02-01-003When a better funnel produces a worse businessA funnel that doubles its leads and halves their cost can still cut expected contribution by two thirds. The objective function comes before the dashboard.02-01-004A prediction is not an intervention effectThe account most likely to buy may be the one that least needs a call. Targeting needs the effect of an action, not a forecast of the outcome.02-01-005The evidence has its own clockA reporting period can end before the outcomes inside it have had time to occur. Cohort age, competing outcomes and statistical resolution decide what a result can mean.02-01-006How much is a better decision worth?An experiment can be valid and still unnecessary. A worked example of the value of information, the cost of waiting and the choice to act.