Scott Kiehl, head of business intelligence (BI) at gaming machine operator Inspired Gaming, explains: ‘Previously, a collector would pick up money from our machines weekly. We would wait to see when revenues for a particular location dropped off, then send over a van with a new machine.

‘Now our networked machines run on a common platform. We can pull data back from them on an almost transactional basis then send an invoice [to the venue hosting the machine] based on that information. This is also processed, allowing us to determine whether content is or isn’t working.’

If a machine rejects a certain number of coins per hour, for example, maintenance staff are automatically summoned to fix it. Similarly, if a game isn’t proving popular, a new one can be uploaded.

It’s hard to put a figure on what the advent of networked gaming machines has meant for Inspired Gaming, says Kiehl. In the seven years since the system was introduced, the company has expanded into seven countries other than the UK, and it listed on AIM last year with a market capitalisation of £115 million (now over £200 million). The international growth would, Kiehl suggests, have been ‘difficult, if not impossible’ without networked machines.

Inspired Gaming shows that BI can have a dramatic impact on a company’s bottom line and its scope for expansion. But before you embark on a BI project, you need to think about the quality of the data you’re starting with, according to David Hobbs-Mallyon, BI product manager at Microsoft.

‘You have to take a strategic decision somewhere along the line that BI is something of importance,’ he says. ‘When you make that decision, that’s when master data management gets on the agenda.’

Mastering your data

As much a management philosophy as a set of technologies, master data management (MDM) promises a seamless integration of data across the divisions or departments of a company.

That may sound more like a vision of IT nirvana than an achievable goal. Tony Jaskeran is the man charged with making it a reality in his role as head of BI at Allied Bakeries, owner of the Allinson, Kingsmill and Burgen brands. ‘We see information as the primary leverage in helping us achieve our business goals,’ he says.

Over the past century or so, Allied Bakeries has accumulated a vast trove of data. The problem is the treasure and the trash is spread across a large number of business systems: 131, to be precise. Of these, the majority were developed in-house.

‘Over the past decades, if you needed a particular process or application, you designed it yourself,’ Jaskeran explains. ‘So you develop more systems without really introducing classical processes.’

Smaller enterprises may not face issues of this complexity, but problems of systems that don’t work together, and the difficulty of getting an overview of the information you have, will strike a chord. Even MDM vendors agree that, for most companies, harmonising the various strands of information is a long way off.

Time to catch up

‘No company has actually finished an MDM implementation,’ says Tony Fisher, chief executive officer of software supplier DataFlux, adding that only ten to 15 per cent of businesses are even ready to start.

‘Once companies make the hard decision to change, we move in a very methodical way to help people bring in elements of the system one at a time,’ Fisher adds.

Fortunately, there is widespread agreement about the first step in this process. An investment in data quality is essential for MDM to be worthwhile – and it can pay dividends all by itself, according to Andrew Larter, data and development director of directory service 118-118.

The Number, which runs 118-118, took the decision to improve its data accuracy two years ago, focusing in particular on duplicate records.

‘Because we take hundreds of millions of calls, even the tiniest errors were unacceptable,’ says Larter. ‘So our data processing volume and error correction operates on a very large scale.’

Larter claims that by using software that provides a framework for data quality and targets duplicate records quickly, bad data in 118-118’s directory was reduced from five per cent to virtually none in six months. ‘We were able to totally eliminate the risk of bad data,’ he states.

Once quality is improved across a business’ various systems, it’s time to bring the systems together. But this doesn’t necessarily equate to centralisation, according to Jaskeran.

‘MDM is not about consolidating all your data into a single set,’ he explains. ‘In supply chains where you’re looking after products, there may not be an impact on the financial strategy, so there’d be no need to integrate those two areas.’

Points of agreement

Where there are ‘touchpoints’ between areas of the business, as Fisher calls them, it’s essential to get people to work together.

‘Cultural change is the biggest issue with any of this,’ he says. ‘People aren’t used to and don’t necessarily like working with other people. A manager might be very successful operating within the confines of what he does. Now he is being told: “You have to work with these other people here.” And that’s very hard to do.’

It’s an irony not lost on Fisher that many of the problems advocates of MDM are now trying to solve, were thrown up by the solutions his industry provided to previous problems.

‘Thirty years ago, business was all about handshakes,’ he recalls. ‘But over the past few decades, mom-and-pop shops have almost completely died out, and the intimate understanding they had of their business and their customers has been lost.’

Systems were developed to plug this familiarity gap, such as data warehousing, enterprise resource planning, and customer relationship management, Fisher adds. The problem was that the systems didn’t work together, and the information on them was left gathering microdust.

Fisher concludes: ‘In a sense, all this effort has been made just to get back to where we were 30 years ago.’