6.4. BI Implementation: Estimation of Return on Investment

The relationship between a company’s value and the implementation of an IT system (including a BI system) may be proved if the implementation is interpreted as an investment project. If the potential increase of a company’s value due to an investment project is measured by net present value (NPV), a direct relationship can be observed:

NPV > 0 → potential increase of company’s value.

We can talk here only about the potential increase of value because a concrete investment project does not exist in a vacuum. Quite the opposite, it shares resources with other processes and coexists with other investment projects. And that’s why the project itself may be successful as an investment, but at the same time it may lead to losses in other areas (e.g., by key workers being overloaded with work), which result in a decrease in the company’s value.

The financial interpretation of BI projects was first thoroughly discussed in an article by Steve and Nancy Williams.²⁴ From a financial point of view, the implementation of IT systems is a material investment, in which both basic and advanced profitability assessment methods can be employed. In practice, only basic methods are used: ROI, payback period, and discount methods, such as net present value, internal ROI, profitability ratio, and discount payback period. The interpretation of NPV is shown in Figure 6.12; the investment cost is in following years compensated by discount cash flows. The main problem with using financial investment assessment methods is the necessity to forecast the effect on the cash flow by a given investment. Such forecasts by definition have a wide margin of error. Positive cash flow can be generated by reducing costs (as in the example of a reporting system), and cash flow is relatively easy to estimate in the case of IT investments. However, cash flow resulting from profit growth (as in the example of a profit forecasting system) is extremely difficult to estimate. The profit growth depends on a combination of many factors that influence it at the same time. It is a highly complex issue to formally single out which of those factors increased profit and measure their share in the profit growth.

Another problem is an attempt to formalize factors that cannot be measured. As the example of an OLAP system shows (see Table 6.4), the accurate use of a new reporting system may give a manager a deeper business insight. Let us assume that a sales manager finds out the next day rather than one week later about an unexpected growth in sales of a given product. Are we able to estimate the value of such information obtained a few days earlier? If adequate operations are performed, we may reduce so-called lost sales. And here the problem emerges of how we can measure the value of information. Let us consider the most vital problems that have an influence on the value that is attached to information:

1. Amount. The classic measurement of the amount stems from information theory (Claude Shannon’s definition) and is based on entropy, which is of no use in business applications. In economic terms, the amount of information (value) should refer to practical value based on the extent to which it can satisfy human needs or to exchange value, that is, the possibility to exchange information for other goods.
2. Accessibility. The value of information decreases when an increasing number of people have access to it.
3. Credibility of the source. The value of information depends on the credibility of a provider.
4. Usability. The value of the information increases if a recipient can use it for practical purposes.
5. Quality. There is a relationship between the value of information and its quality, that is, whether a given message is true and complete.
6. Context. Experimental research proves that the subjective value of the information is subject to an endowment effect,²⁵ in which the value of a good in a salesperson’s eyes is higher than in a purchaser’s eyes.
7. Life cycle. The value of information depends on the time within which it is delivered to a user. Outdated information is usually worthless.
8. Possible applications. Information is of greater value if the same information can be simultaneously used by different people in an organization in different ways.
9. Information symmetry or asymmetry. It is a paradox that unless a purchaser obtains the given information, he or she does not know its content, but when it is revealed, he or she may not be interested in paying for it. It refers to the thesis according to which information is an experience good, which means that it can be discovered exclusively through use.²⁶

Economic models, which deal with difficult but significant problems, are still in the pipeline; here David Lawrence’s monograph is recommended.²⁷