Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

March 8, 2002

/020308/505feat3_1.jhtml">

The Cost of Integrity

Don't let transactional system standards keep you from reaping BI benefits

By Ram Reddy

Continued from Page 1

These categories of BI systems are by no means mutually exclusive. In many firms, elements of the predictive may be meshed with the historical or analytic. The classification scheme helps managers organize priorities when tackling a new BI project. If the proposed BI project covers the entire waterfront, then the recommended path would be to break the project functionality into the three generic categories. Then sequence elements of the project based on business benefit and difficulty of implementation. As I discuss in the next section, a cost-benefit analysis is an absolute must before building BI systems.

REVENUE GENERATOR, NOT COST CENTER

There are other differences between BI and OLTP systems besides the data cleansing and integration issues. These differences stem from the method used to justify funding the project, implementation, and production support. Managers must be aware of these differences if they are to realize the business benefits of deploying BI systems.

Give BI technologies the same treatment as manufacturing investments. BI systems are generally used to make decisions that have the potential for profound financial consequences for the firm. The intelligence for decision-making needs to be delivered before its usefulness expires. Consider the retail example used in the previous section to illuminate historical BI. The purchasing manager for a major retail chain may want the prior year's profitability analysis for a category — such as big screen TVs. The manager wants to use this information to negotiate better rates from the TV manufacturers. If this intelligence has the potential to generate cost savings in a range of $3 million to $4 million, then that should be the sole yardstick used to make the BI investment decision, even if it costs $1.5 million to deliver the system on time with tools that aren't standard across the enterprise.

What's important here is that the intelligence is available at the time of negotiation. An elegant and complete system that's delivered months after the completion of the negotiating cycle doesn't add any business value. Wal-Mart took the approach of using tools (some discarded after system implementation) that delivered the BI system on time without destroying the ROI. BI development and implementation costs should be treated in the same manner as tangible property, plant, and equipment (PPE) investments. Tangible PPE investments are made strictly on the basis of cost-benefit analysis. The tangible asset is amortized and PPE replaced after its useful life. A similar approach needs to be taken for BI systems to deliver value.

Businesses too often burden BI systems with the concept of total cost of ownership (TCO), which really applies to OLTP systems exclusively. TCO is all about standardizing the hardware, software, and application development platforms to reduce the TCO. This approach assumes that IT is a cost center. In contrast, BI systems backed by a sound cost-benefit business case add to the bottom line.

Tools to build BI systems for a particular project should be selected for ease of use, so they can be deployed quickly without the need for extensive off-site training. A large number of BI initiatives get sidetracked by trying to pick the most integrated and technically elegant toolkits that adhere to corporate standards (which is a hangover from the TCO days). Given that BI's focus is on speed of delivery, cumbersome and integrated tools that require extensive training and IT support will be counterproductive. Acquire tools as needed and treat them as expendable after successful delivery of the project. This may sound like heresy, but we have to treat them no differently from PPE.

SUCCESS CAN KILL

A BI system that's successful has the potential to kill the supporting IT department. Successful BI systems, unlike OLTP (deploy and forget for the most part), require frequent enhancements (daily or weekly isn't uncommon). If the system is successful, it implies that frontline managers are using it to make decisions that are improving the firm's performance. This success will make them even hungrier for intelligence and ask for modifications, additions, and refinements that will help run the business better. Plan on assigning full-time resources to support a BI system — it will grow and evolve dynamically if successful.

LISTEN AND SET EXPECTATIONS

To successfully define, develop, and implement BI systems requires communication at a level that's unusual for the IT function. While you're designing the system, it's essential to communicate with the SMEs, who can define the data cleansing, integration, and business benefits accurately. These tangible business benefits subsequently drive the rest of the project.

A mental grasp of the potential business benefits of a BI system leads the business user to tolerate a partially functioning system that delivers value as opposed to an elegant solution that may be delayed. Repeated communication about the sanctioned uses of the BI system is essential. In the event of a successful experience using a BI system, managers may get lulled into using the BI system for looking up operational data that may not be accurate there.

Rate This Article

Comments:

Optional e-mail address:

Improper use of a BI system can derail what would otherwise be a successful implementation. Consider a marketing executive using the retail BI system discussed in the earlier example to inquire about total sales by store. That information is available and required for the purchasing manager to negotiate prices with TV manufacturers. The correct sales figures by store would be in the financial systems. The financial system would have accounted for chargebacks, warranties, and other modifications to generate accurate sales figures. Obviously, the figures from the BI system wouldn't correlate with the figures from the finance system. The marketing executive may be embarrassed in a public forum and blame the "stupid BI system" for giving inaccurate data. Very quickly word spreads that the BI system is unreliable. It happens more often than you think. Therefore, it's imperative to remind business users about the proper use of a BI system and, more important, what it shouldn't be used for.

STOP THINKING OLTP

Exasperated with the delayed implementation schedule of BI systems, the CIO of a major manufacturing firm complained, "By the time my IT group delivers an elegant, integrated, and reliable BI system adhering to the corporate standards, the business may cease to exist." This is the challenge facing IT today. We have demonstrated our ability to build OLTP systems that handle millions of transactions while reducing costs. Can we adapt our mindset in order to build BI systems that assist firms in running their businesses and add value to the top and bottom lines?

Ram Reddy [ramreddy@tacticagroup.com] is the author of Supply Chains to Virtual Integration (McGraw-Hill, 2001). He is the president of Tactica Consulting Group (www.tacticagroup.com), a technology and business strategy consulting company.