Tags Entrepreneur, Data management, Innovation

Turn data into gold

You’ve captured the information and figured out how to let everyone access it. But how do you do anything with it? John Hutchinson finds out.

How much have you invested in gathering and storing data about your customers, products and employees? Probably a lot more money and manpower than you care to think about. And what’s your return on investment from all those pieces of data, which are growing exponentially, both in volume and complexity? Probably a lot less than you would like it to be.

Conclusion? Data is expensive stuff. But it has no financial worth until it is transformed into something else: useful information. Only then does it become a real business asset – and potentially the greatest revenue earner your company can possess.

However, transforming data into useful information is only part of the process. You need to turn it into something even more valuable: business intelligence, ie information that is organised, accessible and relevant. Then the alchemy is complete; the base material of raw data is transmuted into pure gold. Business intelligence, like military intelligence, is increasingly the deciding factor in who wins and who loses.

Companies with the best intelligence have a strategic advantage over those with a less clear view of the business battlefield, outsmarting and outmanoeuvring competitors, and using fewer resources to identify and achieve their objectives more quickly.

Sounds good, but you’re already spending loads on data systems and suppliers, who assure you that the reports and information they create from the oceans of data in their possession are absolutely state of the art. Or are they? Sarko Apcar, managing director of business intelligence specialist the Dot Group, can cite plenty of cases where companies are faced with a flood of uncontrolled data, unable to exploit its true value.

“We recently did a project for a TV channel’s airtime sales department, which had lots of people trying to sell advertising schedules to a very wide range of companies, and needed to put together bespoke packages for each client,” he says. “To create each package, they had to deal with enormous amounts of data from several sources, including viewing figures, demographics, programme schedules and the clients themselves. The data overload was so great that it took an average seven hours to develop each proposal, by which time many clients had booked airtime somewhere else.

“Our job was to get that down to seven minutes – but we cut the time to four seconds. “We combined the data from all the sources and automated how it was transferred. People now have all the information they need to make a sale at their fingertips, in a form that exactly meets the needs of their customers. That’s real business intelligence, delivering real business benefits.”

The Dot Group used IBM’s new Ascential family of software, which is an ETL (extraction, transformation and load) engine. Conventional databases run on matching engines, comparing pieces of data and rejecting those that do not match. Ascential looks for correlations between data. As Apcar explains, it’s a way to glue data together to create useful information, which is the basis of true business intelligence. This kind of approach, Apcar believes, is vital in mergers and acquisitions, where there is inevitably a need to weld together two totally separate empires of data. The task is particularly daunting because each of the empires will have its own, often incompatible repositories of data about customers, suppliers, personnel and products, etc.

“For example, we enabled two high street banks to create a single master reference database involving millions of customers after their merger,” he says. “Ascential looks for similarities in data, so it has dramatically reduced the problem of duplication and errors that normally occur. The merged company now has one record for each customer, which is a much greater achievement than it might sound.”

The capacity to recognise and deal intelligently with similarities in data is a remarkable achievement and has far reaching applications. Whereas a conventional database will see “John Smith of Acacia Avenue” and “J Smith of Acacia Ave” as two different pieces of data, Ascential will recognise that they are probably the same person (and make the connection between multiple databases in which he appears).

The same principle applies to managing databases of products and components. If, say, a car brake calliper is used across several models in several sales territories, with different part numbers, Ascential can spot the connections and automatically create a master database.

For one multinational catering services company, it’s the solution to managing a supply chain of 155,000 products, where there are 300 varieties of baked beans alone. IBM Business Intelligence manager David Beeston believes this is a true revolution in data warehousing and management, because it cuts through the growing complexity and sheer volume of data, to create clarity and resolution.

Nowhere is this more desperately needed than in retailing: “Knowing your customer is the most vital aspect of intelligence in any business involving people. It’s also essential to know who you are dealing with and whether you can trust them,” he says. “For example, retailers typically suffer five per cent shrinkage from theft, fraud and errors, all of which are caused by people, who may be suppliers, employees or members of the public. There are often patterns in dishonest behaviour, which Master Data Management will recognise. In the same way, there are patterns in behaviour of honest customers, including their tastes, interests and purchasing decisions. Understanding them is the key to targeted marketing.

“I always ask businesses, ‘How important is it for you to see into the future?’,” adds Apcar. “The ultimate value of business intelligence is being able to understand both what’s happening now and what’s going to happen next. That’s where the real competitive edge lies.”

Printer Friendly Email This Article