|← Divine Command Theory||Article Summary →|
Managers in the information technology industry are usually faced with difficult task in determining whether it is worth to upgrade their firm's software whenever a new version of the software is released by the vendors. Many factors are considered when deciding whether to upgrade a software or not and the time to do this. Such factors include: the life cycle of the software and the organization's learning curves which are not in normal cases aligned. In respect to the time to upgrade software, a firm may gain both negative and positive effects if the upgrading is done early and when it is done late.
When early upgrade is done, a firm might fail to realize productivity gain and may as well lose investments that were done in earlier stages of upgrading. On the other hand, if late programming is done, a firm may lose the opportunity of gaining productivity gains that are associated with a new version of software. A firm's major concern is to determine the best time to upgrade software while at the same time realize maximum productivity gains. One option of realizing productivity gain that will be criticized in this paper is the use of a learning curve by Ngwenyama et al (2006). This paper will consider the various parameters that should be considered when making decision of the right time to upgrade software. A relationship between these parameters will be used to establish a mathematical framework used to determine the most appropriate time.
Framework of analysis
The increasingly complex decision of determining the best time to upgrade any new IT application has prompted many managers to find out the best option to maximize firm profit. Economist and industrial engineers are some of the professionals who are who have shown interest in the process of learning in the recent past. One of the early studies about the learning curve was done by Bryan and Harter in 1899 when they did an examination of learning patterns of adults using Morse code. The study found out that there was four times improvement in productivity in the interpretation and formation of messages after ten months experience. These conclusions have been used in studying organizational productivity since they withstood time testing. However, these conclusions did not consider the general learning limits when applying Morse code. This is because learning curves are also related to limits that occur naturally; such limits are either in the form of technology or resources.
These limits are associated with complications that lower the pace of learning to do some things in an efficient manner. Such limits are for instance; limits to perfect a process, product or even measurements. Studies on natural limits are concerned with diminishing returns on investment. In his case, the limits can either be financial, political or physical and they affect the system of resource development in general. However, there are also useful natural limits which are meant to measure the return on efforts invested in the learning progress. The ratio of the return on invested effort is denoted as R/I, while the inverse ration is used top measure the learning difficulty.
Learning curve is supposed to indicate the characteristics of improvement in performing certain tasks. During the initial stages of learning a task, the performance is usually low and increases gradually with time. These characteristics of learning curve are evident in both individuals and in organizations. When new skills and routines are learnt collectively, this form of learning is referred to as organizational learning. The need to establish a relationship between organizational learning and productivity prompted many scholars to work hard in the 1930s. One of such early scholars was Wright who tried to maker use of organizational learning curve in his study to produce airframes. In the course of this study, he observed that the labor time spent the production of airframes was decreasing with the number of airframes produced earlier. This kind of relationship is represented by the equation, y =ax-b where y is the labor units required to produce xth unit, a is the amount of labor to produce the first unit and b is the reduction rate in labor. This form of log-linear learning curve presents a scenario of performance whereby the highest level of performance is witnessed during the initial stages. This form of a learning curve proceeds with the performance increasing dramatically until a later stage when the curve takes the shape of a plateau. This is not possible in a in an environment that is full of limit factors that are not predictable. As mentioned earlier, such factors include the level of technology and resource available. These factors affect performance by in that they make the process efficient when available in the correct amount and status. Therefore, when technology and resources are available, they log-linear learning curve will indicate that performance increases slowly at the initial stages and later increases dramatically when more of the factors are supplied.
Learning curve is used in a case of Ontario Life, a medium-size company that deals with insurance matters. In order to perform its value-chain processes, the firm employs a certain software technology. The vendor of the software announced that the software had been upgraded after and replaced with a new version just five months after Ontario life had implemented its software. The firm IT manager was confronted with the decision of determining whether it was worth to upgrade the software right away or wait until the firm had gained enough from the technology it already had. Another option that the IT manager had to upgrade the software when it was already mature and in this case, the time needed to wait for this was not known.
Ngwenyama et al (2006) were interested in developing a mathematical framework that would help firms determine the best time to upgrade software without affecting the productivity of a firm. This they did by considering several parameters and creating a relationship of between these parameters that would help to drive such a mathematical framework. There are basic concepts that a firm should consider when learning the application of a technology since the use of technology in organizations cannot be abrupt. According to this paper, there is quit considerable of time that is required for a firm to launch the use of new technology. This amount of required time can be of negative effects to the firm in terms of competing with other global firms which have flexible decision making processes. Modern firms develop plans that would enable them make flexible decisions in order for them to have their market share at any time. The technology learning process in this paper uses a parameter r to denote dollars/time unit or value as a result of the value gained in the use of technology or cost reduction associated with the application of this technology.
Ngwenyama et al (2006) provides that the value of r would remain low when the new software is introduced into the firm since the uses of the software would take time to make the software compatible with the routine and coordinating mechanisms of the firm. When the new users learn to use the software, the value of r increases gradually until it reaches a plateau which in this case will be the maximum productivity of the firm. A graphical representation of this scenario can be the graph of r against time which would take an s-shape. Such an interpretation fails to consider other factors that determine the r value other than the introduction of new software. The value of r may be affected by the resources available in the firm. For instance, the work force in the firm plays a major role in determining the level of productivity. The firm can still introduce the software as early as possible but the number of personnel working in the various departments of the firm may change and therefore the maximum productivity of the software may fail to be realized. Such a scenario imply that the software would not assist the firm to make more profit than before and therefore the value of r cannot always increase consistently. There are also many other factors that determine the amount of dollar/time made by a firm which are not considered in this paper. This is because the profit made by a firm in every month is not always equal due to the economic factors that surround a business environment. As a result, the graph of r against time should therefore not be a smooth s-graph.
Ngwenyama et al (2006) also indicated that s-curve graph might have resulted due instances where a new software is introduced into a firm before fixing all the defects and before providing enough documentation and training materials. The writer of this paper assumed that any of the mentioned factors would be responsible for the s-curve shape which they referred to as a learning curve. In research, any written statement should be precise and based on some evaluation. Therefore, the assumption by Ngwenyama et al (2006) that any of the mentioned factors would have caused the graph of r against time to be S-shaped is not acceptable in research. Instead, they should have mentioned at least one of the factors and explain why it would cause the effect in order to make a precise statement.
Ngwenyama et al (2006) used several approaches to determine the mathematical framework that would guide a Firm on the right time to upgrade software. In one of the assumptions, they assumed what would happen when a firm starts learning the software and while still in this stage, the vendor makes an announcement that there is an upgrade version of same the software which is come at time t1. In this assumption, they assumed that they were adequate funds to facilitate the upgrading process and they did not care the current value of the upgrading process and the other options of investment. In a real situation, employees in a firm are instantaneous learners implying they do not require learning when a new technology is introduced.
In this assumption, Ngwenyama et al (2006) assumed that all members in an organization are always up to data in matters relating to technology and which is actually not true. Normally, when the software is upgraded, the whole process is modified based on current technology and this calls for the users of the software to undergo a learning process. However, this kind of an analysis is still useful because it provides an opportunity to identify the required parameters and at the same time help to establish the kind of relationship that exists between these parameters. The second reason why this case is useful is because it would provide a good way to approximate the maximum efficiency of upgrading the software (rt) since the firm is considered to have instantaneous learners.
Ngwenyama et al (2006) observed that, in the case of an instantaneous organization, the graph of r against time would be a horizontal line a not S-shaped curve. They argued that, in this case the organization does not require any learning and therefore the new software starts to be used immediately. R1 becomes directly proportional to the time (t). This kind of scenario is not common because many firms would not accept to upgrade the software while they still are in the process of recovering the cost incurred in the application of the first version of the software. As mentioned earlier in the easy, firms' main focus is the maximization of productivity and therefore even if the vendor announces an instant upgrade, the firms would require time to accept it. The total cost of upgrading the software is the amount the vendor would be paid by the firm. Additionally, a firm is expected to incur some cost in the loss of the value of r due to the process of learning new technology. The mathematical representation of the method to determine the best time for a firm to upgrade the software is dependent on the learning curves obtained when using the old version and the curve when using the new version.
Ngwenyama et al (2006) identified that the best time to upgrade the software will be the intersection of the two curves denoted as r1 and r2. This time would be when the firm has learnt how to use the new software better than it had known how to use the old version of the software. However, even after realizing the best time to upgrade the software, a firm still needs to determine when the cost would be recovered. This time is considered as the time when the cost of applying the first and the second version of the software would be recovered.
The mathematical model presented by Ngwenyama et al (2006) is efficient in assisting IT managers determine the time they can upgrade the different types of software introduced in the market. This is because the model has employed a theoretical framework particularly, the learning curve which provides an empirical analysis of considered parameters. The model could be more effective if Ngwenyama et al (2006) provided a method to recover the costs incurred in applying the old version in order to realized maximum productivity. The model has only illustrated the time when the costs would be recovered. This can be realized by the firm increasing productivity in be diversifying its production.