How I Think
Real problems I've spotted, broken down with behavior-first thinking and Point of Performance analysis.
What is Point of Performance Thinking?
Point of Performance is the exact moment and location where an action needs to happen for it to succeed. It's not about the system as a whole — it's about the precise point where someone needs to make a decision, take an action, or receive information.
The goal: Remove all friction between the thought and the result. Make excellence easy. Every case study below starts by identifying that exact point.
Cartma — Turning a $800M Problem Into a Behavior Design Solution
Retailers lose over $800M a year to shopping cart loss. The industry response has been punishment-based: wheel locks, deposits, threatening signs. But nobody stopped to ask why people are taking them in the first place.
Cart loss isn't one behavior — it's three distinct behaviors from three types of people, each requiring a different design response:
- The Walker — Needs the cart to get groceries home. No car. This is a transportation problem, not a theft problem.
- The Forgetter — Left the cart in the parking lot or nearby. Low effort to return, just needs a nudge.
- The Passive Leaver — Doesn't think about it at all. Cart just... ends up somewhere. Zero malice.
A gamified return system that replaces punishment with design. The app creates different return pathways for each behavior type — unlocking rewards, community credit, and walker-specific checkout flows. The point of performance isn't the parking lot. It's the moment someone decides whether returning the cart is worth their time.
- Human behavior mapping for each user type
- Full app UX with unlock, rewards, and walker checkout flows
- Brand identity system built from scratch
- Point of Performance thinking applied to retail
DoorDash / McDonald's Bottleneck
When McDonald's gets slammed with orders, the drive-thru backs up. DoorDash drivers idle in line. Customers inside wait longer. The bottleneck cascades — one slow point breaks the entire system.
The bottleneck isn't the kitchen — it's the single-channel pickup. Every person (drive-thru customer, DoorDash driver, walk-in) is competing for the same window. The fix needs to happen at the pickup point, not the order point.
External pickup window + app notification system. DoorDash drivers don't sit in the drive-thru line. They park, get notified when the order is ready, walk to a dedicated window, grab and go. It fixes the exact point where the bottleneck happens — separating delivery traffic from customer traffic.
The "Moneyball" Hiring Process
The standard corporate interview process screens out high-value neurodivergent talent by prioritizing social camouflage — like eye contact and small talk — over actual technical ability and output. This broken filter costs companies thousands in turnover and leaves specialized talent on the table.
Treat hiring like the Oakland A's treated baseball in Moneyball. You don't need a perfectly rounded employee who networks well at the water cooler; you need a specialist who delivers massive output. Neurodivergent individuals often possess deep hyper-focus and pattern-recognition skills, but the standard corporate environment actively works against them.
A behavior-first hiring and retention model. Redesign the interview into a skills-based, pressure-free audition rather than a social interrogation. Eliminate the sensory friction in the workspace so the employee can actually perform. The Point of Performance isn't the handshake in the lobby — it's the environment where the actual work happens.
The Venue Rideshare Trap
Every time a concert or game ends at a major venue, the rideshare pickup zone turns into a gridlocked nightmare. Drivers are trapped in a single lane of traffic, passengers are wandering around looking for license plates in the dark, and a process that should take two minutes takes forty-five. It's a completely broken loop that frustrates the driver, the passenger, and city traffic control.
City planners and venue organizers design pickup zones for cars, but the actual bottleneck is matching. When you dump thousands of people into a single geofenced lot, the app's GPS gets confused, and drivers get matched with passengers who are physically at the other end of the line. The point of performance isn't just giving cars a place to park — it's streamlining the exact moment a specific passenger finds a specific car.
A behavior-first staging system. Instead of random matching in a chaotic lot, implement a queue system similar to airport taxi lines but modernized for rideshare. Drivers stage in a holding lot and pull up to numbered stalls. Passengers walk to the corresponding stall number when their app pings. You remove the friction of the "hunt" and turn a 45-minute cluster into a continuous, flowing conveyor belt.