Business intelligence, a slightly tainted buzz phrase from the early 1990s, is supposedly making a big comeback — although some say it never went away and others insist that it never really arrived. Cynics of course will tell you that business intelligence is an oxymoron.
We’re not talking about
the alleged content of an MBA’s skull, though. Business intelligence (BI) refers to a set of systems tools for gathering, storing and analysing data from multiple databases and presenting intelligence reports based on it. BI helps managers make better decisions and companies gain competitive advantages.
Insurance firms and tax departments use it to uncover fraud. Wal-Mart uses BI to optimize inventory and shelf management to ensure the right products get on its shelves at the right time. Business intelligence is also a back-end enabler for customer relationship management (CRM) and supply chain systems.
It may not always be easy to calculate ROI on BI projects, advocates admit. “”But from the moment your competitor starts using business intelligence,”” says Bernard Sevigny, Toronto-based vice president of business intelligence at Fujitsu Consulting, “”if you want to stay competitive, if you want to survive, you will have to use this technology too.””
Calgary-based Trimac Transportation Services Inc. didn’t wait for its competitors. Trimac, one of the largest bulk shippers in North America, has for the past two years been using BI tools from Ottawa-based Cognos Inc. to help improve operations processes. It claims to be realizing business benefits of $2 million a year.
With costs, notably fuel and insurance, rising rapidly, competition intense and margins in the trucking industry “”razor thin,”” Trimac had no choice, says CIO Ted Barnicoat. “”We concluded that the only way we could prosper in this business was if we were very, very efficient. And that requires a tremendous amount of analysis of data.””
Analysis of data is at the heart of BI. Exception reporting and trend and pattern analysis performed by analytical tools such as Cognos’s PowerPlay turns raw data into business intelligence.
At Trimac, the Cognos-powered Trip Standards application compares data from the company’s billing and dispatch systems and kicks out exception reports when parameters used to calculate customer contracts — such as time to load and unload, miles driven, etc. — differ from actual performance and eventual billing.
For example, Trip Standards analysis revealed that Trimac trucks moving wood chips from saw mills to pulp mills were consistently hauling loads that were under the contracted weight because drivers had no way to judge volume at the point of loading. Result: Trimac had to bill less per load than it expected — or had budgeted.
This is valuable intelligence, but as Sevigny points out, “”BI supports decision making, it never provides you with a decision. Human managers have to make the proper decisions based on that intelligence.””
Analytics may be the heart of BI, but the user-facing components — typically Web portals and PC-screen dashboards — are vital. Some advanced BI tools even generate actions such as e-mails alerting users to changed circumstances — a shortfall in sales in a region or product market, for example.
At Trimac, the Trip Standards reports go to operations managers at the companies 142 locations. They come up with ways to solve the problems. In the wood chip case, the solution was to install relatively inexpensive air scales on the trailers, which allowed drivers to more accurately load them.
“”In some cases, we’re only talking about savings of a couple of bucks a load,”” says Barnicoat. “”But if you add it up over this scale — 3,000 trucks, 4,000 drivers, 142 branches — you start to get some important benefits.””
Doing the analysis and — most important — ensuring decision makers see, understand and act on it is what delivers BI benefits. But before that can happen, BI systems have to integrate, standardize and cleanse data from multiple sources and create logical structures — data warehouses or data marts — for storing unified data and reports.
As Barnicoat says with some understatement, “”The process isn’t without pain. Getting the data populated and standardized, the discipline needed — that requires a lot of work.””
It’s also where BI-related projects most often flounder. In the past, too many enterprises sank all their resources into enterprise data warehouses, which proved difficult and expensive to build. Worse, in the absence of good analytics and BI reporting, they delivered few tangible benefits. Data warehouses too often ended up looking like white elephants, and BI suffered guilt by association.
That’s changing now, though, says consultant Lian Zerafa, a senior manager at Deloitte Consulting Inc. in Toronto. “”Now the BI tools have been decoupled [from data warehousing] and are no longer carrying all that baggage,”” Zerafa says.
He and others also advocate a “”federated”” approach to business intelligence. Rather than build huge enterprise-wide data warehouses, it’s better to deploy BI at the departmental level first — as Trimac did. Get some immediate pay-backs, then integrate later. This has given BI a boost.
“”There has been a huge resurgence of business intelligence recently partly because systems are more practical and deployable now,”” Zerafa says.
They’re also in many cases cheaper. The price tag for customized solutions starts in the low hundreds of thousands and climbs into the millions or even tens of millions. But vendors are now introducing cheaper, easier-to-implement BI-in-a-box products. They provide data marts, analytics and standardized reporting for well-understood areas such as sales and customer relationship management. And this for the first time is making BI accessible to mid-size firms.
The new products bring the price well below $200,000 — in some cases much lower. For example QlikTech Inc. of Raleigh N.C. recently announced its QlikView Sales Performance Analysis application — at $39,900 U.S.
Sevigny would argue that business intelligence didn’t need a resurgence because it never really went away and, indeed, is already mainstream. It’s just that different, related, buzz words — CRM, supply chain management — grabbed the IT spotlight.
“”There is no true CRM,”” Sevigny points out, “”without a business intelligence infrastructure in the backend. You need to gather and analyse all that information about the customer first.””
In fact, growth in the BI tools market did slow over 2000 and 2001. But according to a June 2001 report from Boston-based market research firm IDC, it was set to accelerate again.
The worldwide market in 2000 was $3.6 billion, the report said. IDC expected it to grow to $11.9 billion in 2005, a compound annual growth rate (CAGR) of 27 per cent. That was probably over-optimistic in light of the post-9/11 recession. A more recent IDC report on the Western European market called for only a 14.9 per cent CAGR between 2001 and 2006.
Still, however you define it, business intelligence appears set to play a larger role in more companies in the future. That doesn’t mean deploying the technology is now a cake walk. No gain without pain, the fitness gurus used to say, and it fits here. (See sidebar this page for tips on minimizing the pain.) But with diligence and a little BI of the human variety, business intelligence can clearly deliver real benefits.
1. Manage expectations. Vendors are aggressive in their claims. BI in fact can’t do everything — or anything without considerable effort.
2. Involve decision makers early and often. You can’t build effective BI solutions if you don’t understand how decisions are made.
3. Develop to design. Users will find it difficult to describe how they make decisions. Build prototypes: system requirements will be defined by user reactions to them.
4. Ensure data quality. It’s vitally important. Cleanse existing data and learn how to modify business processes to avoid erroneous data in future.
5. Be sure you understand what the data means. Implement meta data management functionality to help interpret data. Understanding the answer is more important that getting it.
Thanks to Bernard Sevigny, vice-president of business intelligence at Fujitsu Consulting in Toronto.