One of my first jobs in IT was doing PC installations. It wasn’t the most skilled work in the world but, back then, connecting a video, keyboard, mouse to the CPU unit and checking that all was working was beyond most office workers. As I emerged, dusty and slightly sweaty from under the desk, the person who occupied the office where I was working turned to me and smirked. He pointed to his 1-2-3 spreadsheet and said, “Now that we have these machines, we won’t need you computer guys any longer – we’ll all be doing our own IT.” Well, that was *cough* years ago and I’m still doing IT, albeit with a bit more expertise than I had back then. And we can rely on users to install their own PCs and software.
I was reminded of this episode when I attended Radius Business’s event that posed the same question. The premise is that since IT is so easy, commodified and consumable these days that any old Tom, Dick and Sally can be their own IT expert. The death of the IT departments is imminent.
I do subscribe to the view that things are easier now – but at the same time they’re also more complicated and that’s why I’m still sanguine about my future in Information Technology. Like the answer to the question, “When will we get nuclear fusion?”, the demise of the IT department is often foretold – but it’s always 50 years from whenever you ask the question.
One of the panel speakers, Alan Young, made the point that architecture is going to be key to successful IT strategies in the future. The more I think about this, the more I agree. Let’s take the next big things in technology in turn and see why.
First of all, there’s cloud. The opportunities with cloud are only just becoming clear. It’s already a given that fewer and fewer organisations will run their own hardware but increasingly they’ll also rent their applications and the processes that run on them. The attractions are obvious: avoidance of capital expense and ability to scale depending on demand. Also, you can afford a slice of the time of a real IT expert to run your hardware and software. If you’re a start-up the benefits are even stronger – you get a full suite of applications right out of the gate. And those applications can be the best in class available. That’s why big companies like IBM see the IT decision making increasingly passing to Line of Business executives. The Chief Marketing Officer selects his/her marketing automation applications while the Chief Financial Officer chooses the best business analytics for his requirements.
But, as every large company knows, choosing individual applications is the easy bit. Besides, some departments have been doing that for ages: the finance department has often chosen its own ledger package. What’s difficult is making the applications and processes work together. Cloud is likely to exacerbate that problem, although it also provides the solution: by exposing the application programming interfaces for each of its components. IT departments (not business departments) will be able to weave these components together to create interfaces between those individual cloud-based applications. That’s an architectural way of doing IT.
Then there’s analytics. The drive is to get more and more up-to-date analytics, with the real goal to have a complete picture of what’s happening right now. As Clayten Christenson has pointed out – the problem with most data is that they tell you about what happened, not what’s happening now or going to happen (for that you need predictive analytics). In the somewhat clichéd phrase, it’s like driving a car through the rear-view mirror. Personally, I think some of this is a bit over-egged. While I don’t dispute that knowing what’s going on right now is incredibly useful, a bit of distance from events is useful for understanding what happened. After all, that’s why we study history; the web of causes that contribute to events can only be fully appreciated with hindsight.
On the other hand I do remember the periodic nervousness I would experience when I was responsible for data loads into our warehouse. On the morning after the load was supposed to take place, I would come into the office with trepidation – because sometimes it would have gone wrong. The really crucial loads, for example to enable financial month end, used to be supervised in real time. It was always an all-or-nothing load.
The optimal solution has always been to have the operational transactions post directly to the data warehouse at the time they occur. That way the data warehouse would be up-to-minute accurate. But that’s only possible if you are able to create the interface between the operations systems, create any transformations that you need and load directly to the warehouse. That’s an architectural solution.
A lot of enterprises are embracing mobile. The goal is not just to serve their customers and staff on more devices and to set them free from their desks and offices. Those customers who are implementing mobile in the most interesting ways are using it as a way to bridge between applications. Successful mobile implementations concentrate on what people want to accomplish – in other words, they take a task-based view of application design. A single transaction might very well span several applications. The thinking that’s needed to achieve that is, guess what?, architectural.
Then we come to the last big thing in technology today: social. At first blush, this really doesn’t seem to have much of an architectural component. And, indeed, for many businesses it’s something you do on the side of your real job. But those enterprises that really get social realise that it’s the oil that lubricates all business processes. If you can integrate the knowledge capturing and sharing into your processes, if you can make expertise available to anybody trying to carry out a task at the moment it’s needed, then you can make a significant difference to how easily work gets done in your company. Social systems are applications and integrating them with your other applications is, guess what?, an architectural questions.
So yes, some IT roles can now be done by anyone – software selection and consuming the applications you need to get your immediate job done. But the tough stuff has actually got both simpler and more complex. It’s simpler because the applications are building in the plugs that allow us to interconnect our processes and data flows. And when they do that on the top of open standards, that’s even better. But it’s complex because the number of applications and the places that they interconnect are getting more numerous. So, do I think we’re all IT experts now? Yes we are – but some of us have the architectural skills that I hope mean we’ll continue working in information technology for some time to come. Like Mark Twain, I think the news of the demise of the Information Technologist has been greatly exaggerated.