Nudge unit
Also known as: Behavioural insights team
A government or organizational team that designs, applies, and rigorously tests nudges.
What it means
A nudge unit is a dedicated team that applies behavioral science to policy and operations, typically by designing low-cost interventions grounded in behavioral insights and testing them with large field experiments before they are scaled. The defining feature is the marriage of behavioral design with rigorous evaluation: rather than assuming a nudge will work, these units run randomized controlled trials, often across hundreds of thousands of people, to measure real effects on real behavior and to discard interventions that do not deliver. The approach was pioneered by the UK's Behavioural Insights Team, established in 2010 and originally housed within government, whose visible successes — particularly in tax compliance, where revised letters using social norms and simplification raised payment rates — sparked a global wave of similar units in governments, agencies, and companies. Their work institutionalizes the EAST principles and the broader nudge toolkit while holding it to an experimental evidence standard. Critics question the modest average effect sizes of many nudges, their durability, and the ethics and transparency of government behavior change. It matters because nudge units brought experimental, evidence-based behavioral design into the machinery of public policy at scale.
Examples
A tax authority's nudge unit RCT-tests different letter wordings across hundreds of thousands of taxpayers before rolling out the version that demonstrably lifts on-time payment.
England's Chief Medical Officer wrote to GPs telling them they prescribed more antibiotics than most local peers — and the letter was tested in a randomised trial rather than simply sent.
A utility's in-house team randomly assigns two bill redesigns across customers and shelves the version that dazzled in the workshop but moved nobody; the trial decides, not taste.
First described in UK Behavioural Insights Team (2010).