We frequently struggle with relating to inquisitive clients what the reach method "threshold" in TURF does.
We usually prefer the Threshold because it allows us to be deliberately more conservative than "better than a random set" or "top item". Threshold 80% tells is that the results have at least 1 items which has exceptionally high value [or whatever the framing] for the customer with high certainty.
However, clients regularly trip up and instead understand "the set has 80% probability to reach xx% of the sample" or "at least 1 item is among the top 20%" and other variations.
Are there any other ways/ideas of framing the meaning (and how Threshold operates) here?
Our clients are genuinely interested and generally statistically apt, so the frequent advice of "best not bother the client with the complicated fine print" unfortunately doesn't quite work for us 😅