This article originally appeared on the BeyeNETWORK
In the nineties, many companies faced the difficulties of implementing data marts and data warehouses. After many failures, CXOs finally became aware of the fact that a siloed approach – whether it is IT led, departmental island development or exclusively a corporate initiative – does not lead to success. A Business Intelligence Competency Center is a joint effort, combining initiatives and leading to a business intelligence infrastructure that protects the interests of all parties involved.
The business intelligence infrastructure nowadays is a complex environment of central data warehouse tables comprised of master data solutions, thematic data marts, metadata tables, ETL scripts, audit trails and logging scripts, information delivery algorithms, near real time monitoring, etc. This infrastructure is ideally exploited by a Business Intelligence Competency Center (BICC) containing analytical skills, functional business knowledge and IT knowledge. This is the BICC that Gartner suggested some years ago. The BICC is responsible for the technical gathering of data as well as shaping it into management information that can be used to improve the business processes. The delivery of reliable management information with suitable access is a service of merit but should reflect the cost of producing it.
Costs Versus Benefits
The Business Intelligence Competency Center is under a constant threat of having to prove its value over and over again. Repeatedly, the BICC has to compete with siloed extraction efforts. Although it is recognized that local extracts are less reliable and not reusable and that the outcome and the definitions depend on the person that constructs the report, it cannot be denied that local application developers often have a special status. The local manager appreciates their special skills and most of all the speed of their information delivery.
The arguments to get the BICC budget covered are usually vague and abstract. Goals such as competitive advantages, knowledge of the market and the customer, predictive analysis to prevent commercial disasters, speeding up the plan-do-act cycle are hardly smart and hardly measured as a result of the BICC services. To counter the threat of being sacrificed with every crisis that the company meets, the arguments to run it should preferably be of a more numeric character.
The BICC as a Profit Center
From this situation grew the idea of exploiting the BICC as a profit center. The costs of the exploitation of a BI infrastructure would be covered by the subscriptions to the information delivery services. Running the BICC as a profit center could be viable.
To run a BICC as a profit center, a few principles should be followed. The costs being made should reflect the price being paid. That is, the cost of system processing and maintenance is reflected in the price for management information. It is also important that the price of BICC management information should compete with siloed extraction efforts. Now that is a challenge!
Potential BICC customers could include:
- Report users
- OLAP users
- Application users
- Data miners and other specialist users
Suppose a local manager wants management information for a specific purpose. Looking on the BICC portal and opening the data dictionary for available information, he or she finds a report that might do the trick. He subscribes to the report. There should be a few parameters to define the price. First, there is the depth of the data – the more the depth the higher the price. Secondly, there is the frequency – the more often a fresh report is being sent, the higher the price. The customization (i.e., the adaptation of a standard report) has to be calculated and offered by the BICC.
OLAP users look at the BICC product market combinations with a different view. They usually do not want reports. They want OLAP environments and/or data marts – more or less prepared data that they can shape into useful reports. Their parameters of choice are again the frequency of data refreshment, tool subscription and, inevitably, tool support. If they find any problems in constructing a report (technical tool support) or difficulties in interpreting the data (analytical and business skill support), they need a BICC specialist to support them in their information construction. Parameters could be response time – anything between 7x24 or within a week. These support contracts are a risk. A BICC can have high costs for support people on the payroll, especially for specialists of rare tools or tools that are difficult to use. Also, the need for support is usually concentrated within specific periods. At the time when budgets are being set, everyone wants tool support. At other times, those expensive specialists are reading novels. A way to deal with this is to use the same people for developing applications and tool support. Another construction that might help is a time guarantee contract. For example, the contract with the user department consists of 100 guaranteed hours, whether or not they make use of it. This creates an opportunity to predict and set a budget for your tool support services.
Customization poses other questions. Often the first user of a management information application pays far more for the first iteration; and, what is worse, the implicit development of an overall business solution causes a considerable delay in application delivery. The reusability is an advantage to other future users. It is not an advantage from which this first user can profit.
The Importance of Being Earnest
An important property of the data warehouse is its reusability. Once built, a report or analytic data environment can probably be used by more users. Additionally, these users typically get their management information requests delivered much more quickly and sometimes at less cost than the first user. There are, however, ways to deal with this injustice. Suppose that the BICC investigates the market for the potential of a new management information product and that the cost of developing this application is $400,000. The conclusion might be that potentially 75% of this application could be reused at least five times. The first user would therefore pay 25% for customizing the application (i.e., $100,000) plus one-fifth of the building price (i.e., 20% of $300,000 = $60,000) plus $20,000, which is a reasonable margin for pre-investment and profit. This total price has a far better chance of competing with local development efforts, even though application delivery speed will still lag behind. However, better and faster management information delivery, once the application is up and running, should easily convince the future user.
What is the Information Worth?
Earlier it was said as a principle that the price for management information should somehow reflect the costs of producing it. This is only part of the truth. When management information delivery is considered a product market combination, the commercial price also depends on the need for this particular information by the business. If the need is huge, the overall gain is more evident. As a result, the price can be higher.
The supply of material for the product is, of course, the data. A worthwhile consideration is to actually pay for the data. The data owner has the responsibility for the reliability and the correctness of the data; and when he or she is paid for that data, this responsibility becomes more explicit. And, of course, what goes for the business goes for the BICC: the bigger the need, the higher the price.
Information that is constructed in the data warehouse could be any combination of data of different owners. The overall company turnover could be the departmental turnover of all five departments of five different data owners. The new information item (e.g., overall turnover) is created by the BICC as a company asset. It is constructed of five purchased sub items.
The BICC Profit Center buys data and sells information constructed from this purchased data. The very construction of new information items by combining data of different sources is one of the main activities in a BI infrastructure and thus anchored in the product market exploitation procedures.
You can run your Business Intelligence Competency Center by selling:
- Contracts for tool support
- Reports by subscription
- Application delivery
- Ad hoc analysis on request
- Periodic or ad hoc data delivery
The price for each of those products can be calculated depending on:
- The price of producing the information from raw data
- Future reusability of both application and data
- The urgency of the business need
- The price of the raw data
- The costs of application development and maintenance
Looking at the BICC as a profit center makes it a lot easier to prove its reason-to-be. CXOs are easily convinced by financial arguments.
Author’s note: This article is based on the Dutch article: “Het exploiteren van een BI omgeving” published in: “Bi-Ware, de harde en de zachte kant van Business Intelligence”, ISBN 90-74562-12-4, Karien Verhagen, Array Publications, Alphen aan de Rijn 2006.