There’s a long running tension in IT between centralisation and decentralisation.
Processing more data, after all, simply requires more resources. More storage, more computing power — but the software is (probably) the same. (The “probably” reflects the fact that different products you might by are scaled for configurations of various sizes and priced accordingly — but it’s more the pricing model than the code that affects this.)
So IT tends toward centralised, consolidated solutions. Easier to manage. Easier to deal with.
Clients within the enterprise, on the other hand, tend toward decentralised decision-making, and solutions that “fit them”. In a 10,000 person company, few people actually work for “the company”. Mostly they work for their department, their function, or their location.
Not only are their loyalties more local, their expectations for work are tied to the way the people they work with work.
Now … in your personal life, you have no trouble dealing with diverse systems oriented toward different purposes.
We call it the Internet. Maybe we see it through a browser, maybe through phone/tablet apps, but we deal with diversity all the time.
Then we come to work and are determined to stamp it out. But why?
W L Gore & Associates, for instance, demonstrated years ago that keeping plant sizes small — thereby making them high-trust, high-involvement “villages” — paid off. Rather than grow in size, the company would just build another plant, both of which would be kept small.
From an IT point of view, each plant ought to have the technology that suits its needs best. Sure, there’s likely to be a lot of overlap — but not 100%.
After all, trust has a value, too — even though it seldom shows up in a requirements-gathering session.
What does this mean, therefore, going forward?
It means data unification is something that can happen apart from operating areas.
It means with an ecosystem of parts change becomes easier. So, too, experimentation and innovation.
It takes the whole “centralised vs decentralised” pendulum and kicks it off the table. A centralised service can provide for very decentralised visions.
Just like the Internet.