Evidence Half-Life

Evidence Half-Life is the shelf life of a research finding before it stops supporting the recommendation it was gathered for. Agency research rarely fails because the method was wrong; it fails because the finding aged out while the deliverable kept quoting it. A survey fielded in March can still be accurate in September and still be unusable, because the market moved and the client's own product changed underneath it. The framework asks one question per study: how many weeks until this evidence needs re-collection? Short half-life evidence (pricing sensitivity, competitor positioning, AI citation share) belongs in quarterly refreshes. Long half-life evidence (category vocabulary, buyer role maps, onboarding friction) can anchor a retainer for a year or more. The practical move is tagging every research asset with a re-collection date at delivery, so the next engagement starts from a scheduled refresh rather than a rebuild. Agencies that publish those dates turn research from a project cost into a recurring line item.