Why BI fails

SO WHAT IS GOING WRONG?

According to Van Aswegen, it could be due to mindset. "BI is becoming more entrenched in the corporate mindset in South Africa, although it still has a long way to go. The industry first needs to kill off the memories of the long, expensive, failed projects from years ago when, for example, data warehouses were the flavour of the day."

IT AND BUSINESS DISCONNECT

Caron Mooney, director of IS Partners, says that there is often a major disconnect between IT and business during the implementation stage, and this leads to major hiccups. She says often IT is too quick to say no and should start accepting its role as that of facilitator, instead of clipping the wings of top management. "IT often forgets to listen to top management properly," argues Mooney. "It needs to view these requirements from a business perspective and not a technical one. Once the requirements are realised, the company needs to figure out what is involved in making it available, the costing of it, and how to get it."

At the same time, she says, business should also realise that a job well done takes time. "Any IT project has to have serious commitment from the business side - not only in words, but in time."

GETTING TECHNICAL

According to Martin Rennhackkamp, strategic information director at the PBT Group, its failures could be due to integration issues. "ERP vendors have packaged warehouse solutions," he says. "It is sold to a company as a BI solution and often does not integrate at all into existing structure."

John Olsson, marketing director at Ability Solutions, stresses that often too much emphasis is placed in how data is presented, rather than the data itself.

"While there is always a lot of noise about the front-end information presentation tool that each vendor has in their BI suite, nobody really cares about these applications," he states. "The important aspect of BI is data. If you want to know if a BI solution is successful, simply ask how the data is stored, how it got there, what quality control is in place, how the data cube was designed and if the cube will satisfy the real intelligence needs of the company. How the data is presented is almost irrelevant."

According to Van Aswegen, problems crop up due to bad data management. "A recent Metadata study found that 80% of companies have no centralised data strategy in place," he says. "Any successful master data management programme requires three components: technology, processes and people."

Synergy MD Andrew Connold also feels the problem could very well lie in the way that data is handled, and stresses that the simpler the solution, the easier it will make the process. "People tend to over-engineer BI," he says. "They try to put in too many measures and invest in too much that isn`t necessary."

Intellient BI manager Yolanda Komen agrees, saying: "Companies should not try to measure everything. Instead, they should focus on the key performance indicators that allow them to proactively manage their business."