Centralised or decentralised?

Periodically, over decades now, I’ve had customers and service users of KM ask that “everything be in one place”. They’re mostly talking about documents and other similar resources and tools, and what they want is to see everything they need at one time all readily to hand in one view.

And who wouldn’t?

Yet this is one of those oh so very easy to ask for requirements that bears a bit more scrutiny.

If the customer/user themselves has an idea exactly what “everything” they would need on a particular occasion was, or if this could be analysed, then the chances are that this is a repeating process or situation that calls for a custom solution that fits it exactly. And this is usually very doable.

However, mostly this request is a vague and hopeful request that everything they might need on a special case occasion be somehow assembled in that “one place”. This is really hard, tending impossible.

The repeating, common case situation can have a solution made to fit. But the generic case of searching for all the tools, contacts, resources etc that you need on a one-time, special case situation is different.

Think about it. The good ol’ Internet has pretty much everything in it, but it’s not all there waiting right in front of you for the specific need you find yourself presented with at any moment.

So the concept of just having everything in one place seems like the answer. After all, if everything is there, then everything that you need (if it exists or is at least available) should be there. And so we come to the dilemma of centralising everything in the most general stores and pots and containers that you can, vs. specialising the curation, storage and presentation of resources to particular groups and particular uses (as in the first case).

There’s a bit of a paradox here because, as we saw in the case of the Internet, pretty much everything (at least, everything that you have or that is available) is in that one place, conveniently all accessible from your one screen and search engine. Yes, it’s distributed at many levels of abstraction, but you don’t need to worry about that. But the paradox is, whilst one might argue that everything is there, it very much seems like it isn’t and that you have to find and assemble it.

And it makes real sense to divide resource between separate stores and pots and between different curators, so that each resource can be maintained and developed and optimised in its own terms.

But sometimes making that decision is just a judgement call. Should all my content of type ‘Published report’ be in one Reports library, tagged and segrated by subject, or decentralised and distributed between the different subject matter groups (communities, experts) who care most about it? There’s a judgement call and the best thing is to at least consider the question as to whether the better outcome overall, not just for the customer/user staring at you right now, accrues from a centralised, “all-in-one-place” model or a decentralised, “each curated by the people who care most” one.