Home & Office

A perspective on business analytics

Am in the middle of some research into storage, and the latest technologies that help to ensure that storage doesn't exist as an island or silo. No, don't switch off!
Written by Manek Dubash, Contributor

Am in the middle of some research into storage, and the latest technologies that help to ensure that storage doesn't exist as an island or silo. No, don't switch off! It gets interesting later...

I participated in a panel discussion yesterday with IBM chief scientist Jeff Jonas, from Big Blue's Entity Analytics Group. Jonas is clearly no stranger to self-promotion, so I'm a little loathe to give him more publicity but I'm forced to admit that I've never heard the problem of business analytics explained so lucidly.

If you're not a CEO, business analytics or BI can appear -- well -- a bit dull. There's no shiny box to look at, no complex new technology that does miraculous things with piles of bits that have never been done before. There's nothing sexy about BI. Or is there?

Maybe it's because the problem appears more organisational than technological. There's lots of information scattered around the enterprise but how do you get to it?

Even if you can, how do you know what's important until it's in the context of something else? That's the gist of the message that Jonas has been pushing for the last 20 years or so.

The problem according to Jonas is that, even though computers get faster, organisations get dumber. That's because they can make sense of only a limited volume of information. But data is growing faster than the speed of data assimilation, posits Jonas, so the volume of data you can make sense of becomes a smaller proportion of the whole over time.

And it leads to business amnesia.

An example he gave was that one huge retailer found via BI that two of every 1,000 employees hired had been convicted for theft from the same store. Another example: your bank has suffered a fraud involving an outsider and an insider. The outsider is caught, convicted and gaoled, but you don't know which of your 25,000 employees is the insider. Then an employee moves to the outsider's address. Bingo!

The key here is that you didn't know that an employee's address was important until it was put into the right context. That's what business analytics does.

What's key is to assemble all the data you can, because context is all. And while I'm told there's a lot of very smart algorithmic work being done by Jonas's inventions both prior to and during his tenure at IBM, underlying business analytics is a need for all corporate data to be accessible, not in silos around the organisation. Which takes me back to storage.

I'm convinced of the need for data access: as a writer, siloed data has always been a huge barrier to well-researched journalism. Not to mention deliberately hidden data - but that's another story. Yet so many organisations seem still to be struggling with providing universal access to information (assuming correct authentication), with the finger of blame often pointed at data storage in departments and branch offices that doesn't hook up either physically or logically outside its creation point.

It highlights the need to look at both ends of the problem -- the storage and the way that the information that storage contains is used.

Is your organisation well joined up?

Editorial standards