011

Parallel Universe Syndrome?

I have become increasingly concerned that in the typical enterprise the major business improvement disciplines operate to some extent as silos. EA is often disconnected from Business Architecture. BPM is frequently not well connected with EA and Application Modernization. Governance is commonly applied at discipline level rather than being coordinated across discipline. I have blogged on these topics recently in Beware the New Silos! and Silos What Silos!

The term Parallel Universe comes to mind in which a hypothetical self-contained separate reality coexists with one's own. Although strictly we should refer to this as a Multiverse, let’s not get too technical. Think about it for a moment; disciplines are inward looking, they have their own language, industry standards, specialists, champions, culture and sponsors. In fact they have a significant critical mass of their own. Furthermore we can observe standards bodies for the various disciplines expanding their scope to encroach on the natural space of other disciplines, instead of establishing agreed boundaries and coordination mechanisms.

In the BPM world there appears to be a profusion of frameworks, all of which cover similar ground, but no real convergence. Other disciplines have equivalent frameworks, and they are similarly discipline centric. No surprise here. They all use different perspectives to manage dimensions of a common business objective. The issue is that all these frameworks need to work in a coordinated, networked manner. But in addition to the culture, language and other areas of difference mentioned above, they operate on different life cycles and timescales in a concurrent but often conflicting manner frequently with very fragmentary coordination.

I don’t believe we need yet another framework to address this question. There are a huge number of possible dependencies and interfaces, and given the heterogeneous nature of the overall business improvement environment, it seems preferable to encourage outcomes not techniques. What I propose is that because governance is typically exerted for each discipline, what’s required is a set of governance review criteria for each discipline that raises the key questions that allows the organization to monitor and govern across discipline.

In this month’s CBDI Journal I have introduced such a cross discipline governance system. I have suggested a simple approach which extends the widely used COBIT governance framework. There’s no attempt to say “how” the disciplines should work together; simply to focus attention on the outcome for the collaboration to ensure that appropriate coordination is considered.

Friday, April 8, 2011

So Nobody is Doing EA Eh!

I get rather tired of the “Enterprise Architecture doesn’t work” debate. The latest comment from ZapThink is predicated solely on the fact that “architecture must precede build”, therefore because an enterprise, by definition already exists, architecture is redundant. In the line of facile arguments this one must rank very highly.

Simplistic analogies never work. Architecture for bridges or buildings doesn’t scan. In the physical world there are any number of examples of evolutionary enterprise architecture in which the product is evolved on a continuous basis. Think airports, highways, networks etc.

For years we at CBDI have been advising (and indeed practising) how to modernize portfolios of applications and technology in order to establish continuously evolving capabilities. Of course we practice enterprise architecture, but it’s about clear understanding of the levels of abstraction that provide minimum necessary detail to plan and execute the next phase of activity in an agile, iterative process that is common to initial build and ongoing evolution. And even more important, it’s about triage – understanding what elements must have an enterprise context and what elements are entirely local.

All this bottom up and top down, inside out, emergent and so forth are so much rubbish bandied about by those that need to spend some time breaking rocks. Think continuous evolution of business architecture, enterprise architecture and solutions. Rant over!