BUS303: Strategic Information Technology

Most important management activities can be viewed from a decision making perspective. Within organizations (especially larger ones), numerous decisions are being made on a continuous basis. For example, decisions about how to design an assembly line in a production facility, and how to structure work tasks for employees working on the line, will have direct and substantial impacts on the efficiency and effectiveness of the use of resources (employees, production materials). Not surprisingly, many organizations have expended considerable resources to acquire or develop information systems that are designed to help improve the efficiency and effectiveness of organizational decision making.

A wide variety of terms have been used to describe information systems that are designed to support the decision making of organizational members. These include decision support systems (DSS), group decision support systems (GDSS), executive information systems (EIS), knowledge management systems (KMS), and business intelligence (BI). Additionally, the term expert system (ES) is often used to describe systems that attempt to augment human knowledge by providing access to reasoning used by experts in reaching their decisions.

On occasion, the term managerial support system or management support system (MSS) is used as an umbrella term to encompass these diverse (yet related) types of information systems. While each type of system has some unique aspect (e.g. DSS are designed to support one individual, while GDSS are used by groups; EIS are geared toward the unique monitoring and control needs of individuals that are higher in the organization; and so on), they also share some common elements. At their core, all are designed to improve decision making within organizations.

Rather than examining each of these related types of system in detail, we will focus on the functions that organizational members need to perform and decisions they need to make, and then show how information systems may be used to support them. In doing so, we'll also see how different types of management support systems can come into play.

Controlling

One important function that needs to be performed within organizations is that of control, and managers are frequently charged with controlling certain organizational processes (or functions). Data, and information, are generally essential components for aiding control.

As an example, consider the task of managing an assembly line in a production facility. For the purposes of illustration, we'll use the example of a production facility that assembles office chairs. The facility obtains parts from suppliers, assembles them into chairs, and releases the chairs to customers. The customers are distributors who then sell the chairs to the ultimate buyers (mostly larger companies that buy office chairs in bulk).

Note that for the moment, it isn't important to distinguish who has the responsibility for ensuring that the assembly line uses resources efficiently in creating chairs that are of sufficiently high quality; it could be a shift manager, or it could be the employees working on the line. Either way, certain data need to be captured, and certain information created, in order to control the operations of the assembly line.

To be more specific, assume that a decision has been reached to keep track of each part (chair back, right chair arm, etc.) that is used as input; this is accomplished by ensuring that each part has a UPC (Universal Product Code) bar-code affixed to it when it is received from a supplier, and each part's bar-code is scanned by a bar-code reader before being used in a chair's assembly. When a part is scanned, the information contained on the bar-code is copied and stored in a production database. In addition, as each part is added to the chair moving through the assembly line, a record is kept (in the database) of the time at which the part was scanned. When the chair has been completely assembled, it is placed inside a plastic bag, and either a UPC bar-code or an RFID (radio-frequency identification) tag is attached (depending on the needs of the customer). The bar-code is then scanned (or the RFID tag is read by an RFID reader), which records the time at which the chair was completely assembled and ready for storage (or shipment). This record is also added to the production database.

One way of using data to control this process is to constantly monitor the length of time it takes from when the UPC bar-code for the first part is scanned to when the bar-code (or RFID tag) for the assembled chair is scanned. By recording this information over a period of time, it is possible to obtain a distribution of observations (e.g. the mean [average] length of time taken to assemble a chair is 15 minutes, and the standard deviation is 1.5 minutes). Using this information, it would be possible to write a computer program to monitor the times taken as each chair is produced, and notify someone if the time taken is excessively long or unusually short. Note that a person (or group of people) would determine the rule for identifying exceptions (based on past experience), and the software would be programmed to enforce the rule.

Using this approach, a shift manager could be alerted, for example, when a chair takes longer than normal to be assembled. The shift manager could then investigate possible reasons for the delay (e.g. a temporary delay occurred when there were several defective left chair arms in a pallet; the immediate supply of left chair arms was depleted faster than the right chair arms, and a fork-lift truck had to be sent to the parts storage area to retrieve another pallet of left chair arms). As a result of this delay, the shift manager might institute or revise a policy to reduce the possibility of a similar delay occurring in the future.

Note that the scenario described above is only one of many possibilities of how this business process might be designed and controlled, and hence how an information system could be designed to support it. For example, an alternative would be to have the employees on the assembly line responsible for controlling the assembly process. Instead of notifying someone of the time taken after a chair has been completely assembled, it would be possible to compare the time from the start of assembling a chair until each part is scanned, and therefore it would be possible to know much sooner if a problem is occurring. As a general rule, the business process should be designed first, and then the information system should be designed to best support the process.

Automating decisions

Whenever possible, organizations strive to automate certain types of decisions. Automating decisions can be much more efficient, since you are essentially allowing a computer program to make a decision that was previously made by a human being. In addition, automating decisions can lead to more consistent and objective decisions being reached.

Earlier in this chapter, we discussed the issue of the degree of structure for decision situations. Basically, the more highly structured the decision situation, the easier it is to automate it. If it is possible to derive an algorithm that can be used to reach an effective decision, and the data that are needed as input to the algorithm can be obtained at a reasonable cost, then it typically makes sense to automate the decision.

For example, the chair assembly company discussed previously might find it possible to automate the decision as to when to request the transportation of a pallet of parts from the parts storage area to the assembly line. This could be done by scanning not only the parts that are used in the chair assembly, but also any defective parts (which could be tagged as defective and noted as such in the database through a touch-screen monitor, keyboard or mouse at a workstation on the assembly line). By monitoring all parts that are removed from the temporary storage on the assembly line, the information system could determine when a pre-determined re-order level has been reached, and issue a request to the next available fork-lift operator to deliver the needed parts to the assembly line.

Davenport and Harris offered a framework for categorizing applications that are being used for automating decisions. Most of the systems that they describe include some type of expert system, often combined with aspects of DSS, GDSS, and/or EIS. The categories they provided include:

Solution configuration-these systems are employed to help users (either internal staff or customers) work through complex sets of options and features to select a final product or service that most closely meets their needs. Examples might include configuring a large or medium-sized computer system or selecting among a wide variety of cellular telephone service plans. The underlying computer programs would involve a type of expert system, including a set of decision rules that have been obtained from experts from the decision context.

Yield optimization-describes systems which use variable-pricing models to help improve the financial performance of a company, or to try and modify the behavior of people in some way. One example would be an airline, where 10 different people on the same flight might pay 10 different amounts for their tickets, depending on when they purchased the ticket, how frequently they fly with that airline, how full the flight is when they book their ticket, and so on.

Routing or segmentation decisions-these systems perform a type of triage for handling incoming requests for information or services. Examples include systems that balance the loads on Internet Web servers by routing requests to different servers, or systems for insurance companies that handle routine insurance claim requests and only route exceptional (unusual) requests to human claims processors.

Corporate or regulatory compliance-these systems ensure that an organization is complying with all internal or external policies in a consistent manner. For example, mortgage companies that want to sell mortgages on the secondary market have to ensure that they comply with all of the rules of that market when they are preparing the original mortgage. Similarly, insurance companies have to comply with federal and state regulations when writing insurance policies.

Fraud detection-these systems provide a mechanism for monitoring transactions and noting possible fraudulent ones. The approach used might be very simple, such as checking for a credit card's security code (in addition to the credit card number, to prove the person physically has the card in their possession). Other approaches can be quite sophisticated, such as checking for purchases that seem to be out-of-character for the credit card holder (based on past purchasing history). By automatically identifying potentially fraudulent transactions, and then having a human operator contact the card holder to verify the transaction, the credit card company can reduce fraud losses and increase their customers' satisfaction.

Dynamic forecasting-organizations all along a supply chain can decrease their costs of operations by reducing the amount of product (raw materials, work-in-process, finished goods) that they hold in inventory. Dynamic forecasting systems (that use historical sales data etc.) help manufacturing companies align their customers' forecasts with their own internal plans. This in turn helps them to reduce their inventory carrying costs and make more efficient use of their production resources (facilities and people).

Operational control-these systems monitor some aspect of the physical environment (such as wind speed or rainfall amount) or some type of physical infrastructure (such as an electrical power grid or a communications network). If an unusual event occurs (such as a sudden surge in electrical power at one point in the electrical grid), the system automatically performs some type of action (such as shutting down some nodes and re-directing power over others).

Many management decision situations are not highly structured, however, and hence cannot (or should not) be completely automated. Next we describe systems that are designed to support decision making, rather than automate it.

Supporting complex decisions

In an unstructured decision context, there may be numerous factors or variables that need to be considered. Often, an attempt to find the "best" decision with respect to one factor will lead to a poor solution with respect to another. Even when the situation is very complex, however, it is often possible to use information systems to help support the decision-making context.

An entire branch of management theory and practice, termed management science, has evolved to try and bring more structure to unstructured decision situations. Management science is based on the application of mathematical models, and draws fairly heavily on the use of statistical analysis techniques. Examples include the use of regression analysis (to assess possible empirical relationships), simulation (to identify potential solutions by varying certain assumptions), and optimization models (to generate a "best" solution when resource constraints exist).

When a mathematical model fits well with reality, it may be possible to create an information system to help automate the decision. When the fit is less than perfect, we need to augment the use of the model with the judgment of a human decision maker. The term decision support system is sometimes used to describe information systems that are designed to help address unstructured decision situations. Many decision support systems use management science techniques to provide decision makers with alternative options.

Decision support systems do not necessarily need to be large, complex information systems. For example, a sales manager for the chair assembly company might use spreadsheet software to develop a forecasting model that could be used to predict demand levels for a product (line of chairs). After building the model to include criteria believed to impact demand (price, success rates for promotional [marketing] campaigns, etc.) the sales manager could use it to help forecast demand and then decide what demand levels to forward to the production group.

Consider a more complex example. In the early 1990s, American Airlines was faced with the daunting task of scheduling about 11,000 pilots and 21,000 flight attendants on close to 700 airplanes on flights to over 200 cities. In addition, they had certain constraints, such as the maximum time pilots and flight attendants can be in the air during a specific time period. This problem could generate between 10 and 12 million possible solutions.

The scheduling challenge facing American Airlines (and every other major airline) is very complex. When you consider overtime costs, labor contracts, federal labor mandates, fuel costs, demand for routes, and so on, it is obvious that there is no perfect solution. If a solution is derived that is "best" for one dimension (e.g. reducing overtime pay), another dimension (such as holiday preferences) will likely be compromised. To address the situation, American Airlines spent over two years working on a scheduling system that used management science techniques. The result was an information system that saves the company between USD 40 and USD 50 million per year, by reducing wasted flight crew time.

The scheduling information system uses data such as flight crew availability (e.g. federal mandates concerning necessary "down time" between flights, etc.), flight crew capabilities (e.g. pilots licensed to operate certain types of airplanes), flight crew preferences (e.g. base airport, requested vacation days, etc.), airplane characteristics (seating capacity, range, etc.), and route characteristics (e.g. distance, historical demand, etc.) as input. It then uses the optimization rules and logic embedded in the software to generate possible schedules that satisfy as many of the constraints as possible. The proposed schedules could be viewed as information, which is then used by decision makers (those responsible for scheduling airplanes and flight crews) to produce final schedules.

Decisions concerning the scheduling of airplanes and flight crews to flight routes has observable outcomes (e.g. the number of times a flight is delayed because a flight crew was delayed, the number of flights crew members complete over a given time period, and so on). These outcomes can be measured and examined, leading to greater insights (knowledge) which can then be used to fine-tune the rules and logic in the scheduling information system.

Knowledge management

When experienced people retire or leave an organization, typically their knowledge leaves with them. In addition, many larger organizations (e.g. major information technology consulting firms) have many people who have similar responsibilities (e.g. IT consulting) that could benefit from each others' experiences, but because of the numbers involved (and geographical separation) personal communications among the employees is not practical. A type of information system that is designed to help address these situations is often referred to as a knowledge management system (KMS).

Knowledge management systems can take many different forms, but the basic objectives are to (a) try and facilitate communications among knowledge workers within an organization and (b) try to make the expertise of a few available to many. Consider an international consulting firm, for example. The company will employ thousands (or tens of thousands) of consultants across numerous countries. It is quite possible (in fact, quite likely) that one consulting team in, say, Spain is trying to solve a problem for a client that is very similar to a similar situation that a different consulting team in Singapore already solved. Rather than reinventing a solution, it would be much more efficient (and effective) if the team in Spain could use the knowledge gained by the team in Singapore.

One way of addressing this situation is to have case histories for all client engagements posted to a case repository, which employees from all over the world can access (using the Internet) and search (using a search engine). If the case documentation is of good quality (accurate, timely, complete, etc.), then the consultants will be able to share and benefit from each others' experiences and the knowledge gained. Unfortunately, however, it is often difficult to get employees to contribute in a meaningful way to the knowledge base (since they are probably more concerned about moving forward on their next client engagement, rather than documenting their experiences with the last one). In order for such systems to have any chance of working successfully, management may need to considered changes to reward systems and even to the organizational culture.

A different approach to knowledge management focuses more on identifying (and storing) details about the expertise of employees, and then allowing other employees to locate and contact these internal experts. This approach also has weaknesses, however, since the "experts" may spend so much time responding to requests and educating other employees that they have little time for completing their own work. As a result, employees may hesitate to volunteer information about their expertise.

Business intelligence

The term business intelligence (BI) is generally used to describe a type of information system which is designed to help decision-makers spot trends and relationships within large volumes of data. Typically, business intelligence software is used in conjunction with large databases or data warehouses. While the specific capabilities of BI systems vary, most can be used for specialized reporting (e.g. summarizing data along multiple dimensions simultaneously), ad hoc querying, and trend analysis.

As an example, consider a large grocery store chain. Similar to most competitors, the store keeps track of all purchases, down to the item category level. By that, we mean that the store keeps track of all the items that are purchased together (e.g. a package of diapers, a bag of potatoes, etc.) on a single receipt. The detailed data is captured and stored in a large database (and possibly copied into a data warehouse). Once that data is available, data analysts use business intelligence software to try and identify products that seem to be purchased together (which can help in product placement decisions), evaluate the success of marketing promotions, and so on.

When you consider the extent of the data that is captured at a check out, (goods purchased, prices, combinations, what is not purchased, date and time, how payment was made, and that much of that data can be combined with other data such as information available about the purchaser on loyalty cards, advertising campaigns, weather, competitors activities etc.) you begin to see the extent of the possibilities.

As with knowledge management systems, the value of business intelligence systems can be hindered in several ways. The quality of the data that is captured and stored is one concern. In addition, the database (or data warehouse) might be missing important data (for example, the sales of ice cream are probably correlated with the temperature; without temperature information, it might be difficult to identify why sales of ice cream increase or decrease). A third challenge can be that while data analysts may know how to use the BI software, they may not know too much about the context for the organizations operations. In contrast, a manager may know the organization, but not know how to use the BI software. As a result, it is not uncommon to have a team (a manager paired with a data analyst) to try and get the most information (and/or knowledge) out of a business intelligence system.