The Skunk Works Method
Small elite teams, total autonomy, rapid delivery.
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 7
- Confidence
- —
Developed by Lockheed's Kelly Johnson during WWII and formalized into 14 rules, the Skunk Works method is a management approach for building breakthrough technology far faster and cheaper than conventional corporate R&D. It combines a tiny hand-picked team (10-25 engineers plus a similarly small shop crew), a single empowered program manager with end-to-end authority, physical proximity between designers and the production floor, compensation tied to responsibility rather than headcount, and radical secrecy with minimal management interference. Born from the existential urgency of WWII and the Cold War, it produced the P-80 Shooting Star, U-2, SR-71 Blackbird, and F-117 Nighthawk. The model later diffused into Silicon Valley management culture but was largely abandoned in modern defense procurement.
Origin
In late 1944, with German jet fighters appearing over Europe, the U.S. government gave Lockheed 180 days to build a prototype jet fighter. Kelly Johnson hand-picked 23 engineers and 30 shop workers and housed them in a rented circus tent next to a plastics factory in Burbank, whose smell inspired the 'Skunk Works' nickname (borrowed from the Li'l Abner comic strip). The team delivered the P-80 prototype, the Lulu Belle, in 143 days.
Core principles
- 01Rapid delivery of a superior product is the only thing that matters
- 02Keep the team ruthlessly small and co-located
- 03Give the program manager total, unchallenged authority
- 04Keep designers a stone's throw from the shop floor
- 05Reward responsibility and scope, not number of people supervised
- 06Keep overhead and cost low so management doesn't need heavy oversight
- 07Let an existential, high-stakes threat supply the motivation
How to run it
- 1
Restrict the team to a small, elite group
Cap the team at roughly 10-25 of the best engineers/designers, plus a similarly lean shop crew, rather than staffing a 'normal' large organization.
Pro tip Keep everyone physically together so relationships, collaboration, and full-stack thinking happen naturally.
- 2
Delegate complete control to one program manager
One person -- from engineering decisions to negotiating with the government customer -- must own the entire program end to end, with no committee or middle-management layer diluting authority.
Watch out Without single-point accountability, speed and clarity collapse into consensus-seeking.
- 3
Put designers next to the production floor
Physically locate engineers and designers a 'stone's throw' from the machinists building the product so design flaws get caught and fixed immediately instead of through a slow game of telephone.
Pro tip Let designers watch parts being built in real time, not just review diagrams.
Watch out If the design team is not within walking distance of the shop, the feedback loop breaks down.
- 4
Reward responsibility, not headcount
Base pay increases on the scope of responsibility a person takes on, not on how many people they supervise -- the opposite of typical corporate empire-building incentives.
Watch out Most executives default to rewarding headcount growth; this must be actively resisted.
- 5
Minimize overhead and cost exposure
Fund the program small enough that it doesn't require heavy shareholder reporting or scrutiny, keeping financial risk to the parent company low even on ambitious projects.
- 6
Operate under strict secrecy with minimal outside interference
Get the sponsoring agency to state requirements clearly, provide resources, and then 'get out of the way' rather than micromanaging the project.
- 7
Anchor the mission to a real, urgent threat
Use a genuine existential stake (e.g., losing a war) as the team's motivation rather than artificial incentives -- this was assumed, not written into any rule, but essential to why the model worked.
Watch out Without a real threat-based need, teams can drift into engineering for its own sake and lose commercial or strategic relevance.
In the wild
Given 180 days and a spec to beat Germany's 550mph jet, Kelly Johnson's 23 engineers and 30 shop workers built America's first jet fighter prototype, the Lulu Belle, from a rented circus tent in a Burbank parking lot.
→ Delivered in 143 days, becoming the P-80 Shooting Star, the U.S.'s first operational jet fighter.
The CIA needed a plane that could fly above Soviet radar and fighter ceilings; Skunk Works had to invent new jet fuel, a pressure suit, and a 70,000-foot airframe from scratch.
→ Delivered in about 18 months for $3.5 million, giving the U.S. unprecedented intelligence on Soviet capabilities for five years.
To outrun any missile, Skunk Works built a titanium airframe with spike-inlet 'superchargers' to sustain Mach 3+ flight, solving fuel leakage, extreme heat, and navigation-by-stars problems along the way.
→ Never successfully shot down despite roughly 4,000 missiles fired at it over its operational life.
Common mistakes
Ignoring the threat-based/market need
Kelly Johnson tried to give the Air Force the 'best' fighter rather than the cheap, mass-producible plane it actually wanted for the F-16 program, and Skunk Works lost the bid to General Dynamics.
Assuming engineering excellence alone guarantees success
The 'only rapid delivery of superior products matters' ethos worked because of WWII/Cold War urgency; applied to commercial ventures without that urgency (e.g., Lockheed's L-1011 airliner), it led to a $2.5 billion loss.
Letting bureaucracy replace the model
Post-Cold War procurement politics forced manufacturing to be spread across 46 states and thousands of subcontractors for jobs and votes, driving costs up while abandoning the small-team, high-autonomy approach that made Skunk Works work.
Is it for you?
Best for
Small, autonomous engineering teams tackling an urgent, high-stakes technical challenge with a single empowered leader and minimal bureaucratic interference.
Not ideal for
Large, politically distributed programs where funding, jobs, and manufacturing must be spread across many stakeholders/regions (e.g. modern multi-state defense contracts).
From the transcript
“the number of people having any connection with the project must be restricted in an almost vicious manner use a small number of good people…”
“the Skunk Works manager must be delegated practically complete control of his program in all aspects”
“in the main plant they give raises on the basis of the more people supervised I give raises to the guy who supervises the least…”
From the episode
Lockheed Martin