Reaping business intelligence's harvest

For investments in business-intelligence technology to be worthwhile, they must avoid creating silos

If you've been working in business for any length of time, you would have heard countless war stories. There was, for example, the one about the notebook manufacturer whose salespeople won a contract for 600,000 systems only to find its parts department couldn't source more than 300,000 notebook keyboards to make them.

Or the sales manager who set himself up for a bonus by setting conservative sales targets that were easily beaten -- until it came time to deliver, when manufacturing was struggling to meet real sales because the production manager had set her strategy based on the low sales forecast. And what about the toy manufacturer that forgot to consult its logistics provider when ramping up production, ending up with loads of unshippable orders that kept shelves empty two weeks before Christmas.

These sorts of stories reflect the type of all-too-common problems that can result when different parts of a company simply can't communicate properly. This was almost understandable years ago, when business managers used to plan strategies based on gut instinct and information that was often months out of date. Yet with the wealth of information available to today's managers, the problems have simply morphed into a different form as managers have come to rely on information systems that often reflect only part of the whole truth -- and, in so doing, drive poor business decision making.

Business intelligence (BI) software was, from its inception, designed to fix that situation by allowing first executives, and eventually large numbers of middle managers, to analyse data pertaining to activities that were not within their immediate control. This allowed the correlation of figures in a way that recognises the implicit interconnectedness of the operating divisions of modern business: nobody operates in a vacuum, and BI reflects that fact.

Sort of. The importance of good business reporting has helped the BI segment enjoy strong demand regardless of the varying fortunes of IT on the whole, yet most of the estimated 80 percent of businesses that have implemented BI systems have done so in what analysts like to call a "stovepipe" or "silo" manner.

You'll probably call it something else, though, after you realise the system in which your department has invested months of time -- and tens of thousands of dollars -- isn't giving you a complete picture of what's happening in the parts of the business outside your immediate control.

BI = Business Imperative
Problems with the distributed nature of business intelligence systems are not new, but finding a workable and universal solution has been much harder than simply identifying the problem. Data is inherently tied with the operational units where it's stored; even though the rise of enterprise resource planning has centralised business data, individual parts of the business still find many different ways to use that information.

Even more fun for the IT strategist, those ways often involve different tools that reflect each tool vendor's individual strengths. Cognos, for example, has long been strong in ad hoc data query processing, while SAS Institute's business performance focus has made its systems more of an applied business tool. Hyperion's roots in financial analysis are reflected in the structure of its current tools lineup, although its recent acquisition of query-based BI vendor Brio has rounded out Hyperion's capabilities.

In many cases, the tools' respective strengths provide systems that are most appropriate for the needs of a particular department. Yet multiplicity is the enemy of consistency, and the large number of BI platforms installed in the average company impedes efforts to standardise the handling and management of information across various parts of the business.

It's not just the information that's problematic, points out Russell Evans, sales and marketing director of Hyperion Australia. "Most companies have embraced BI and see a value from it, but it tends to be rear-window looking and [falls into] silos of power and requirement," he says.

"There's pressure on CFOs to greatly improve the clarity of their financial forecasts, but ERP is the domain of the CIO, and CFOs need to convince IT to give them the tools to provide greater levels of functionality. Unless chief executives and boards have really got their heads around behaviour from both a people and a systems perspective, getting past these silos presents a very big challenge for most organisations."

Although the problem is clear, the solution is not. Gartner recently predicted that this issue would persist within Global 2000 enterprises, which will continue to adopt "disparate and unrelated" BI systems well through 2005. By that point, the analyst firm warned, lack of integrated information could spell disaster as increasingly information-hungry executives find themselves unable to maintain the real-time, top-down perspective they so desperately need.

Timely reporting, after all, is the fuel for modern business -- particularly in public companies catering to risk-averse investors -- and it is often the companies with the best visibility of their performance that are able to maintain the most trust from investment markets and stakeholders. This trust cascades through the supply chain to suppliers and agents, and its dissolution can spell disaster.

Building BI competence
Clearly, the ideal of a central data repository provides the most promise of meeting the demands of enterprise reporting by storing one version of every piece of information in a single accessible place. In recent years, delivering on this vision became physically possible through storage area networks (SANs) that consolidated data from multiple servers; ERP systems that logically centralised key business information; and data warehouses, which provided data solidarity by organising key data into a single database that could be readily analysed.

None of these technologies has succeeded in overcoming internal business multiplicity, however. To this end, BI advocates are now pushing the consolidation of business intelligence assets into what Gartner calls a BICC (Business Intelligence Competency Centre).

In Gartner's explanation, the BICC becomes a centre of excellence for all things BI-related -- a focal point for advocacy of greater business intelligence. Rather than allowing individual departments to pursue their own strategies as they feel like it, the formation of a BICC represents executives' recognition that a more rigorous and stringent approach to BI has become necessary.

It provides both technological guidance and something of a moral high-ground from which BI advocates can force an analytics-driven way of working down through the entire company. In other words, the fight is for information efficiency, and effective BI should be the executive's weapon of choice.

But how to build a BICC? Business intelligence systems, after all, have long been the responsibility of IT departments that -- ideally -- work with individual departments to build systems that suit their needs. Even the fact that BI has been handled this way reflects the inappropriate structure of most businesses.

Maintenance teams, after all, are given responsibility for changing locks or fixing air conditioners; actuaries are called upon to formally quantify risk exposure; and production managers are charged with making sure supplies get where and when they're needed. So why should BI, which is arguably important as a source of business information in the same way that manufacturing drives business sales, be treated any differently?

Don't be afraid to put some effort into it. A proper BICC should be staffed by technology experts skilled in both the whys and wherefores of specific BI technologies, and in the over-reaching business needs that drive adoption of those technologies. Train your technology people in business understanding and analytical skills; create BI policies that route authority and responsibility for analytics projects through the BICC department; and give the BICC the authority to push the use of BI tools as a means of improving interaction between operationally separate business units.

This approach will give both policy backing and moral weight to the BICC, which will be in a better position to centralise data management and track the technological discrepancies between departments. BICC, BIN, BAM: the new BI ecosystem
In the long term, the centralisation of business intelligence expertise will pave the way towards what is collectively being referred to as the business intelligence network (BIN).

In the same way that the SAN centralised data storage and local area networks (LANs) centralise physical interconnects, the BIN centralises analytical knowledge and technology at a single point -- then uses ever-present relationships and technological links to spread that knowledge into different parts of the business. With the BICC as mediator, however, business analytics can be rolled out in a much more useful way.

A major part of the BIN concept is the use of access control and collaboration tools, which are being built into major BI systems as quickly as programmers can write code. Collaboration platforms can, for example, use instant messaging to tie business information and expertise to specific people, then indicate to users whenever they could reach the people responsible for that information. They also allow for the customisation of content to suit particular business needs whilst retaining a uniform BI infrastructure at the core of the company.

While customised analytical platforms may be the conduit for information exchange throughout the company, people are also a critical part of the business intelligence network. Their expertise is invaluable, and their ability to add value to financial, production, or other reports far exceeds the abilities of BI tools. An effective BIN, therefore, will need to address both technological and operational issues at a people level.

One company addressing this issue is New Zealand telecommunications company TelstraClear, with 350,000 customers taking a full range of residential and business services, which has used this approach to maintain service levels despite its heavy investment in monitoring and management technology that's directly tied to business performance.

TelstraClear used Micromuse NetCool to improve visibility of its call handling and problem escalation system, which is essential for the company to meet the service level agreements (SLAs) that it promises its customers. Ability to meet those SLAs is directly related to the company's bottom line, a fact that led the company to emphasise the people element even as it improved visibility of its network and operational performance.

"Correlating information about problems with the topology of our services helps us prioritise things, but that has to be done in near real time," says William Lee, head of network operations with the company. "This real-time information has to be presented to an operator, technician, or manager for them to be able to take either proactive or very fast reactive action to fix the problem. You get the base level of information from the system, but you still need smart people to react. At the end of the day, you still rely on good people and technicians."

Application of business intelligence analytics to TelstraClear's issues handling reveals patterns in performance and lets managers highlight areas for potential future investments. This information is then fed back into the company's analytical processes, driving decisions about funding and budgeting as well as future hiring, training of technical experts and other decisions. In this way, TelstraClear's BIN produces meaningful information about the company's operations and ties in the expertise of people into a continuous feedback loop aimed at continuous service improvement.

BIN doesn't operate in isolation, however: as if one three-letter acronym wasn't enough, many in the industry also point out the importance of business activity monitoring (BAM), which is a composite of BI, network systems management, and application integration and middleware.

If BIN provides the sheer data analytics and human decision-making to channel information processing, BAM embodies the business layer of the decision-making process. Gartner recognises 10 key activities that fall under the scope of BAM:

  • spelling out business processes;
  • achieving consensus on data and process standards;
  • identifying key parameters and establishing historical trend data;
  • a community-based approach to user interface design;
  • a high-payback, incremental approach;
  • change management and technology adoption;
  • ongoing measurement of business results;
  • integration of the process model within a BI framework;
  • optimising the BAM solution;
  • and
  • data and application integration.

  • Only three of these areas are directly driven by BI technology; the rest are business decisions, or reflect the application of BI to business activities. Gartner describes BAM's role as "to reduce or eliminate delays, bottlenecks and inefficient use of labour and materials, while providing real-time financial and performance data. BAM will be an analytical tool that enterprises must wield to gain a near 'zero latency' position -- that is, to reduce the time between capturing new data in one area and making it available and usable somewhere else."

    In other words, BAM is the life force of the business intelligence network. Inventory levels, for example, will be available in real time outside of the warehouse alongside orders and sales data; metrics like customer satisfaction, cash on hand, online experience, and on-time delivery will be tied to real financial results; and so on.

    By straddling the business and technology areas, BAM becomes a catalyst for turning BI analytics into real business process change. The BICC is the heart of the system, BIN the arteries, BAM the muscles -- and people, the brain.

    BI's big future
    After a quite ordinary 2002, the BI market heated up a bit during 2003, as vendors alternatively acquired and built new technologies pushing their technologies into new areas like collaboration and complex analytics. Rampant consolidation has narrowed the field considerably: Gartner's August 2003 Magic Quadrant industry snapshot identified just four leaders in the BI suite market -- Cognos, Business Objects, Information Builders, Crystal Decisions (since purchased by Business Objects) -- and only seven other companies in various stages of market maturity.

    This consolidation makes choosing a BI platform a much easier proposition for customers, who will find today's offerings richer and more capable than those that came before. Given the breadth of functionality in today's BI suites, it should be easier than ever for companies to tie their reporting, query and analytical capabilities into the broader enterprise context of the business intelligence network.

    IDC is forecasting continuing strong growth for the BI sector, predicting that Asia-Pacific BI revenues will grow at 12.2 percent annually from $588.1m (£328.7m) in 2002, to over $1bn by 2007. This reflects continuing strong adoption of BI solutions, with many companies strengthening their investments after enjoying success with first-generation deployments.

    Ultimately, the BIN will be so well entrenched within company strategy that it becomes virtually invisible to users. Some vendors have already pushed in that direction: Alphablox, for example, has enjoyed strong success with one Australian telecommunications company where its eponymous BI analytics framework has taken on a life of its own -- spreading across 18 applications and thousands of users -- as it progressively proves its value.

    The key to that growth has been Alphablox's strategy of providing a BI framework that's buried within internal business applications, says Elisabeth Whitelock, the company's sales and marketing director.

    "Our approach is to personalise it to the individual level so the experience they have is very specific to them, their role and the level of information they're entitled to access," she says. "It's going to be really difficult for people to say that they should only use one thing enterprise-wide, and I'm not necessarily convinced that one size fits all. But customers have a bigger view of where they want to take things, and a vision that this sort of technology can fulfil many [needs] across the organisation."

    By building the BICC as a business abstraction layer and the BIN as a BI technological abstraction layer, companies will be able to progressively increase their use of business information of all kinds. By 2006, Gartner believes the stratification of BI architectures will allow users in different departments to work on the same models simultaneously, or to have multiple BI applications communicating seamlessly.

    Whether the concept is right for your business depends, well, on your business. But as increasing numbers of corporate information users come to see BI as a tool for analysing information -- rather than an oracle for simply being given the answers -- there is no limit to the change that good intelligence can bring.

    Executive summary
    Business intelligence speculators envision BI technology that's going to become a fundamental enterprise service, like storage and network connectivity. Here are a few pointers to help you keep your BI strategy on track:

  • Executives reap what they sow… So they should plant their seeds well. Their day-to-day management is going to depend on the effective delivery of information, so it's important to take a keen interest in the company's BI strategy -- and not just let departments pursue their own strategies willy-nilly. Executive mandate can work wonders for data consistency.
  • . . . but silos are for farms. What use is data if the people that need it can't get to it? Departments often operate in isolation, but they need to approach business intelligence in a way that will bridge those gaps.
  • People are a great source of business intelligence. Make sure you don't depend on BI so much as provide it to the right people. Although BI networks will become invaluable to staying on top of the business, always remember that these systems still need to be designed for usability and relevance.
  • Data handling policies need to mirror business requirements. Just as BI is becoming a fundamental operational unit at the boundary between business and IT, it can also be a conduit for shaping the flow of data according to business rules. Utilise the BI network to enforce broader reporting and analytical requirements on the handling of data.
  • Nobody does better with too much information. Enterprise portals, and many BI platforms, allow customisation of system outputs based on their job role. Use this feature extensively as it will allow even casual users to take advantage of the BI network without drowning in information.
  • Collaborate. Just as it's assumed that everybody is reachable on email, instant messaging and collaboration systems are becoming increasingly important. Tomorrow's BI platforms will recognise this, tying in with knowledge management and sharing policies to encourage group action on key business information.
  • Plan. Don't get caught up in the technology; remember that BI is all about execution. If you can't translate its outputs into effective business action plans, your investment has been wasted.
  • Any BI is better than none. Because reporting is so essential to corporate functioning, BI tools are a great investment because they can provide such valuable information to so many people. As with many technologies, start small -- targeting low-hanging fruit -- and grow from there to see the most benefits.