Nothing is guaranteed to bring the wolves to the door faster than criticising Agile. Apart from maybe criticising Enterprise Architecture. So strap in. This is going to be fun.
I recently spoke at a conference in London for IT Architects. I told them the horrible truth. That Architecture is dying. That the twin diseases of “cargo cult” Agile and Enterprise Architecture were taking their toll, and that businesses were rightfully questioning the purpose of Architecture as the hailstorm that is ‘Digitalization’ destroys the comfort of doing things the way they have always been done, and rewards companies that actually get things done.
Like the cable companies currently struggling to keep customers paying, the Architect’s insistence on fixing problems only relevant to Architects, is going to lead to en masse cable cutting. It’s time to take a step back and rethink our approach. We are concentrating on how to combine EA and Agile in yet more frameworks. This is the question keeping Architects up at night. It’s like preventing cable cutting by laying more cables.
Every industry is currently undergoing its own digital transformation. It is insane that Architects don’t see this coming to our own industry. We are clinging to our frameworks and ideas the same way that the music industry is clinging to copyright, in the face of all evidence that things aren’t going our way. The seesawing arguments from the perspectives of EA and Agile are allowing bad engineering to exist in a vacuum –something that impacts credibility on all sides, and something we need to urgently address.
The shocking thing was not the resistance of the audience to these ideas, the emperor was always going to protest that he was indeed fully clothed. The shocking thing was that with a few minutes of logical argument, heads began to nod. That Agile and Waterfall are very similar ideas (building static systems to answer dynamic problems, with a slight disagreement on the nature of planning and execution), with no real empirical evidence on any concrete axis to prove Agile’s efficiency, was accepted. That Enterprise Architecture, after 30 years, still doesn’t even have a definition anyone can agree on, let alone a sense of purpose. It doesn’t matter about your anecdotal evidence of how great your EA practice or Agile model are. Architecture is being questioned, and dismissed, at the point in time where it could provide enormous benefit. Unless we step up as a community, architecture will become a by-product of business analysts and scrum masters talking to each other. Architecture will become accidental – and it will be our fault, for frameworking while Rome burned. The proof of this is the rise of Shadow IT. The business wants another way, and the consultants smell blood in the water.
Someone will point out in the comments that I’m missing the point and that EA is about much more than IT and is strategic “something something” alignment. That’s a discussion you need to have with your CEO, I’m trying to make an entirely different point here. Architecture is dying, because the “either/or” focus on EA and Agile is preventing the real work of Systems Engineering from being done.
Digitization projects are suffering from this lack of Systems Engineering. The audience was in total agreement on this. In review after review we see the same things. Is this system scalable? Answer: Yes. How scalable, at what cost? Shrugged shoulders. Somewhere along the road we forgot to do engineering. We’re doing Agile programming. We’re doing EA. There’s a gap, and no one seems to want to or be able to fill it. So we have lots of people with the title “Architect” and a system that will fail. Who is responsible?
Businesses are being seduced by Shadow IT. There’s a reason for that. The whole world knows that technology has evolved. Cloud is not a secret. Business people know that the lead times on IT projects aren’t inevitable. Clever consultants tap into this frustration and give the business what they want. The vendors know this too. The market is converging around Shadow IT. And no one there is doing TOGAF. Centralised governance for the sake of it is the primary driver of the frustrations that lead to Shadow IT. And Shadow IT is the model to beat. Business loves it, IT Consultants love it, what’s not to love? It’s easy and simple. Like Napster, Netflix, and Spotify. Architects are still stuck suing people for ripping MP3’s, in denial that they can stand still while the world around them changes dramatically.
‘It will work; you’re just not doing it right’ you say. This is the erstwhile cry of Agilistas, EA enthusiasts, and 20th century communists.
When it does work, when something gets delivered that works and delivers benefit (the only measure we should be using), it’s usually because the EA or the SA stepped up to the job of Systems Engineering. Making sure, validating and verifying, that something is the right thing to build, and that is built right, is the only real provable value that an architect can bring. There is no agreement on where this function should be executed, and maybe there doesn’t need to be, but we need to be honest about the gap. The crafty developer who smuggles real architectural work into Sprint 0, or the EA who refuses to sign off until quality attributes are verified and validated, are the heroes here.
We need to become more efficient. Fewer ivory tower architects and more doers is a start. Solid Systems Engineering that can deliver quality is a must. Self-titling needs to go away. Certification needs to be a standard thing, as in all other industries. Most of all, we need to deliver, deliver, deliver.
We need to separate the practice of creating strategy and advising the business on business matters, from engineering, and engineering from programming. We need to be clear in every project where engineering responsibility lies, and how we answer engineering questions. Only then will we have something that can presented as having value. So now we get to choose. Do we let the cables get cut, and hope that Netflix steps up to pay for season 10 of TOGAF & Agile? Or do we proactively take steps to make sure that we give our customers what they actually want?