
Previous implementations of business intelligence (BI) solutions, justified suspicion about new claims, the entrenchment of spreadsheets, and a lack of executive buy-in aren't helping BI's adoption.
Today's CEOs are harassed individuals. Demands from shareholders, business partners, staff and the board often make it an unenviable job. But what if they could get an instant bird's eye view of the state of their businesses on their computer desktop? That has been the promise of BI since 1989. The term, coined by a Gartner analyst, encompasses all decision-based information systems. As with many technologies, some parts of BI work brilliantly, some parts are probably an impossible dream to all but those secretly building them, and the rest would work just fine if the older versions already installed could be ripped out tomorrow and started from scratch.
Historic issues
Martin Rennhackkamp, technology director at Cape-based Prescient Business Technologies, agrees that BI suffers somewhat from its past. "We have seen a lot of clients who still suffer from their legacy and past mistakes, which they almost refuse to correct due to politics or business pressure, or a combination of both," he says.
We have seen a lot of clients who still suffer from their legacy and past mistakes.
Martin Rennhackkamp, technology director, Prescient Business Technologies
"There are some clean slate implementations which work very well, to be sure, or in some instances part of a BI solution does deliver business value, but there is definitely a legacy problem. It's a very difficult technology to rip-out and replace, because it gets locked in and absorbed so easily - baggage included.
"Despite major problems with the implementation, it delivers the chairman's report, so of course nobody wants to change it. I liken it to nursing a broken car across a gravel road with spit and bubblegum," he says.
Dirty data
One of modern BI's biggest problems is not its fault. All too often a successful pilot project will expose for the first time - and much to the horror of management - just how poor the underlying data really is. This unwelcome surprise often forces organisations to perform months, sometimes years, of work on their data before attempting to extract meaningful information from it.
There is absolutely no reason to abandon [Excel when] you have relied on [it] for so long.
Jeremy Waterman, MD, Softline Accpac
"Data cleansing is the expensive part of our consulting process," says Yolanda Komen, BI manager at Intellient. "Seventy percent of a BI project is spent cleaning up data. Twenty percent is the front-end and the remaining 10% is spent building the cubes and doing the reports and front-end scoping."
Andrew Brown, BI manager at CSC, confirms that any value BI analysis tools have is often sabotaged by the poor quality of data contained in corporate databases. "The most common example of dirty data on a large scale is only uncovered after CRM systems highlight how often clients have multiple addresses pulled in from different systems," he says. "And there's a good chance that none of them are current."
Bow to the master
Something that attempts to address this is master data management (MDM), a central store of customer and product hierarchies, business rules and business processes that is auditable and traceable. FirstRand's banking division has implemented Hyperion's MDM server solution from Intellient for precisely this reason.
Data cleansing is the expensive part of our consulting process.
Yolanda Komen, BI manager, Intellient
"MDM helps us cut the time our experts need to spend updating systems with changes in master data," says Tim Green, IT manager at FirstRand. "It also teaches people at ground level how to ensure data integrity through version control."
And it's often the people at ground level that are the next biggest hurdle. Once a company has clean data and what most BI experts refer to as a "single source of truth", the next most important factor is not the technology but rather the people and processes.
"The technology is there," says Prescient's Rennhackkamp, "but you have to have the commitment in the organisation that 'we're going to do this'. There are enough queuing and data warehousing and enterprise architecture components to make the technology work. The key is to have somebody at a high enough level to make it happen."
Soft issues
Marianne Vosloo, MD of Icentric, agrees. "BI solutions are not always successful due to bad implementation, but often it's about soft issues," she says. "The culture of the company is very important; some executives drive their organisation without realising the benefits that good information can give them. That information and knowledge must come from BI. A lot of people think they can solve problems by putting in the right software. But you have to train people to understand that data is important, that quality is important and to make them comfortable with using a new tool.
BI solutions are not always successful due to bad implementation.
Marianne Vosloo, MD, Icentric
"Often, people have been using Excel for many moons because that's what they're used to and it can be extremely difficult to move them onto something like Business Objects or Cognos because they are quite different," she notes.
Ampie Swanepoel, services manager of the corporate performance management unit at Dimension Data SA, is not a fan of Excel. "Many companies use Excel for budgeting purposes," he says, "but not only is this method riddled with possible errors, but it is also very time-consuming. In many instances, an organisation's strategic objectives are also not aligned to, or reflected in its budget. Using archaic budgeting methods does not allow the organisation to link its short-term objectives to long-term strategies."
Keep those spreadsheets
But the popular spreadsheet does have its defenders [see the round table discussion for more perspective]. Jeremy Waterman, MD of Softline Accpac, says the perception that companies need to throw out spreadsheets is a myth. "Since many of today's more affordable BI solutions integrate seamlessly with Excel, there is absolutely no reason to abandon a trusty tool that you have relied on for so long," he says. "As long as you can access the right data that can help you measure your company's operational performance, the spreadsheet is still the way to go. Of course, if you are accessing data from disparate databases whose information is out of sync, then it won't matter what BI solution you use; you're not going to get accurate results."
Disparate information - even if clean - still remains a challenge for successful BI.
Mind the gap
Rennhackkamp points out that the problem is often that a large gap exists between the BI consumers and the original data producers. "The staff who use BI are far removed structurally or otherwise from the staff that manage the operational systems. In some organisations, there's still an education gap at the top. Management needs to be educated on the importance of information, believe it or not. That's where we see the difference between organisations who are content to fly their Boeings by the stars and those who want to practise good airmanship."
Getting that gap reduced is the next challenge for the BI systems integrator, given that the technology is moving down from the boardroom and onto many more desktops.
Dr Willem Gertenbach, business development consultant at T-Systems SA, says as BI extends its reach to ordinary users, there will be a new focus on simplified and interactive real-time process-oriented tools.
"The CEO still needs a balanced scorecard; the sales executive needs an up-to-date sales pipeline," he points out. "But immediate action is required if RFID [radio frequency identification] systems detect that a truck is incorrectly loaded."
Gertenbach recommends CIOs ensure they understand the big picture. "BI is now being extended to the masses and this may require new skill sets, although it is still important for the data farmers, miners and analysts."
Clearly, access to operational back-end data is becoming even more important. Vendors that have a grip on both back-end software and BI software will have an undisputed competitive advantage.
* Article first published on brainstorm.itweb.co.za
Share