Little’s Law state that the average number of customers in a stationary system (L) is equal to the long-term average effective arrival rate (λ)multiplied by the average time (W) that a customer spends in the system. It is represented by the following formula L = λ x W (Hendijani, 2021). Where;
- Work in Progress (L) is the number of items in process in any system.
- Throughput (λ) represents the rate at which items arrive in/out of the system.
- Lead time (W) is the average time one item spends in the system.
The Law was developed by a professor at the Massachusetts Institute of Technology (MIT) called John Little in 1954. He first applied the Law to queues in shops, where it was revealed that anything that can queue, and even software tasks, felt under this Law (Rusek & Chołda, 2018). His initial trial to publish this work never succeeded because he did not have enough proof. However, after gathering enough evidence, he published this work in 1961.
Once the management has accorded an agile project to a specific team, it will require that you measure the performance of that team. The administration may want to measure the team’s performance to determine future iteration, velocity or the management may simply want to report to stakeholders to make an informed decision towards the business (Potter et al., 2020). Therefore, Little’s Law can be very effective when measuring project performance and planning for the future.
The Law states that L = λW. This can be applied to the performance of an agile team as follows, assuming the team follows the below process: Start (Receive User Story) —> Develop User Story—> Test User Story—>Deploy User Story (Finish). Where L, the average inventory/WIP in a system, represents the User Stories. λ the arrival rate represents the rate at which the team can complete the user stories, and W is the average time each user story takes to be developed from start to finish (Whitt & Zhang, 2019). For example, a team can finish a project by completing L – 100 user stories, in λ- 20 stories per iteration, where each iteration is four weeks. Then W (Average number of iterations) = 100/20 = 5 iterations. Hence, the team can finish 100 user stories in 5 iterations. What if, in 2 iterations, the team only completed 30 user stories. Then, the team’s new rate of completion λ is 30/2 = 15 user stories per iteration, showing a reduction in iteration. Therefore, the project will need L – 70 (100-30) user stories remaining, at the rate of λ – 15 user stories per iteration, which is W – 70/15 = 4.66 iterations more to complete the project. This means the project will take approximately 2+4.66 = 6.66/7 iterations instead of the initial 5 to be completed. Therefore, a relevant course of action for the project is necessary.
Little’s Law was initially being used as queuing theory majorly. However, after further research, the Law is being applied in almost all kinds of systems. Little’s Law is basically tied to the following three flow metrics, throughput, cycle time, and work in progress, all relating in one formula (Medonos & Jurová, 2016). Understanding the relationship between these three flow metrics allows one to measure the work process.
References
Hendijani, R. (2021). Analytical thinking, Little’s Law understanding, and stock‐flow performance: two empirical studies. System Dynamics Review.
Medonos, M., & Jurová, M. (2016). Implementing lean production application of Little’s Law. Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis, 64(3), 1013-1019.
Potter, A., Towill, D. R., & Gosling, J. (2020). On the versatility of Little’s Law in operations management: a review and classification using vignettes. Production Planning & Control, 31(6), 437-452.
Rusek, K., & Chołda, P. (2018). Message-passing neural networks learn little’s Law. IEEE Communications Letters, 23(2), 274-277.Hayati, F. (2017). Some perspectives on the application of Little’s Law: L= λW. International Journal of Modelling in Operations Management, 6(3), 141-152.
Whitt, W., & Zhang, X. (2019). Periodic Little’s Law. Operations Research, 67(1), 267-280.