Big data decisions: Why C-level is wrong to tell IT to butt out

When it comes to exploiting data, businesses are generally ineffectual — and it's a failing that isn't helped by attitudes to IT teams and poor communication.
Written by Toby Wolpe, Contributor

IT departments have been told so often they don't understand the business that many are now fearful of standing up against bad C-level technology decisions.

That reticence, added to a lack of access for IT to top executives, is hampering businesses' ability to make the most of data. Only 11 percent of CIOs feel strongly they have enough access to business leaders, according to research from Hitachi Data Systems.

"The biggest problem with IT is sometimes we are scared to challenge the ideas that come out of the CEO or the CFO," Ray Ford, CTO at accident management service provider Accident Exchange, told a recent Hitachi round table in London.

"You go in there and say, 'Hang on a minute. That is just not a good idea — as we say in polite circles. They'll think, 'Well, what the hell are you? If we're not here, you won't be in business'."

Even in areas such as which data can be legitimately mined, flaws and issues can be traced back to the earliest stages of a project, according to Ford.

"The problem is that is that many projects are kicked off with a fait accompli handed to IT. 'This is what you're doing. Here's the spec, go do it'," Ford said.

"No one says, 'Actually, if you do it like this, you can dissociate this type of data, which will hang you from a compliance point of view. You can put an index there, you can put that there. If you do that, you can derive value.

"You can keep some bits for a long period of time but you can abstract away or delete or remove personal data, which means you're compliant."

Hitachi's research, conducted among 200 IT decision-makers in UK organisations with more than 1,000 staff, also suggests that 73 percent of businesses are failing to exploit their data because of the way it is stored and managed.

Nine out of 10 CIOs also think IT could be doing more to support business leaders in using data more effectively.

Ford said this was an area where IT leaders should step up and actively show business units how analysing data can provide financially important insights before they need them.

As an example of this approach, Ford cited an exercise that involved dividing the UK up into 20-mile radius circles and looking at every car delivery in each circle over the past 15 years, asking where each vehicle should optimally have been delivered from.

"It took about a month of processing — we borrowed a few machines — and we gave the answer, 'That's where your depot is. That's where your car should be delivered from. No more like this.

"The the CFO said, 'I didn't ask for that'. I said, 'No, but we decided you needed to know and that's your strategy going forward'."

Ford said there is sometimes a gap between business functions and IT but it's nothing that improved communication cannot solve.

"Talking about customer metadata when somebody doesn't even know about data is like trying to put somebody on Mars who doesn't know what it's like to go to Europe," he said.

"It is a big divide and IT is guilty — we go into our silos. We don't explain it in words that ordinary people in their day-to-day job can understand. We're not very good at that.

"There is a gap but it's relatively easy to close if people have an open dialogue, a civilised discussion. It's more to do with people than it is with technology."

But Ford said it is also important that business leaders understand limitations on the exploitation of data in terms of relevance and legality.

"That's a big problem from a legal point of view. Some data you can only use for the purpose for which it was gathered. You haven't got a free hand to go away and mine it," he said.

The historical relevance of data also needs to be taken into account when it is being used in analytics.

"You've got to bear in mind that some of your data was gathered four, five or six years ago. The business environment today is a different business environment to what it was then," he said.

"It's an historical view. It's fraught with difficulties using history to predict the future but if that's your best way you can do it that's fine but you've got to be careful."

Ford said not all data needs to be retained but what does matter is context and metadata.

"Some bits we need to think about but some bits we've got to get rid of and what matters is the metadata," he said.

"If I've got the metadata of a Word document that's more important than the Word document unless I need the words or it's on legal hold or it's on a particular trial track and we have to keep it for the next 20 years or whatever it might be. Metadata is everything. The context is what matters not the actual file."

 Read more on Big Data

Editorial standards