Business Intelligence ROI: It's a Figment of Your Imagination

Business Intelligence ROI: It's a Figment of Your Imagination Vijay Sankaran September 4, 2001

Business Intelligence

The big themes in corporate IT projects these days are "business value" and "ROI." Everyone is trying to justify the business benefit behind their IT projects. That in itself is a good thing. Every project should determine the business benefits that it delivers, whether that is in operational savings, increased sales revenues or improved customer satisfaction. We've already run through an era where IT jockeys ran rampant and implemented solutions without consulting their business partners. Today, IT focuses much more on identifying business value in its projects but has gone completely overboard in its definition of business value. The focus is to quantify business benefits, and this is not an easy thing to do on an IT project.

The key metric that IT uses in its business quantification is return on investment (ROI). In justifying capital projects or new businesses, ROI is an invaluable justification. ROI is simply the revenues or savings that a project will deliver net of its costs. For example, if you were building a new manufacturing plant, the cost of that plant might be $1 million, but the capacity could increase by 1,000 units. The incremental profits over the next five years for those 1,000 units might be equivalent to $3 million in present dollars. Thus, the ROI for the project is:

profits - investment costs = $2 million

However, let's use a data warehouse project as an example of an IT project for which you want to build an ROI justification. Let's assume we want to build a customer data warehouse that will allow for targeted marketing by building analytical profiles of individual customers. The project will cost $10 million over three years. We start with our high-level business benefits: increased customer acquisition, increased customer wallet share, reduced mailing costs and a higher rate of retention. So far so good; you've defined valid benefits for your project. Now, the tough part: quantifying those business benefits:

• How many new customers can you acquire because of the data warehouse?
• How much more are those customers likely to spend because of the data warehouse?
• How much can I improve retention because of the new technology?
• How many mailings can I reduce?

How do you even come up with numbers for those first three questions? At least with the mailings, you can use data mining to do targeted mailings and reduce the total mailing population. However, reduced mailings are the least value-added proposition and would take years to recoup the costs of an expensive data warehouse. For the rest of the questions, let me know when you have an answer, and I'll publish your method on my site.

Let's just suppose for a second that you could come up with numbers for the three difficult questions. What's next? You could probably determine what the incremental profit was for each new customer based on an average profit per customer. You could calculate incremental profits based on increased wallet share, and you could probably determine the value if customers were retained for a longer period of time. This could be the additional profit generated by the warehouse. This part of the calculation is actually fairly logical. However, the question really is whether your business is willing to commit that it will acquire X new customers next year because of the data warehouse or whether it can commit to increase average customer revenue by $Y per person. I bet you that not one business executive would be willing to stick his or her head under that guillotine. If you know of any executives who have, please send me their name because I will post their name on my site as well.

So what am I trying to tell you? Well, what I am NOT trying to tell you is that business justification is a bad thing. Business justification is a GREAT thing for IT projects. However, quantifying specific ROI for IT projects and specifically business intelligence projects is an incredibly difficult task. Rather than predicting cost savings and increased revenue based on fantasy scenarios, use figures that other companies have captured for the specific project that you are delivering. For example, try to find out customer acquisition and customer retention improvements that other companies have realized and use that as a basis for your predictions. Second, don't hold a gun to anyone's head about committing to specific numerical benefits. If your organization is so stringent about committing your life to specific project-level ROI, no one will commit to anything. Just don't lose sight of those business benefits that make sense and have logic behind them and you will be much more successful than you ever have been.