Learning Curves and Capacity Analysis

Learning curve analysis serves as a valuable tool in capacity analysis by enhancing the accuracy of capacity projections (Smunt, 1986). When a firm has consistent learning curves and can gain detailed data about its processes, data analysts can build statistical models that predict the efficiency of production at any given unit. This effect is commonly called the economies of learning, echoing to the widely known economic term of economies of scale. The general idea is that as a firm produces more of a product or service, the firm, both institutionally and on an individual employee level, learns more about the processes and becomes more of an expert, leading to greater efficiency.

Let's say you want to build a house. You hire ten random people who may or may not have any experience building houses. You could safely assume that house #1 and probably all of the first few houses are going to take longer than expected, go over budget, and have a high defect rate. By the time you get to completing an entire subdivision of 50 houses, your ten-person team that started out not knowing a screw from a nail will likely be really good at building houses.

This isn't an example of economies of scale; after all, there are only ten people building one house at a time; instead, it is an example of economies of learning. As your "crew" built more and more houses, they were able to learn from mistakes and gain valuable experience that made their processes well-oiled and polished by the time they got to number fifty.

In capacity analysis, the principal or manager seeks to efficiently allocate resources to provide the most valuable outcome within the given limitations. By acknowledging, measuring, and predicting the impact of economies of learning or learning curves, the lower initial productivity, later rising productivity, and eventual leveling can be planned for and expected. If, in our house building example, we expected that houses took three months on average to build and our first three houses took an average of five months each, we may believe that we've made a terrible mistake and decide to cancel the project. But if we acknowledge that the production time is not constant and is instead subject to learning curves, we can predict that our production times will (hopefully) eventually start to fall and average to a more profitable level as our builder learn more about their craft. This, of course, assumes ceteris paribus.

Why did I post this? This is a slightly edited version of one of my weekly discussion post contributions while completing my MBA. Concepts like this are simple to illustrate but not always easy to understand. I decided to share their excerpt to perhaps help shift the mindset of a reader from reactive decision-making to strategic planning. I'd love to hear your thoughts, too; just add a comment below.

References

Smunt, T. L. (1986). Incorporating Learning Curve Analysis into Medium-Term Capacity Planning Procedures: A Simulation Experiment. Management Science, 32(9), 1164–1176. https://doi.org/10.1287/mnsc.32.9.1164?