IT departments are having a serious identity crisis and it’s time to do something about it. I’d say it’s an epidemic. The symptoms of the affliction change over time. First, there’s a strained relationship between IT and the business. Then there’s a disconnect – people hardly speaking across business and IT. Finally, in the most advanced stages, there’s an odd acceptance. It’s sort of a “You do your thing, we’ll do ours, and we’ll all get along fine” situation. But who does what thing? Redundancy, complexity, and risk are everywhere, and ignoring the problem won’t make it go away.
This crisis has been building for a long time. As far back as dBase (an early PC programming language and database management system), and probably earlier, people outside of the formal IT department have had the ability to create their own solutions with just a little bit of initiative and a book or two. Then Microsoft took end user empowerment to a new level with MS Access, sometimes underpinned with a SQLServer database running on a server situated literally under someone’s desk. Things proliferated from there.
Were they production solutions? Well, what does “production” mean? If it means the end user could call the help desk when something went wrong, then, no, they weren’t production. You had to call Carl and he might be on vacation or worse. But if “production” means that a problem with the system could lead to a critical business issue, then yes. Some of these systems were used to write checks, manage manufacturing schedules, issue purchase orders – important stuff.
And potential disruptions in operational processes were only the acute risks. There were also deeper strategic issues. For example, how can you have coherent and integrated resources (like data resources) if there is no coordinated approach to development and deployment? You can’t.
This leads us to today. The level of end user empowerment has never been greater. Same with the risks and the issues. In addition to all the previous capabilities, end users can now whip out their credit card, stick it in the cloud, and get something up and running faster than ever. No need for the rogue server in the office anymore.
I often hear data management leaders say something like, “We really need to get control of our data – it needs to be integrated; it needs to be consistent; we need to look and act like one company. That’s what our ‘One <fill in almost any company name> Initiative’ is all about!” And then a few minutes later, “Sure, lots of our departments build their own solutions from top to bottom [and we are clearly ok with this and don’t see the contradiction between this and our stated goals.]”
I’m not saying this is all bad. End user empowerment is a good thing. Self-service reporting and analytics is a good thing. Experimental data provisioning to support prototyping and hypothesis testing is a good thing. But if skilled developers in various business areas are deploying core production data, they are basically IT employees, even though they don’t officially report into IT. In other words, by just allowing these activities to grow haphazardly with little to no conscious thought, organizations have simply drawn their IT organizational lines and boxes in different places. Actually, it’s more like random, scattered circles. Sometimes we play with words to make it easier to ignore the overlap – “Oh, that person? She’s not a ‘data integration developer’ she’s a ‘data engineer’; totally different [if you don’t think about it.]”
“Designing” an IT department this way will never result in coherence and integration. And make no mistake, that’s exactly what’s happening. The proliferation of pseudo-production development isn’t “getting around IT”. It’s just a careless re-design of IT.
So, what’s the answer? Big challenges like this usually call for revisiting the fundamentals. What should an internal service department like IT (or HR, Finance, or any other) look like? It should have a central hub for coordination and shared resource development with distributed teams supporting – and empowering – their “customers” as directly as possible.
But primarily, the cure for this identity crisis is intention followed by action. It’s not that organizations are making bad decisions about how they should be organized and which groups should perform which functions; it’s that they aren’t making decisions at all.