Want to know a secret? I loathe the phrase Digital Transformation. Not only is the word Digital silly and misleading, but Transformation ain’t much better. The mere act of naming a business initiative Digital Transformation suggests to people that once the transformation is complete, we’ll be done. Look at us! We’re Digitally Transformed!
Hogwash. A single transformation is better than nothing, but what we really want is the ability to transform our business as needed. Furthermore, we need to be agile with our transformation initiative itself. The last thing we want is to institute, say, a three-year transformation effort, only to find at the end of the three years that the world changed a year into the initiative, rendering it a useless waste of money and effort.
So, who are the right people in the organization to address this sorry state of affairs? Driving the change needed to make our organizations more successful at such ongoing transformation should ideally fall to the Enterprise Architects. Unfortunately, the way Enterprise Architecture (EA) is done today falls well short of this mark.
Traditional EA starts with documenting the initial state of an enterprise. Then Enterprise Architects document the final state as required by the business. Finally, they help the organization plan how to get from one state to the other.
There are only three things wrong with this way to think about EA: The initial state. The final state. And yes, the plan to get from one to the other. It’s no wonder most organizations who were fool enough to sink money into EA believe EA is completely broken. Even EAs think EA is broken, and they do it for a living!
But no matter how fun it is to bemoan this sorry state of affairs, this Cortex must help point the direction to a better way of thinking about – and doing – Enterprise Architecture. The success of any Digital Transformation effort hangs in the balance.
It seems so intuitive that the way to solve a problem is first, understand the problem, second, understand the solution, and third, solve the problem. Unfortunately, many EA initiatives don’t get past step one. Documenting the initial state of an enterprise – or even simply documenting the initial state of an enterprise’s IT environment, which is usually where EAs focus their efforts – is an exercise in futility. There are simply way too many bits and pieces, moving parts made of people and technology, for such an activity to ever complete, let alone provide sufficient detail to add clarity to the situation. Remember, the organization is always in flux. Even if you could take a snapshot of everything, it would be out of date a moment later.
What about the final state? That’s the nirvana situation where all your problems are solved, right? OK, raise your hand if there was ever a time in your career where all the problems were solved, all the bugs were fixed, all the projects were complete. Anybody? Sounds rather unrealistic, wouldn’t you say? Even if you narrow your scope to a single aspect of your business or even a single project, there is nothing final about a final state, if you could even refer to such an implausible scenario as a state, which is asking a lot. Remember, everything is connected to everything else, and everything is always changing. The notion of a “final state” is an unrealistic simplification that misleads people into thinking the approach they’re taking will actually solve problems.
Very well, let’s say you’ve still reading, in spite of thinking my comments on initial and final state are a bunch of hogwash. Certainly, you may ask, the EAs can at least concoct some kind of plan for getting from point A to point B? Sorry to disappoint. If what you mean by a “plan” is a description of the steps you must take that you can put together before you take those steps, then no. You’ll never come up with such a plan that ends up having any relationship to reality by even the halfway mark. There are simply too many unknowns, and too many things can change along the way.
The core problem with this traditional approach to architecture is that it is simply too literal. It deals with things as they are and things as you desire them to be, without dealing with the fact that everything is subject to change. But the good news is, we architects have a wonderful tool in our box of tricks for dealing with excess literalness: the tool of abstraction. If we move up a level of abstraction, we’re no longer thinking about initial and final state. Instead, we’re thinking about moving from being less agile to being more agile.
Literal-minded thinkers (and I know you’re out there) will be protesting at this point that “being less agile” is our initial state and “being more agile” is our final state, so we’re just playing word games. And if you squint and hold your nose, there is a minuscule portion of truth in that overly literal perspective – but you’re missing the point here. “Being less agile” describes any number of possible initial states we may find ourselves in, and “being more agile” means being able to deal with change better, no matter which state we started at or where we happened to end up.
Put another way: moving from A to B describes a change (or actually, a set of changes) in the organization. Moving from less agile to more agile describes a change in how the organization deals with change, from being less able to deal with change to being more able to deal with change. In other words, we’re now thinking at the Meta level. We’re not doing architecture any more, oh no. We’re doing meta-architecture: we’re architecting our architecture.
Time to put some meat on these bones. As I’ve explained before, the Bloomberg Agile Architecture (BAA) Technique is a particular approach to implementing EA that is business agility driven. It’s not a framework you need to subtract from but a technique that complements other architectural efforts. Today I’m rolling out a core BAA artifact: the BAA Maturity Metamodel.
Try as I might, I couldn’t fit the entire chart into this article, so instead, I’ve linked to a pdf version of the Metamodel. Click here or the image below to download the pdf.
Unfortunately, there isn’t room in this Cortex to fully explain the BAA Maturity Metamodel (for that, take one of our classes to learn all the details, or drop me a line). But there is room here for a few of the highlights.
First, notice the four levels: Chaotic State, First-Gen SOA/Centralized, Next-Gen SOA/Cloud-Centric, and Agile Architecture. These designations refer to the organization’s context for architecture as it matures from less agile to more agile. In other words, it maps how your architecture is evolving – and thus the chart represents an aspect of your meta-architecture.
Second, note the twelve dimensions (rows), sorted neatly into four areas: Organization, Process, Information, and Technology. This organization of dimensions is more a matter of convenience than anything else. Feel free to rearrange the dimensions as you see fit.
In fact, feel free to change anything in the metamodel you see fit, based upon your situation – and always be willing to change it as your architecture evolves. Remember, there are two sides to being agile: the core driver of business agility (increasing in the metamodel from left to right) and the fact that BAA is Agile in approach (in other words, generally follows the principles of the Agile Manifesto). And as ye old Manifesto sayeth, thou shalt respond to change over following a plan. Even if the plan is to be Agile!
Finally, never forget what this diagram is for. It’s for helping architects coordinate change across the organization as driven by business needs. The metamodel – or any other aspect of the architecture – never tells the business what it should do. In the immortal words of a well-known sage: “business needs drive the architecture, and architecture drives the technology. Never let technology drive the architecture or the architecture drive the business.” Couldn’t have said it better myself.
I know that many of you who took the time to download the Maturity Metamodel will have many questions about it – and over time, I’ll get to the answers. But just to whet your appetite a wee bit further, let’s take a moment to discuss the first dimension: Business Transformation.
The starting point for business transformation, of course, must be “no transformation.” As the organization gets a handle on their architecture and moves to first-generation SOA, IT moves into the role of service provider – not simply IT services, but also Web Services that automate the interaction with existing systems of record. Instead of simply integrating those systems, now the IT organization publishes interfaces that can be consumed as needed, transforming the relationship between IT and the business (which was always easier said than done, hence the relegation of first-generation SOA to Level One).
As the architecture matures to next-generation, Cloud-friendly SOA (which in many cases loses the “SOA” name, but is Service-oriented nevertheless), the transformation story centers on the move to the Cloud. Cloud Computing, you see, is far more than a change in technology deployment. It represents a force for transforming the business.
But even Cloud-driven transformation isn’t the end of the story. The move to Agile Architecture is a move to continuous business transformation – where the organization is as agile as it wants to be, and is able to deal with change as a routine part of how it does business. Continuous business transformation, of course, has always been the center of the Agile Architecture story – and is what we really mean when we say we want Digital Transformation.