The mean means nothing
As I asked before: don’t summarize, but plot the data.
Of course, the need to summarize comes sooner or later, and lately I’ve been using quartile plots, with perhaps outliers, as method. A new idea appeared this week! A post about the same irritation with the overuse of the mean: The mean means nothing makes the case for the CDF: the cumulative distribution function. It can be plotted without modelling, and does something very similar to plotting quartile ranges, without forcing you to show only a few arbitrary amounts (25%, 50%, 75%). The CDF shows you any amount, so you can use it to find inflection points (or not).