Lies, damn lies and statistics

Architecture Rant


One must occasionally rant. And when that cathartic wave has crashed and some of the built up cynicism washed away, hopefully the receding waters leave a perspective for a pragmatic path forward.

For me, enterprise architectures should represent the very best of what we can and should strive for, encourage the business and staff to reach higher and push us to see beyond the boundaries and limitations of the current landscape. I guess what it comes down to is that there's good IT and bad IT and we have become overwhelmed by a sales machine - both internal and external, which advertises one but actually delivers the other. Its not a matter of subscribing to the latest method, programme or ethos - it's about understanding the subtle idiosyncrasies and the sympathetic treatment of people, process and technology. Leaning into the complex problems, and sometimes having the courage and integrity to defend work which doesn't have an immediate payoff.

I'm a natural contrarian, and I love putting early thinking out and hearing other opinions and will heap praise on anyone who challenges or improves on the ideas - even if it doesn't end up swaying my opinion. I value the honest, plain speaking of other IT experts who have the courage and talent to engage with things that we don't know the answer to yet.

I sometimes get criticized for using emotive language. However, I do this for a reason and an architect should know when to challenge and when to inspire - both are done most effectively through emotive language. Far from being something to change or neuter, I see this as an absolute strength. I never use an ad hominem turn of phrase or belittle others and I make a point in being able to backup every claim I make with more than being a google search ahead.

A constant deluge of busy work, 'quick wins' and bureaucracy, forces architects into the weeds and it feels the architecture practice rarely has the trust of senior management and the exec who rather engage with consultants instead of valuing their own staff.

It's kind of sad and disheartening to watch tired rehashes of buzzwords tarted up and sold to management, clinging on through a paper thin veil of legitimacy, hoping no one lifts the skirt before the contract is signed. The same promises and haloed reverence that used to be attributed to the likes of Agile, data-centric development or SOA are now bestowed upon Micro-services, SAFe, Block-chain, Service Models, Spotify, RPA - anything that can be wrapped into a product and sold by the IT equivalent of a life coach.

It's an effective strategy that I've tried to use myself - notably Data Mesh, but I'm no salesman so it generally falls flat. This is a real catch-22 as the best architects often make the worst salespeople. Architects can become jaded and cynical through constant restructures and new management trying to make their mark. Constant change means there's no time do do the vastly more important thinking and discourse which was the reason people across the corporate world used to be attracted to the architecture disciplines.

It's only through the lens of time and experience that we can see the wood for the trees - the repeating patterns, the hollow promises of product sales teams, initiative slogans and consultant shenanigans motivated by what can be advertised as a success story for the lowest price point. This is the consultant version of the discussion by R.M. Bastian in Look Ma, It Works!

The reality of what should be done is not simple or productizable. Quick wins, MVPs and tactical solutions imposed on an already bloated and creaking IT landscape in the name of some mission statement or laid on the altar of Agile, looks good to the board and scratches an immediate itch but does nothing to alleviate the underlying issues.

Architects have become brow beaten or have a naivety about how the business will de-scope the strategic solution or remediation of technical debt with carefully contrived excuses and priority shifts. We have to be better than this.

blog comments powered by Disqus