Self-Report Decay Rate

Self-Report Decay Rate is the gap between what a respondent says in a form and what the same person does when observed. Every research tool in this category captures stated data: a Typeform survey, a moderated test, a feedback widget. That data is cheap to collect and fast to report, but its predictive value falls as the decision gets more expensive. The framework asks agencies to score each client decision on two axes, cost of being wrong and distance from the moment of behavior, then route only low-stakes calls to self-report alone. A useful calibration: an Anthropic study gave 50 experts an AI agent whose preferences were set by coin toss, and those experts adopted the agent's views up to 86 percent of the time on judgment calls. If trained specialists defer that readily to an ungrounded opinion, client stakeholders will do the same with a 40-response survey. Pair stated data with observed behavior before it reaches a retainer deliverable.