Reading a number honestly, saying what it actually supports, and presenting it so the decision gets made.
Product decisions are routinely made from small samples read with more confidence than they deserve, and the fix is a working understanding of statistics rather than a formal course in it. This list starts with the two plain language introductions, moves through framing a question and sizing user research, and ends with the presentation and ethics material. None of it asks for maths beyond what you already have.
Wheelan covers descriptive statistics, probability, inference and regression through examples rather than notation. The best first book here if the last statistics you studied was at school and you would like that to stop mattering.
Spiegelhalter works from real questions and real datasets, and is unusually clear about uncertainty, p values and what a confidence interval genuinely claims. The chapter on how statistics goes wrong repays a second reading.
Shron is about framing the question before anyone touches the data, using a structure of context, need, vision and outcome. Short, and it prevents the analysis that answers something nobody asked for.
Sauro and Lewis on putting defensible numbers on usability work, including confidence intervals for the small samples research actually produces. This is where the question of how many users you need to test with gets a real answer.
The applied statistics of a live product, covering statistical power, novelty effects and metrics that move for reasons unrelated to the change. Read the relevant chapters before you interpret your first significant result.
Knaflic on chart choice, removing clutter and directing attention, with before and after examples on almost every page. Analysis nobody acts on is usually analysis that was presented badly.
O'Neil on models that are large in scale, opaque in workings and damaging in effect, drawn from lending, policing and hiring. Useful for spotting the same three properties in something sitting on your own roadmap.
Why you will see a pattern in noise and then stay confident about it, from the researcher who named the effects. The background that makes everything above land harder.