Clearly we want one. A thread on LinkedIn a couple of months ago attempted to define the value of EA, and produced a tirade of over 1,300 entries! But while individuals were busy chatting, the Enterprise Architecture Research Forum took a different approach. This body is a collaborative including the Open Group South Africa, the Meraka Institute, Real IRM, Telkom, Unisa and the University of Pretoria (link).
This group started with a “definition of definitions.” In other words, they thought about the requirements of a definition before producing the solution. (What a novel idea :-). The requirements that they used come from a paper by Dr. Sam Vaknin. Dr. Vaknin’s definition of definitions indicates that a definition should explain the meaning, use and function, essential characteristics, and the differentia of the concept. (link)
Using this model, the EARF created the following definition:
Enterprise Architecture is the continuous practice of describing the essential elements of a sociotechnical organization, their relationships to each other and to the environment, in order to understand complexity and manage change.
Let’s compare that definition to the “compromise” definition that emerged out of the “160 character challenge” on LinkedIn. The LinkedIn thread that I’m referring to started as a post, by Kevin Smith (author of the Pragmatic EA Framework, or PEAF). Kevin asked participants to define the “value” of Enterprise Architecture in a message short enough to send in an SMS text message. While the length constraint was arbitrary, it was useful for insuring short responses. Kevin then collected the data and, in a fairly rigorous process, analyzed it to produce a more succinct definition (link).
Using the responses as a guide, Kevin found that three different aspects of Enterprise Architecture were appearing over and over in the various definitions so his construct includes all three. He refers to these aspects as the “why, how, and what” perspectives. Kevin’s combined definition follows:
The purpose of Enterprise Architecture is to enable an enterprise to realise its Vision through the execution of it’s Mission, whilst enabling it to respond to change and increasing its effectiveness, profitability, customer satisfaction, competitive edge, growth, stability, value, durability, efficiency and quality while reducing costs and risks by Strategic Planning, Architecture and Governance supported by a Decision Support framework in the context of aligning all parts of the enterprise using Models, Guidance, Processes and Tools.
Now, to be completely fair, I participated in Kevin’s “160 character challenge” on LinkedIn. I do not have a single definition of Enterprise Architecture… I have three. Just as a dictionary may show you many definitions for a word, I have found that the term Enterprise Architecture is used in three different ways: As the name of a business function, a reference to a team of people, and as a reference to a model that describes an enterprise.
When creating my submission, I was not being particularly rigorous, so I see flaws in my definition as I type this blog entry. (to whit: EA is not limited to businesses). That said, I will cite my original words rather than revise history. Note: In order to fit into Kevin’s challenge, I submitted each of my definitions independently. Recombined as a single definition, my contribution was as follows:
-- Noun 1. A business function that collects and manages business information for the purpose of improving the way that a business responds to current or future challenges and opportunities. 2. A rigorous model of the motivations, structures, information, processes, and systems of an enterprise created for the purpose of decision support. -- Adjective (used with object) 3. A team of influencers and thought leaders within an enterprise chartered with understanding, optimizing, and improving the way the business operates.
Is there a right answer? Which of these is a better definition? Which misses the point? Could we improve one of these entries and recognize it as canonical? Is there a better definition of Enterprise Architecture, and if there is, what would it be?
What is your opinion?
The elements that get mentioned are description , understanding and management of change. I guess the definition does cover codification and rendering explicit the description of systems but it is also important is the collaborative part of enterprise architecture development and the aim of achieving coherence
hemm..The purpose of Enterprise Architecture is to enable an enterprise to realise its Vision through the execution of it’s Mission.
Hi Guna88: "to enable an enterprise to realise it's Vision through the execution of its Mission" is generic, and does not differentiate the EA function from any other business function. The Sales team enables an enterprise etc. The Customer Service team enables an enterprise etc.
Please read the referenced sources for full analysis of a definition and see if you would agree that your definition meets the criteria that Dr. Vaknin describes, or that Kevin Smith discovered through analysis.
Hi Nick: I'm very happy with all three of those definitions (though slightly surprised that you described 3. as an 'adjective', because that definition [a team'] is kind of halfway between a noun and a verb, surely?) What I'm delighted to see is the _absence_ of any reference to IT as such - that's a big shift from where the industry was even just two or three years ago.
I _really_ like the point about 'definition of definition' - not surprised it comes from the RealIRM crew, to me they're some of the most perceptive and forward-looking in the industry.
In my own work I tend to describe EA as "custodian of a body of knowledge about enterprise structure and purpose". It sort-of incorporates all of your definitions, but misses the very important point about dynamics, about guiding change, and also the necessary emphasis on rigour. So although your definitions are longer, they're probably better and more useful than mine.
Useful and thought-provoking - many thanks.
I have to say that this MSDN blogging framework is pretty primitive. I lost twice my whole reply when deleting a small section. I give up. Read my blog.
Hi Adrian, Sorry for your frustration. The software does do weird things with very long responses. I'm watching your blog for a response. Looking forward to your insight.
Hi Tom, Thank you for the feedback. The 'very-short-form' of the definition of EA that you propose is useful, especially in situations where brevity is more important than purity. (e.g. Business :-). I like it.
I've never understood why we don't define EA in the most obvious way possible -- an enterprise architecture is the architecture of an enterprise.
Admittedly, this just pushes the interesting questions (what is architecture? what is an enterprise?) under a different part of the rug, but regardless of how you answer those questions, it still accurately defines enterprise architecture.
Indeed, I find that in most proferred definitions of EA it's difficult to tease out the enterprise part and the architecture part.
I had a lot more to say, (about the ways we use the phrase "enterprise architecture", and about the EARF definition in particular) but it all disappeared when I tried to post it, so I'll stop here for now.
I agree with most of the definitions, but there is another important point most of the definitions missing.
Enterprise Architecture is defined by most of the Architects is Uni-Directional like improve efficiency, control, profitability, value etc as if no end for that. Don't forget one more important aspect. For every aspect there are positives and negatives.
An Enterprise Architecture must survive in situations in least/worst case scenarios too. And the real stoppers for all positive terms you used above are "Set of Constraints", who take active part in most of the situations and definitely over period of time. These constraints are regulatory, budget, skills, an undocumented political views etc. Before I assess any architecture value or survivorship in ever changing environments, I consider to measure how much each constraint will contribute to Enterprise Architecture success factor.
Well, at least here in Holland every client I've worked with (and I've been working with some of the biggest companies there are) sees EA as an IT function, whether we EAs like it or not. So maybe that's the to-be conceptual definition, but I don't think it accurately describes the reality of what EA is today.
And I'm not sure these definitions would be meaningful to someone who's not already familiar with EA - they're so high level. They certainly wouldn't pass the "explain it to my (really smart but not technical) wife" test. The closest I've come to one that does is:
I look at what the business is trying to do, model the different parts of the business (the information they use, the tools they use, how they're organised, etc.) and figure out how to change them so that they work together better.
@Len: If you look at Dr. Vaknin's definition of definitions, you would see why the definitions are all longer than "architecture of the enterprise." There is a simple elegance to your idea, but unfortunately, it wouldn't meet the requirements.
@Sree: I grant that one of your practices includes considering the negative constraints, but it is not clear to me how (or why) you would want to change the definition of EA to include that particular practice. Is that practice so much more important than the other practices that you would consider it "definitional?" Would you make the claim that it is not possible to practice "true" enterprise architecture without the single practice you illustrate?
@Derek, Most clients in the USA also see EA as part of IT, but the place where EA lands is not the same as it's definition. (Your explaination does not say that EA is in IT either.) The explanation that I use is even simpler: "I draw pictures of business structures to make it easier for business people to fix things."
Interesting that I like all four :). Kevin's canonical definition seems to be exactly that. It has everything you could think of wrt the stuff we try to do every day. Like a description more than a model. Yours on the other hand are a bit simpler and maybe a bit more tactical or usable. More like a model than a description.
I don't know.. I was just thinking about it a bit. Could be wrong. I posted something on linked the other day asking members for an opinion of what a good "Mission Statement" would be for an EA team. I feel like your "Adjective" definition could actually be that. Hmm....
dang.. my coffee is cold now..
Thanks.. GREAT POST.. you are my FAV.
So, I participated in that LinkedIn thread as well. I shared a prior tweet of mine, “The ultimate outcome of Enterprise Architecture is change-friendly capability delivery”.
Interestingly, most replies in that thread interpreted “purpose” as describing the function, rather than the outcome. I find this problematic.
I think the number one question Enterprise Architects and Enterprise Architecture Practices need to answer is “What do we contribute to the business”. What is the ultimate outcome of Enterprise Architecture? And therefore, what would be missing (or more difficult) without Enterprise Architecture.
For me, it’s getting to “change-friendly”. If I’m reading correctly, your definition and the EARF contain a similar theme. So, it seems purpose-wise, we coalesce on enabling change.
Additionally, I agree with the EARF purpose of reducing complexity, and as Aleks Buterman often calls out, EA plays a large role in technology investment management.
So, I propose we think of EA as a business and work backwards from the desired outcomes -- ease of change, reduction of complexity, and better technology investment return.
To achieve those outcomes, what capabilities, policies, people and tools are required. And then, how would we describe (classify) that?
Would it be a rev of Enterprise Architecture? Something else? I know it’s not the Linkedin Group output.
Your comrade in the EA revolution,
@Brenda, As always, your comment is too insightful for me to respond in the thread... I'll have to grok it and respond with a new post. I've thought a good bit about "what we contribute to the business" and there are soooo many twists to answering that question in a way that encourages the development of the EA discipline without derailing it in politics.
@Mitch: think you for your kind remarks! It is interesting that you mention the mission statement. Our EA team just went through a process of creating a new Mission and Vision statement. As soon as it is baked, I'll probably blog it, along with my insight into the process of creating a mission and vision statement, and how that applies to EA practices.
Good articale as ever. I like the first statement although something that I do not think comes out of it (or out of yours) is the forward looking and planning aspect which I regard as crucial. They seem to be a documentation of the as is state.
The second definintion seems a bit long winded and, dare I say it, looks like it was written by committee. It seems to define what an enterprises vision would be, eg improve profitability, but how would that work for a non profit making organization, e.g. a charity? It does I think though show that the future landscape and route map are part of EA.
The first definition uses the word "sociotechnical" instead of enterprise, how far do you think EA can be taken out of the technical environment? Is it valid in an environment with no IT or no technology?
@Erik, Excellent points. You are correct in that the definitions say very little about one of the key values of EA (in my book): the consideration of new innovative ideas, both technical and non-technical, in the generation of new business strategies. That is a very valid criticism.
The second definition, while it was written by a single person, was an intentional compromise composition of the concepts and memes submitted to LinkedIn. In effect, it is worse than something designed by committee, because a committee can at least work to reduce the definition... no committee existed in this case to produce a more viable compromise.
The point of "sociotechnical" is to describe the elements that are being captured by the EA function... they are both social and technical elements in a single interrelated model. This does not exclude technology but rather includes it in the business context.
One valid consideration in the use of that word (sociotechnical) is to recognize that the boundaries of an enterprise architectural model do not have to be defined by the boundaries of the enterprise. The model can extend to include many aspects beyond the enterprise, including competitors, partners, customers, suppliers, influencers, and potentially influencial proposals for legislation and regulation.