Building Uber's Seattle Driver Onboarding Operation

The Ask

In early 2015, Uber was expanding into Seattle and needed to onboard drivers fast. Their immediate problem was vehicle inspections. Every new driver needed their car inspected before they could start driving on the platform.

They came to me with a simple request: rent our shop space at night so they could run inspections.

They were already doing this at other local shops around Seattle. The setup was the same everywhere: rent a facility after hours, bring in their own people, and try to push drivers through. They were paying $4,500–6,000 per night and getting 10–15 inspections done. Roughly $500 per driver.

That's what they asked for. It's not what they needed.

The Real Problem

I spent time understanding what was actually happening during these inspection nights, and the problems were structural:

1. No standardized process. There was no checklist, no defined sequence, no consistency in what got inspected or how.
2. No trained inspectors. The people running inspections didn't have a repeatable method. Every inspection was improvised.
3. Unclear compliance requirements. The transportation department had specific regulations for vehicle inspections, and it wasn't clear that those were being met consistently.
4. No path to scale. Everything about the approach was ad hoc. You couldn't make it faster by throwing more people at it because there was nothing to teach them.

Uber didn't have a space problem. They had a process problem.

What I Built

I converted the space rental into a consulting engagement and took ownership of the entire inspection operation for Uber's Seattle market.

The Inspection System

I started with the regulations. I clarified exactly what the transportation department required for a vehicle to pass inspection, then built a checklist inspection form that standardized every step. No ambiguity, no judgment calls. Every inspector followed the same process.

The Staffing Model

I designed the operation around bays. Each bay had one inspection technician who performed the inspections. A master technician oversaw the operation and certified the results. This separation meant we could scale linearly: add a bay, add a tech, increase throughput.

I hired the staff, trained them on the process, and created operations procedures that made the training repeatable.

The Scheduling System

Instead of having drivers show up to a chaotic nighttime session, I built a scheduling system that controlled flow. Drivers booked time slots. The pipeline was predictable. No bottlenecks at check-in, no idle bays.

Built for Transfer

Every decision was made with scale in mind. The inspection checklist was designed around standard taxi inspection criteria, the major sections of a vehicle that any jurisdiction would require. Different cities might have minor variations, but the foundation was universal.

Nothing about the system was specific to my location. The only equipment required was standard auto shop gear that any facility would already have. The operations procedures, the staffing model, the scheduling approach: all of it could be picked up and deployed at any shop with bays and lifts.

I packaged the entire operation into a handoff-ready system: the inspection process, the training procedures, and the operational playbook. After I stepped away, Uber continued running inspections in Seattle exactly the way I'd set them up, just at other locations.

The Results

The numbers told the story:

• Throughput: From 10–15 drivers per night to 10 drivers per hour per technician
• Peak volume: ~600 drivers per week
• Total onboarded: ~2,000 drivers at our location
• Cost: ~$250 per hour for 30 inspections, compared to $1,500 per hour for 3 inspections under the old approach

That's a cost reduction from roughly $500 per driver to roughly $8 per driver.

I managed the operation from January through June 2015, then handed off a system that was ready to run without me.

What This Illustrates

1. The Stated Problem Is Rarely the Real Problem

Uber asked to rent space. The actual constraint was that they had no process, no trained people, and no system. If I'd just rented them the space, they would have gotten the same results they were getting everywhere else.

2. Process Design Is the Leverage Point

The facility didn't change. The regulations didn't change. What changed was that someone built a defined, repeatable process and trained people to execute it. That's what created a 60x throughput improvement.

3. Build for Scale on Day One

I could have optimized for my location. Instead, I designed around bays, an abstraction that works anywhere. That decision is what made the system transferable rather than location-dependent.

---

Have an operation that needs a system behind it? Let's talk.