Consulting » Software Decisions – The art of smart.

Business intelligence consists of a wide range of applications aimed at different functions. Its mission is to provide “access to, and analysis of, quantitative information sources to deliver insight as a means of aligning people and processes with an organisation’s mission”, says Frank Buytendijk, research VP at analysts Gartner. “It can include a data warehouse and involves reporting, query and analysis. “It is management information systems for the 21st century.”

But Nigel Pendse, lead author of The OLAP Report, an independent research resource for organisations buying and implementing financial applications, believes that business intelligence has become just another sexy marketing term applied to a wide range of solutions for the management of information. “It is an ill-defined category and each vendor defines it based on the modules it has built or acquired,” he says.

Although business intelligence can include procurement, manufacturing and human resources, the two main groups of users are marketing and finance. Marketing applications primarily use individual transaction-level detail, often in a data warehouse. Typical marketing applications include data mining for customer segmentation, targeted campaign management, customer profitability analysis and web analytics.

In contrast, finance applications within business intelligence use summarised data in a multidimensional online analytical processing (OLAP) database. It can be fed either from transaction systems, especially the accounting system, or from a data warehouse. All users share the same data and the database gives the organisation a single version of the truth.

Applications include budgeting, forecasting, reporting, consolidation, activity-based costing and management, scorecards, dashboards, portals, modelling, analysis, and tax and treasury planning. Financial consolidation systems often sit uneasily with the rest and some use a relational database. These applications all correspond to management processes with which they must be integrated.

Thirty years ago, mainframe-based decision support systems replaced paper in large firms. Then spreadsheets became widely used, but they were cumbersome, undocumented and prone to error. Organisations replacing them had a choice of either implementing a series of best-of-breed packaged software applications or building a bespoke system.

The latest trend is producing a single integrated software package that provides the widest possible range of functionality from a single interface and a single multidimensional database. Some organisations have started from scratch, but most have assembled their solution through a series of acquisitions. The resulting suites are variously described as business performance management (BPM), enterprise performance management or corporate performance management.

Business intelligence for the finance function can be thought of as BPM for the planning and control cycle. A recent survey of UK finance directors carried out by FT Research on behalf of Cartesis found that the main drivers of BPM projects are: to guide day-to-day decision-making (84%); to develop a long-term strategy (73%); and to control the organisation (69%).

Most FDs will have been living with the whole subject of turning data into information for years, starting with mainframe-based decision support systems in the 1970s, and then spreadsheets. Then, in the early 1980s, executive information systems (EIS) brought interactive screens with graphs, slicing and dicing, traffic light reports and drill-down. This basic functionality has persisted into OLAP, digital dashboards and portals.

Although business intelligence functionality has been around for 30 years, research by Accenture reveals that 91% of Fortune 1,000 companies believe they need to improve their business intelligence capabilities in order to manage economic and political uncertainty effectively and drive growth.

Companies of this size will have had a range of individual BPM applications for many years, so it is clear that their needs are still not fully met. Indeed, the Cartesis survey found that 88% of respondents are unsatisfied with their current BPM implementation and are still developing it. The fragmented origins of the BPM market are evident in the wide range of tools the respondents used, such as spreadsheets (95%), bespoke systems (57%), suite of products with multiple databases (44%), enterprise resource planning (ERP) (40%) and fully integrated solutions with one database (30%).

“Only recently have best-of-breed and ERP players gone after the enterprise performance management space holistically, looking at how to take the strategies and turn them into forecasts,” says Dan London, global partner for finance and performance management at Accenture. “Many of the solutions addressing this space are only just becoming mature enough to deliver on their promise.”

Even then, it isn’t enough for the software itself to be integrated. Management processes must be integrated with each other and with the other functions to produce an organisation-wide business intelligence process. “The difference between EIS and the scorecards and dashboards of today is that the cost of failure has never been lower,” say Buytendijk. “The problem has always been that they are not aligned with the other streams of management reporting so they introduce yet another version of the truth. The whole exercise is useless.”

A related problem is selecting key performance indicators (KPIs). Robert Bittlestone, managing director of Metapraxis, an EIS pioneer, points out that the finance fraternity still regards non-financial KPIs with some suspicion. “They have no real framework for managing them, are not part of the double-entry system and are not formally recorded in their lower level systems,” he says.

Nigel Youell, applications marketing manager at Hyperion, points out that the philosophy of EIS was to give a handful of people at the top a clear view of the business so they could steer it in the right direction. “The problem was that while the KPIs related to the people at the top, they didn’t relate to the people at the bottom,” he explains. “The guys at the sharp end, in sales or on the shop floor, were not aligned to go in the same direction and didn’t know how to contribute. It was not a good way to drive the business forward.”

Whereas numbers and ratios are naturally hierarchical, making it easy for software to allow the user to drill down to lower levels of summary, leading KPIs are not. Because different roles within the business impact different resources, the KPI at each level must be translated at the level below into one or more different contributory drivers that can be controlled at that level.

The KPIs at group level will differ from division and from operating level. Different functions and levels of management in the same business unit will also have different KPIs. All of these need to be aligned if they are to be successful.

“Balanced scorecards and KPIs alone won’t explain if the figures and percentages reported are drivers of the business or mere output measures,” says John Taylor, UK managing director of Cartesis. It is particularly important to align the budgeting process with all the other management processes, including incentive and reward schemes.

London says that clients who have taken an integrated approach to planning and reporting against the true levers of strategy report that their management review meetings require much less preparation. “The time they spend in the meeting is less focused on the past. They look at the key drivers the team can effect to drive the business forward and improve future performance.”

Developing an integrated set of KPIs has always been an enormous challenge and implementing interactive reporting systems has a long history of problems. “We are still an immature industry with immature terminology,” says Taylor. “Sometimes there is a sad waste of effort as business and IT people argue semantics. Different terminology is masking the simple truths that we could all probably agree on if we could just understand each other.”