Federal agency
Grant program was rewarding paperwork, not outcomes.
Restructured the reporting incentive so applicants competed on measurable impact, not narrative length.
For Government & Policy
Map the unintended consequences before they ship.
For federal, state, municipal, and multilateral teams designing programs, regulations, or policy. We red-team the second-order behavior your impact model is missing — and rewrite the rule so the right behavior is the easy one.
Why this work
Every well-intentioned rule is also a behavior-design system.
Goodhart's Law isn't an academic punchline. It's the operating reality of every program with a target attached to it. The moment a metric becomes a payment, an audit trigger, or a press number, the population in front of you starts optimizing it — often in ways the original policy memo never imagined.
We bring a structured way to anticipate that drift. Every program is read across the twelve dimensions of the Human Behavior Taxonomy™ and pressure-tested against a catalog of 200+ documented perverse incentives from public-sector deployments worldwide.
Who it's for
- • Policy teams designing or revising a program with measurable targets.
- • Regulators worried about the behavior a new rule will provoke.
- • Public-sector innovation units running pilots and RCTs.
- • Multilaterals and NGOs designing cross-border programs.
- • Elected officials and chiefs of staff stress-testing a flagship initiative.
What you get
- • A second-order consequence map for the proposed policy or program.
- • A redesigned incentive structure with the perverse loops named and removed.
- • An evaluation plan instrumented for behavior, not just outputs.
- • Stakeholder-ready briefing materials, written for the room you'll defend it in.
- • Optional closed-door working session with your stakeholders or coalition.
Recent work · anonymized
Policies we've quietly redesigned.
State workforce board
Job-training metric was creating ghost placements.
Reweighted reimbursement around 90-day retention. Placement volume held; ghost-rate collapsed.
Municipal · public safety
Patrol allocation was being gamed by call-volume targets.
Replaced the target with a community-outcome composite. Calls dropped; trust scores rose.
Multilateral · climate
Carbon offset program was paying for what would happen anyway.
Re-tied issuance to verified additionality. Issuance volume fell; integrity premium rose.
State · housing
Affordable-housing voucher utilization was stuck.
Diagnosed the landlord-side disincentive; redesigned the carry, lifted lease-up rates two quarters.
City · transit
On-time rate was being achieved by skipping stops.
Reweighted to a rider-experience composite. On-time held; complaints fell.
ROI
3x floor, guaranteed.
We have seen up to 200x on policy redesigns.
2026 Roster
Only 10 engagements.
20 in 2027 across business, government, and orgs.
Engagement
$100k retainer · $50k sprint.
Annual policy advisory, or a two-week redesign sprint.
Bring us the policy.
Tell us the program, the population, the metric, and what could go wrong.
Start an inquiry →