The Value of Business Intelligence

According to the PricewaterhouseCoopers Global Data Management Survey of 2001, “Companies that manage their data as a strategic resource and invest in its quality are already pulling ahead in terms of reputation and profitability.” [1] This statement implies a quite subtle yet radical notion: Data should be treated as a strategic resource. According to a traditional view, data is the “fuel” driving the automation of a business operation, implying that a company uses computers to help run its business. The forward-looking view of data internalizes the notion that strategic knowledge is embedded in the collection of a company’s data and that extracting actionable knowledge will help a company improve its business.

This leads to another intriguing idea—that a company may acquire a competitive edge by viewing itself as an information business instead of taking the traditional industry—or vertical—view. Consider this: Is a supermarket chain a business that sells food, or is it a business that exploits knowledge about customer preferences, geographical biases, supply chain logistics, product lifecycle, and competitive sales information to optimize its delivery, inventory, pricing, and product placement as a way to increase margin for each item sold? The answer to that question (and its corresponding versions in any industry) may ultimately determine your company’s long-term viability in the Information Age.

So how do you transform data into a strategic resource? A part of that process involves properly applying new technology to your data, but the most important part is being able to understand and subsequently build the business case for the value of information. This is partially an abstract exercise and partially a discrete one, and in this chapter we look at the difference between the traditional use of data in a transactional environment for the purpose of effecting operational side effects and the modern view of data as a valuable resource that can be used for analytical purposes. We introduce the savvy manager to the value of information, discuss what the aspects of information are that make up this value, and indicate what kinds of processing can be performed to add value to data. We also look at some examples of business intelligence (BI) applications to guide your understanding of how to build the business case for a BI program.

The Information Asset and Data Valuation

Is data an asset? Although I have never seen a company’s data listed as a line item on its list of assets and liabilities, there are some reasons to consider it both ways. Certainly, if all a company does is accumulate and store data, there is some cost associated with the ongoing management of that data—the cost of storage, maintenance, office space, support staff, etc. This should show up on the balance sheet as a liability.

Alternatively, data can be viewed as an asset, because data can be used to provide benefits to the company, is controlled by the organization, and is the result of previous transactions (either as the result of data creation internally or through a data purchase). But organizations do not treat data as an asset; for example, there is no depreciation schedule for purchased data. Treating data as an asset is important, though, because it allows us to build the business case for investing in BI when we can show how the value of the data asset is improved.

That implies that we must have some way to measure the value of data, and this is where we get stuck. There are relatively few situations where we can accurately assign a discrete price for information, and this pricing structure is more frequently value based than if data were treated as a commodity. As an example, the telephone company charges you, say, $1 for each directory assistance inquiry, although that same telephone number could be acquired free through an online directory. The difference in cost is based on the convenience value of being able to pick up the receiver and get the number right away.

In most cases, though, the value of information depends on a number of contributing factors, and my discussion of these factors is adapted from research by Daniel Moody and Peter Walsh [2]. What is interesting is that as we are made aware of these factors, we can get a lot closer to developing a model for information valuation. This is not to say that we can accurately enumerate data as an asset on the balance sheet, but it does give some parameters to understanding the value of information and, consequently, the value of BI.

The Time Value of Data

There is a timeliness or currency component to the value of information. Here’s a simple example: If it is March 1st and some prophet told you today that for a certainty the March 2nd closing price of IBM stock is going to be $10 higher than the March 1st price, you can exploit that piece of information today to buy as much IBM stock as you can and sell it at the $10 profit on March 2nd. Yet if I gave you the same information on March 3rd, you could not exploit it in the same way. In this case, a large portion of the value of that piece of information is bound to its timeliness.

This example may appear to be a bit extreme, yet the concept is clear that a portion of the value of information is related to time and that value may degrade as time elapses. Because stored data represents a snapshot of a real-world state at a particular point in time, then in the absence of continuous maintenance, as the world changes our snapshot grows more out of synch with reality. For example, our direct mail database may be of high value at its initial creation, but because it is estimated that about 20% of the population changes addresses each year, then if that database is not updated with new addresses, its value declines with time. Not only that, bad data actually can be viewed as a liability: There is no reduction in the fixed costs associated with managing that bad data, and using the data as part of an integration or linkage process will yield incorrect answers, and the value of the data set as a whole declines.

Information as a Sharable Resource

As opposed to any other raw resource used in a manufacturing process, data is a raw resource that are not used up. That means that information is sharable, and the value of information increases as more people use it. An example of this is the knowledge of a process for sales professionals to alert them to the best time to contact a prospect. The knowledge of this process can streamline the process for any particular salesperson. But even if that salesperson were to share that knowledge with other members of the sales staff, there is no degradation in the value that can be achieved by any one of the individuals aware of that knowledge. This means that the value of that information is multiplied by the number of people who know it.

Increasing Value through Increased Use

For most assets, as usage increases, there is some depreciation in asset value. For example, every mile a car is driven decreases the value of the car. On the other hand, the value of data does not decrease with use, because there is no degradation in information based on the number of times it is viewed. When everyone in the organization knows what information is available, how to access that information, and how to exploit that information, the value of that information rapidly increases. If data is stored, managed, and never used, there is no added value, and, as I mentioned in the introduction to this section, it actually becomes a liability.

Increasing Value through Quality

Let’s look at the stock price example one more time, but let’s change it a bit: This time, the prophet is a false one and tells you that the stock price will fall $10 by the close of business tomorrow, although it really will rise $10. The way you will have exploited what you believe to be true will ultimately result in a significant loss of value instead of an increase. This highlights the value of accuracy of information and the requirement for not only expecting high levels of information quality, but also having a means for defining quality metrics and measuring using those metrics. Having some understanding of the measure or quality of data being used for a decision support process lets the decision maker determine the risks associated with relying on that data.

Increasing Value through Merging

The process of combining bits of knowledge provides significant leverage when increasing the value of information. Having sales channel information is of value; having supply channel information is valuable; combining supply channel information with sales channel information provides knowledge about the movement of products from supplier to customer.

Information increases in value when it can be used to enhance and expand other pieces of information. The BI process revolves around the ability to collect, aggregate, and, most importantly, leverage the integration of different data sets together. As we will discuss many times in this book, there is a large increase in the value of information if it can be used as leverage in increasing an actionable knowledge base. In other words, if we can take two pieces of information, link them together, and infer something new that could not have been learned independently, we can exploit that inference for competitive advantage.

Value versus Volume

Contrary to the behavior of other assets, we do not necessarily gain greater value by having more information. The sheer amount of information that is produced every year is almost unbelievable; trying to integrate that with what has existed before seems like a gargantuan effort. And in fact, the more data an individual is presented with, the less likely he or she is to absorb any of it.

The complexity of data integration grows steeply as the number of data sources is increased. Data from alternate sources infrequently share data models, representations of the same kinds of entities, or even coded reference data. Each data set that is added to the mix must be integrated with all the other sets already extant in the repository, bringing along all the problems associated with that integration.

On the other hand, there is a perception that the more information there is, the better. So there is some fine line between having the right amount of information and having too much. Having the right amount implies that this information supports he defined business requirements and assumes that you are able to integrate that information and provide and present it within the required time.

Not only that, there is a qualitative difference between having lots of data that comes from disparate data sources and having lots of data that derives from the same source. For example, maintaining a large amount of historical transaction data (such as point-of-sale data or call detail records) may prove to be more valuable when it comes to analyzing trends over longer periods of time.

Measuring the Value of Information

One way to assess the value of information is to look at some traditional valuation models for other assets.