How many projects do you work on that the leadership ask you two quarters later if you are sustaining the gains? Hopefully, more than I have through my career. One tool that I find that is not used often enough are the 1-sided and 2-sided t-tests. These tests can prove that the team’s work has been statistically significant from the perspective. I will be setting up a Vlop shortly to explain, with an example of a t-test but below is a discussion related to the uses of the tool rather than the mechanics.

t-tests should be used when testing if there is a change. This is part of the baseline data / measure phase. This will show you when your pilots have a statistically significant change in the metric. Many statistically astute leaders know this but don’t use this as a tool to track process improvements. It is not as slick as DOE’s and ANOVA’s to see where there are issues but those tools aren’t as easy to understand and, from my perspective, limited in use.

t-tests can easily be shown throughout the process life cycle. From baseline data through control, the t-test can show improvement. This can be used for almost any audience too. I have used this with Finance teams to see if there has been sustained improvement that leads to an individual G/L account. I have used this with parts of HR organizations related to talent acquisition, Outsourcing effectiveness, and payroll transactions. I have had to explain this with some groups but I have not had any say that this doesn’t make sense or not a key part of understanding improvement.

t-tests as trigger points. As an example, every week we check downtime as it relates to production. If the t-test shows that the average downtime is greater than the standard, move into corrective action, root cause analysis, countermeasure.

t-tests should be used to recognize improvements during a project, report out findings at the end of an improvement cycle as well as a dashboard metric to know if the process is staying on track or not.