Do Operations Research (OR) tools and techniques have applications for product leaders?
What is Operations Research?
Modeling, statistics and optimization are tools that can aid effective decision making. Operations Research (OR) is a discipline of mathematical sciences that brings these aspects together to help make optimal or near-optimal solutions. Common features of OR include critical path analysis, network optimization and queuing. Previous threads have covered cognitive biases, conviction and over commitments that can negatively influence the quality of decisions in an uncertain environment. OR is commonly leveraged in the military and supply chain to reduce the impact of the above pitfalls, but can it also be used in the software industry?
“The concern of OR with finding an optimum decision, policy, or design is one of its essential characteristics. It does not seek merely to define a better solution to a problem than the one in use; it seeks the best solution. - C. West Churchman
The history of OR
Operations Research gained prominence through its use by the Allied Forces in World War 2 and it has since gained traction in a wide range of industries globally. The Statistical Research Group (SRG) at Columbia University was tasked with solving complex military problems during the war and legend has it that Abraham Wald of the SRG used OR to more effectively armor Allied bomber aircrafts.
Several Allied bombers were failing to return from missions due to catastrophic damages from German air defense and initial solutioning involved inspecting the damage on aircrafts that successfully returned. The preliminary proposal was to add additional armor to the areas on bombers receiving the most damage, but Abraham Wald concluded additional armor should be added to the areas of the planes without any damage as those were the areas where damage was leading to aircraft not returning.
Wald was able to leverage statistical analysis to incorporate information from the bombers that did not return into the decision of where to add additional armor to aircraft. Wald’s intuitiveness enabled the Allied forces to escape the perils of selection bias and decide based on all bomber damage not just the selection of bombers that returned from mission.
Selection bias appears when a sample of data does not reflect the wider population. It can result from multiple reasons including the time frame when a sample was collected, the size of the sample and attrition as in the case of bombers that did not return. Upgrading the armor on bombers based on the damages or lack of damages to planes that returned fails to account for the catastrophic damage on planes that did not return. Using math to estimate the damage and criticality of damage of all planes in the population reduces the uncertainty in the correct locations to add bomber armor.
Applications of OR
Operations Research continues to be utilized today in a wide range of industries including manufacturing, construction, farming, distribution and craft brewing among others. There are dozens of tools and techniques under the OR umbrella and critical path analysis, network optimization and queuing are three that have clear applications for product leaders.
Critical path analysis identifies the activities necessary to complete a task, time each task will take, order they must be completed and optimal way to complete them. Critical path analysis is commonly used by construction managers to optimally plan out the activities to construct large buildings but it could also be used by a product manager to identify their backlog, estimate work, identify successors and predecessors and account for team dependencies.
A common use of network optimization is calculating optimal placement of distribution facilities and warehouses around the globe based on set constraints but you could also use it to plan office locations and where to hire software engineers using inputs including cost, perceived value, overlapping work hours, etc. Network optimization could go as far as picking an office location that maximizes proximity to airports, customers, talented employees while minimizing costs.
Queuing theory in OR optimizes queues balancing level of service and cost. Aspects of queuing theory are commonly utilized in call centers but they could also be used to efficiently staff a product support team or IT service management team. Queuing theory can help product leaders decide when additional team members should be hired, the hours existing team members should work and time zone new team members should be hired in.
Deeper explorations into the tools and techniques of OR and its application for product leaders will appear in the toolbox at a later date.