Best Test

Find the cheapest real-world observation where your leading options would predict different results.

After the offsite, your team has 22 sticky notes of ideas for growing the newsletter. The follow-up meeting is chaos: someone champions referral bonuses, someone else defends a podcast, comparisons ricochet between ideas that have nothing in common, and an hour later nothing is decided except to meet again.

Family Sort says: before comparing options, cluster them by their underlying mechanism or form, so you compare families rather than duplicates. The cue is a long list where head-to-head comparison keeps stalling because items differ in kind, not just quality. Run it in three steps: label each idea's causal mechanism (paid acquisition, referral loops, content expansion, partnerships), collapse near-clones inside each family since they would succeed or fail for the same reason, and hold the real debate at family level: which mechanism fits your goal and resources? Only then pick the best specific idea inside the winning family. What visibly changes is that 22 comparisons become four, the meeting produces a direction, and champions stop arguing past each other because they can see their pet ideas were siblings all along. Skip the sort when the list is already short and structurally varied; sorting five distinct options adds ceremony where a straight comparison would do.