back to notes

Companies need to cut through big data hype

May 9, 2013 | The Financial Times |Michael Skapinker
Along with more than 1,000 other north Londoners, I recently lost my landline. Others were worse off: they lost their internet connections too.

Throughout the disruption, which lasted 10 days, many of us knew more about what was happening than the BT people we contacted.

At first, the telecommunications company's call-centre said the problem was outside my house. When, after four days, and in response to increasingly angry messages, a BT functionary called me on my mobile phone, all he could tell me was that the company had suffered from vandalism.

By then we all knew, from the technicians who arrived three days after the incident, that a group dressed as telecoms engineers had dropped down a manhole, cut the lines in several places to get at the copper, and had fled empty-handed when the police arrived.

A BT spokesman eventually confirmed all this (except for the thieves' disguise, on which he declined to comment). He said cable-cutting triggered an alarm to the police, which is why they managed to prevent theft but not extensive damage.

When I asked why this information did not appear on BT call-centre screens, he sheepishly agreed the company should share what it knew more widely.

This is a modern pattern. Companies have never had so much information - on sales, weather and service breakdowns. And yet the more data they have, the less they seem to know.

People on the spot - local technicians, shopkeepers or bank managers - understand their patch. When companies began to centralise information, they lost detail, nuance and often the basic facts.

We should remember this as management and information technology consultants hype "big data" and attempt to panic companies into spending millions.

"Analytics has arrived, is in wide use and is securely on the C-suite agenda," says an Accenture report .

McKinsey Quarterly adds: "When companies inject data and analytics deep into their operations, they can deliver productivity and profit gains that are 5 to 6 per cent higher than those of the competition."

Yet the same consultants report big data is a big disappointment. Accenture says only 22 per cent of companies are very satisfied with their analytics programmes, 35 per cent are quite satisfied and 34 per cent are dissatisfied. Only 39 per cent say the data they have are "relevant to the business strategy". Companies have accumulated "a large set of metrics", Accenture says, but adds: "Most organisations measure too many things that don't matter, and don't put sufficient focus on those that do."

McKinsey agrees that "companies are buried in information" and are struggling to use it.

What should managers do? The consultants recommend they intensify their efforts. Accenture suggests companies start "embedding analytics into business processes in a robust, industrialised way". McKinsey advocates "a large investment in new data capabilities". And although they are too sophisticated to say so, the consultants clearly know the people who can help.

The ability to study large trends and patterns is clearly useful. But it is no substitute for getting around the business and asking people, particularly frontline staff, what they think is happening.

The data can tell you how many customers you have lost each month, but not necessarily why they have gone. You can do online surveys, which some may bother to complete, but you would learn far more from contacting and talking to some of those lost customers yourself.

Computerised weather forecasts and sales data can tell you how many people are likely to be in your shops this weekend, but experienced store managers were making those calculations long before big data and iPads were thought of.

The same goes for predicting individuals' desires from their online behaviour - what Accenture calls "the market of one". I have been an online shopper for 15 years and find vendors' suggestions and targeted discounts plodding and predictable. The Financial Times cafeteria staff know more about my likes and dislikes than Amazon does.

Big data can be useful, but only if you know what to look for and how to get it to the people who need it. You do need to know your business well to do that, but I suspect you already know it better than any IT or management consultant does.

michael.skapinker@ft.com

By Michael Skapinker


last updated may 2013