Simple Linear Regression

Simple linear regression is one of the most practical and accessible tools in the Improve phase because it helps you move from understanding relationships to actively predicting and influencing process behavior. While the Analyze phase focuses on identifying potential drivers, the Improve phase requires you to quantify those drivers and determine how adjusting them will affect your outcomes. Simple linear regression provides that bridge. It gives you a mathematical model that describes how one input (the predictor) influences one output (the response), allowing you to make informed, evidence‑based decisions about where to intervene. 

At its core, simple linear regression fits a straight line through a set of data points in a way that minimizes the distance between the line and the actual observations. The line is defined by two parameters: the intercept, which represents the expected value of the response when the predictor is zero, and the slope, which represents how much the response changes for each unit change in the predictor. The slope is the heart of the model. A positive slope indicates that the response increases as the predictor increases; a negative slope indicates the opposite.

The power of regression lies in its ability to quantify relationships. Instead of saying “cycle time seems to increase with workload,” regression allows you to say “for every additional unit of workload, cycle time increases by 0.8 minutes.” This level of clarity is essential in the Improve phase because it allows you to prioritize interventions based on measurable impact. It also helps you communicate findings to stakeholders in a way that is both intuitive and grounded in data. 

Regression also provides a measure of how well the model fits the data. The R‑squared value indicates the proportion of variation in the response that is explained by the predictor. A higher R‑squared means the predictor is a strong driver of the response; a lower R‑squared means the relationship is weak or that other factors are more influential. While R‑squared should never be the sole basis for decision‑making, it provides useful context for interpreting the model’s strength.

Another important component is the p‑value associated with the slope. This tells you whether the relationship is statistically significant—whether the observed slope is unlikely to be due to random variation. A significant slope indicates that the predictor truly influences the response. However, statistical significance must always be paired with practical significance. A slope may be statistically significant but too small to matter operationally.

Simple linear regression also supports prediction. Once you have a model, you can estimate the expected response for a given value of the predictor. This is particularly useful when evaluating potential improvements. For example, if you know how cycle time responds to staffing levels, you can estimate the impact of adding or reallocating resources. Prediction intervals provide a range of likely outcomes, helping you understand the uncertainty around those estimates. 

In the Improve phase, simple linear regression is not just a statistical tool—it is a decision‑making tool. It helps you quantify relationships, prioritize interventions, and forecast the impact of changes. When used thoughtfully, it provides a clear, evidence‑based foundation for improvement actions.