Watching somebody use the thing, which is the cheapest way to find out that it does not work.
Every team believes its own product is obvious, and a morning spent watching three people use it settles the question. These cover how to run a session, how many people it takes, what to measure when somebody wants a number and how to report the result so that it gets acted on. Ordered from the one to read before the first session.
Start here. One morning a month, three participants and a screen recorder, with the facilitator script printed in the back.
The full treatment for when the stakes justify a proper study, covering planning, running and analysis at length.
The design principles that testing keeps confirming, which makes reading it the cheaper way to learn them.
For the moment somebody asks whether the finding is statistically real, with the sample sizes and tests to answer them.
The wider reference, for choosing between a usability test and the method that would answer the question better.
A usability session becomes an interview the moment somebody explains why they did that, and this covers the second half.
How much research the decision actually warrants, which is the question to settle before booking anybody.