A leading Australian telco uses it to examine the likely impact on its business of customer churn. At the NSW Department of Community Services it helps HR proactively address staffing issues. Goodman Fielder has harnessed its power to help bridge islands of information and outline the "big picture" on sales, while at Austral Energy it simplifies tracking of repairs on power poles -- as vital from a safety and liability point of view as it is to cost savings. A pharmaceutical company near you is almost certainly using it to continually evaluate their vulnerability to competition: tracking patents and trawling through other open source information like financial reports, Web site information, press reports on market activity and postings from academia.
Business intelligence (BI) is undeniably hard to do well, but it's also one of the most potent tools available to executives today, with the capacity to generate substantial bottom-line results. Organisations that harness BI to anticipate external forces gain a razor-sharp competitive edge and may avoid those unfortunate decisions that can trip up even the smartest of companies.
According to Vernon Prior, vice president, Society of Competitive Intelligence Professionals, a recent study of 24 US defence and aerospace companies found that those three using business intelligence obtained outstanding results in several categories of performance. Against an industry-average bid success rate of 18 per cent, the top three won 87 per cent, 75 per cent and 57 per cent of bids respectively. Better yet, while the industry average return for every dollar spent on proposals was $US78, the top three averaged $US225. Those are substantial benefits.
BI has given organisations advanced warnings of competitive threats in a globalised market place and allowed years-long lead times on new technologies through monitoring the regulatory and political environment and the activities of potential partners and existing and potential competitors. It has application to almost every functional area of a business: from use of sales and marketing intelligence for help in preparation of tenders to the use of technical competitive intelligence for tracking technical advances. Typically its most important function lies in responding to questions from senior management to support their decision-making activities.
Retail electricity distributor Integral Energy's BERT project is all about implementing business intelligence tools from Cognos in preparation for a fully competitive electricity market in 2001. "BERT has made a huge difference in the way we use our data," BERT project manager Warren Thrush says. "We now spend more time analysing business data and less time generating the data. Marrying our information systems with our corporate vision makes for a smarter organisation. For example, in finance we've moved from generating reconciliations, to analysing trends in cost drivers."
Smart companies exploit BI to explore all the areas that fundamentally influence their strategic plans, from competitors and the business environment through to the market place. And they do it continually and cyclically. "BI is ideally a continuous process," says Woodlawn Marketing Services principal Adrian Farrell. "It is designed for efficiency and effectiveness in a world where decision makers are flooded with information but starved of intelligence. The intelligence aspect involves, through various methods of analysis, the conversion of multiple pieces of information into meaning. It is akin to solving a jigsaw puzzle."
Intelligence operations are hard to detect, difficult to prevent and perfectly legal -- most of the information you could ever need is available amongst your own data resources and from open source information. BI can even have a bonus effect of serving to alert the organisation to the dangers of making its own information too freely available to the world. Yet most companies have yet to fully exploit this potent weapon. One reason is that BI based on first generation host-based query and reporting or second generation data warehousing technology saw attempts to extract business benefit from data degenerating into over-budget, under-performing BI applications.
"A lot of companies have been doing that first part, which is the data warehousing, where they've been building enterprise data warehouses, or they've been trying to build enterprise data warehouses," IBM marketing manager, data management and business intelligence software Anil Reddy says. "Some companies have succeeded, some have failed"In recent times there's been a move towards building departmental data marts taking data from different sources, but feeding it into a data mart just for a particular part of the business. A lot of that was because people had problems trying to feed into that data warehousing -- you know, extracting data from the different sources." Now, Reddy says, explosive growth in the number of tools available to provide that back-end integration and extraction from operational systems and different data sources was making BI vastly more feasible.
At the same time CIOs have come to recognise the futility of building a data warehouse without a strong analytical platform to support it. Traditional business intelligence tools were structured to provide very specific kinds of query, analysis and reporting, but they were expensive and difficult to deploy across the enterprise. These days a new emphasis on information democracy is coinciding with the bringing of the new enterprise reporting tools to the masses. These tools are designed to help companies turn raw customer data obtained via the Internet and other sources into useful, strategic business knowledge.
The coincidence with the new explosion of interest in customer relationship management (CRM) is giving BI efforts a new impetus as companies look to transform transaction data, generally from enterprise resource planning systems (ERP), into information and knowledge. GartnerGroup describes business intelligence as a "hot new industry sector". Independent research company Ovum says demand for business intelligence products and services is expected to top $US50 billion worldwide within the first few years of the next millennium.
There's still no foolproof remedy, but experts say that the sector has evolved to the point where the risks of implementation are dropping, and the corporate benefits of BI are becoming far easier to obtain. Those barriers that remain are becoming easier to surmount, with forethought and careful planning. "I guess one of the biggest barriers that a lot of these companies have found is just setting up that common view across the organisation," Reddy says. "I think they've been finding it difficult to try and extract or pull that information out of their existing operational data systems and then putting that into a common view. There can be a lot of redundancy of data and duplication of that data."
Here extraction utilities like IBM's ETI (which allows extraction from multiple different sources of data) and data cleansing and data transformation tools can make a difference. "When you have six or seven different operational databases, you need to then pull the information out of those different databases. I think a lot of companies have tried to do this with hand-coding and it just takes too long to have a team of people trying to hand-code and extract this data, because it's constantly changing," Reddy says.
Sophisticated and Refined
Database giants such as IBM, Microsoft and Oracle have all refined their data warehousing wares, transforming their databases into platforms for business intelligence solutions rather than simple data repositories. Meanwhile, vendors like Cognos and Informix are offering sophisticated suites of BI tools designed as end-to-end data warehousing frameworks that provide everything from extraction to analysis.
Computron Software has launched Cool iMergence, an enterprise information portal designed to let users leave the existing reporting infrastructure in place and create a "report repository" to help navigate corporate information and extract new information. Computron claims that despite the relative maturity of the reporting, query and BI tools market -- and the proliferation of OLAP, ROLAP and other analytical tools --market research in Australia consistently identifies "flexible reporting and analysis" as the top issue with CFOs.
At Goodman Fielder's baking division the information required to get an understanding of the sales situation resides on three totally separate sales systems repositories: Oracle Express, Informix Relational Database and Cognos PowerPlay Cubes. This information needs to be consolidated into one report. IT manager Jim O'Toole says: "Cool iMergence will address the issue of 'bridging these islands of information' and will enable us to overcome our immediate problem without having to build a data warehouse and with a short implementation time of between four and six weeks."
GartnerGroup says the enterprise BI suites (EBIS) segment has been forming through a natural evolution of query, reporting and OLAP viewing functionality driven by IS departments seeking to provide basic or common BI on a large scale.
"EBIS' predecessor, the 'low-end' BI segment, has been the most valuable of all BI segments, with a market value of around $800 million for 1998. As EBIS takes centre stage, we expect that it will drain most of the revenue from discrete query and reporting tools. A source of significant market energy, large software vendors will increasingly invest in this segment," GartnerGroup analyst Philip Gill predicts.
Such tools will undoubtedly help to transform data warehouses into useful BI repositories. But when it comes to the wealth of "soft" information available to the business from open sources, which can provide the real gems for intelligence purposes, the most basic BI tool remains the contact manager software used to keep track of sources of information, including industry contacts.
"Many companies go to considerable effort and expense to build massive databases -- designed either to cover everything they will ever need, or as a tool for database mining. The most useful, however, is one that gives access to contacts, experts, networks, existing and potential customers or clients, and so on; to sources of soft information, in other words," Prior says.
Data "Mine" Field
Of course, the business intelligence market is littered with failed multimillion-dollar projects; even today, implementing BI solutions remains a hazardous proposition. There are numbers of factors to consider if you want to get your BI implementation right.
According to Ovum, many users underestimate the importance of the "back-end problem -- getting information into the data warehouse or analytical application". The report adds: "Some vendors are still encouraging them in this attitude." There is also a risk of the data warehousing approach generating disparate data pools providing different business information to different users.
Reddy says organisations are particularly likely to run into problems, if they try to do data mining or decision-support off operational data. "You can then have the situation where you have a board meeting and the sales manager comes in and says: 'I've made my numbers', and the finance manager says: 'No you haven't', because they're looking at two different views of that same data.
"That happens all the time in a lot of larger companies," he says.
A successful BI implementation also requires companies to understand the potential of the technology, and it demands line-of-business buy-in, Reddy says. Further, many executives overwhelmed by the weight of information crossing their desks may find little time to consider intelligence reports. Many employees may be reluctant to share the intelligence they have.
In fact, corporate culture can often prove another significant barrier to implementation, says Cognos marketing director Peter Kokinakos. "One of the barriers is the culture of some of the organisations out there that want to retain that knowledge and that power. But if they can understand that the leveraging of information at all levels in an organisation is really essential to managing a business, there's now a way around that. You arm everyone with the right information, so that everyone understands the impact of all the different activities on the bottom line, then everyone can help to achieve improvements in the business. It's all about trying to get everyone batting for the team," he says.
It's equally vital, according to Farrell, to define a direction or framework of BI. With so much data available, it's essential to put some boundaries on the BI activity in order to focus on the most important areas of interest.
"The real point in competitive intelligence is that it's purely there to support the decision-making process, such that executives are making wise decisions based not just on how it used to work when they were in the business five years ago," he says. "What we're trying to do here is to complement that valuable experience with up-to-date relevant information so there is a disciplined approach to gathering and answering the questions that a senior executive will have."
Always start by clarifying the information the executive wants and the reasons he needs it, Farrell advises. At Integral Energy, Thrush has found it important to have people involved in the project who come from the business yet who also understand the "information objects" and how they relate to each other. "That's an important head start in constructing things that other mere mortals will use, because that acts like a useful transformation layer," he says.
Inchcape Motors has learnt to its cost that user involvement is also critical. Business analyst Andrew Bateup told CIO a previous small-scale attempt at BI where users were little involved led to a lack of ownership and some misunderstandings about definitions of concepts. "That was almost a pilot and we learned from that; now we're coming back with a very structured approach with a lot of consultation with users, defining all the issues upfront. We now have a high level of user involvement and a bit of ownership; BI has been embraced more by the users that were involved with that methodology, as opposed to the previous one where it was: here you go and good luck," he says.
One final golden rule is worth inscribing on your desk. Never forget that published information (including that available on the Internet) is not always the only source of intelligence, nor is published information always accurate. Check your sources, then check then again, to avoid leading your decision-makers astray.
Getting It All Together
A survey conducted by The Futures Group in October 1997 asked respondents to identify where intelligence is needed to make decisions. These areas were ranked in order of importance:
1. Competitive activities
2. Changing market or industry structure 3. Customer or supplier activities 4. Emerging technology initiatives 5. Global economic conditions 6. Regulatory climate 7. Political climate Leading on from there, Adrian Farrell, principal of Melbourne-based consultancy Woodlawn Marketing Services (who prefers the term competitive intelligence to business intelligence), identifies five steps to a successful BI project.
1. Define the direction or framework.
"Because there is so much data, it is vital that some boundaries are put on the CI activity, so as to focus on the most important areas of interest. For example, "Find the answer to this problem" or "What can we do to take advantage of this opportunity?"
CI practitioners rely on publications, suppliers and customers as the most popular sources of information, followed by company employees, industry experts, the Internet, industry conferences, and commercial databases.
Discovering information about what's happening in a market is not new. All businesses do this to some degree, Farrell says. However, particularly in small-to-medium sized businesses, this is usually an ad hoc and disorganised activity. It might occur after a company has lost a bid to a rival company. Or it might occur when there is a prospect of buying out another business and due diligence on information from external sources is required.
"A key word in this definition of CI is process, meaning this is a structured way of going about gaining competitive intelligence."
Analysis is the means of making sense of the myriad data available. It involves evaluating the data for usability by taking into account the relevance, truth value, understandability, sufficiency, significance and timeliness of the data. Information is then collated and synthesised according to the target and priorities set by users.
The filtered information is stored and disseminated to the relevant interested parties as written reports, via e-mails, or verbally. Online facilities are often provided for users to access stored intelligence reports.
Farrell says this important element is often missing from the intelligence cycle. Learning takes place through reflection on what occurred and by identifying ways to improve the process next time. For example, did the result match the expectations of the initiator? In what ways could the results be presented to improve communications? Debriefing is the way to ensure quality in the process.
"Although computers take a major role in CI systems, there is a large component of human involvement in all phases of the CI process," Farrell says.
-- S Bushell
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.