ExpertsFMLockheed Martin
Acquired · May 30, 2023
5 Selected Lessons:
heuristic · 2 case · 2 framework · 1
Kelly Johnson's Skunk Works Rules
Source
When this applies
when building a small high-performance team inside a larger organization to deliver something extraordinary fast
The takeaway
Elite rapid-delivery teams work by minimizing headcount, maximizing proximity between decision-makers and the people executing, giving the leader full authority, and rewarding efficiency rather than organizational size.
Why it works
Restricting headcount forces every member to be a high-leverage generalist, co-location collapses communication lag between design and execution, single-point accountability removes committee drag, and compensating for output rather than headcount removes the incentive to build empires.
One level up
Structure follows the mission's true constraint (speed/quality), not the organization's normal bureaucratic defaults.
Where it stops holding
This stops working when the mission doesn't require extreme speed/innovation (favoring instead broad coordination or risk-spreading) or when the small team lacks intrinsic, mission-driven motivation to compensate for lower resources and oversight.
Detail
procedure:
- Restrict the number of people connected to the project 'in an almost vicious manner' — 10-25% of normal staffing.
- Co-locate engineers and designers a stone's throw from the production floor so changes happen immediately.
- Delegate the program manager practically complete control over all aspects of the program, including external relationships.
- Reward performance based on output/responsibility taken, not on number of people supervised.
benchmarks: {"peak_team_size":"~50 designers/engineers and ~100 machinists at Skunk Works' most productive"}
rationale: A small number of highly capable, cross-disciplinary people collaborating tightly, led by one empowered decision-maker, outperforms large bureaucratic teams on speed and quality.
In their words
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, 10% to 25% compared to the so-called normal systems.
Don't build what doesn't make the plane fly faster
Source
Acquired — Lockheed Martin (May 30, 2023) · at 20:06 ▸
When this applies
when operating under an extreme deadline or resource constraint for a complex product
The takeaway
When time or capital is the binding constraint, outsource anything that doesn't directly create your differentiation, even if it means depending on an outside (even allied/competitor) supplier.
Why it works
Vertically integrating into a non-core, already-commoditized capability consumes scarce time/capital that could go to the actual differentiator, and a capable outside partner can supply it faster.
One level up
Ruthlessly define your core competency and let everything else be bought, not built.
Where it stops holding
This stops working when the 'non-core' capability is actually strategically critical or when no reliable external supplier exists at the needed quality or speed.
Detail
rule: If you only have a short runway, don't become a manufacturer of a component that isn't your core differentiator — buy it from whoever already does it well.
threshold: 180-day build deadline left zero slack to develop an in-house jet engine.
rationale: Lockheed and essentially the whole aerospace industry has never made its own engines, outsourcing to specialists like Rolls-Royce and GE even on commercial programs.
In their words
this 180-day thing is a very interesting constraint placed on them, and it means that they immediately need to go to an acquired axiom that we've talked about forever. don't do something that's not your core competency.
Pay for responsibility taken, not headcount managed
Source
Acquired — Lockheed Martin (May 30, 2023) · at 28:14 ▸
When this applies
when designing incentive/compensation structures for managers in any organization
The takeaway
Reward managers for the results and responsibility they take on, not for the size of the organization they control, to avoid incentivizing bureaucratic growth.
Why it works
Tying pay to headcount supervised incentivizes empire-building and bloat; tying pay to output and responsibility incentivizes managers to do more with less.
One level up
Incentive design, more than stated values, determines organizational behavior.
Where it stops holding
This stops working when the nature of the work genuinely requires scaling headcount to produce more output, making headcount growth itself a legitimate signal of value created.
Detail
rule: Give raises to the person who supervises the least, because that means he's doing more and taking more responsibility.
threshold: Applied specifically to a small-team, high-leverage R&D operation.
rationale: Most executives are empire builders; this policy counteracts that tendency and mirrors capitalist efficiency — achieving the most with the least rather than a fixed amount with fixed margin.
In their words
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. That means he's doing more and taking more responsibility.
LMSC's margin advantage over Skunk Works via computing/software vs. hardware manufacturing
Source
Acquired — Lockheed Martin (May 30, 2023) · at 1:47:14 ▸
When this applies
when comparing two business lines within a company that differ mainly in whether value creation centers on computing/software vs. physical manufacturing
The takeaway
Within the same company and customer relationship, the business line whose core value is computing/software will tend to be structurally more profitable than the one whose core value is physical manufacturing.
Why it works
Value embedded in computing, guidance, and software carries inherently better margins than physical airframe manufacturing, so a division centered on the former outearns one centered on the latter even within the same parent company and customer base.
One level up
Margin structure follows the nature of the underlying value creation (information vs. atoms), not just the prestige or scale of the program.
Where it stops holding
This stops working when the computing/software is commoditized or when external forces (e.g., cost-plus contracting that caps margins) neutralize the usual software margin advantage.
Detail
situation: From 1960-1972, Lockheed as a whole did $26B revenue and only $255M profit; LMSC accounted for over a third of revenue but 128% of profit (meaning the rest of Lockheed lost money).
decision: N/A — observed structural pattern across divisions.
outcome: From 1983-1992, LMSC's share grew to 46% of revenue and 72% of profit.
lesson: A computing/software-centric division embedded within an airplane manufacturer can carry the whole company's profitability for decades.
In their words
During the 12-year period from 1960, when Corona first launched to 1972, Lockheed as a whole did 26 billion in revenue over that 12-year period and just 255 million in total profit.
The Last Supper: government-mandated industry consolidation
Source
When this applies
when a dominant buyer in a concentrated industry explicitly instructs suppliers to consolidate because the market can no longer support them all
The takeaway
A monopsony buyer facing structurally shrinking demand may rationally sacrifice competitive dynamics and directly orchestrate supplier consolidation to preserve critical capability, rather than risk losing it to uncoordinated market exit.
Why it works
When a single dominant buyer controls nearly all demand and strategic capability preservation matters more than market competition, that buyer can directly engineer supplier consolidation rather than let market forces potentially eliminate capability entirely.
One level up
Market mechanisms are sometimes deliberately overridden by a powerful stakeholder when the stakes (losing a strategic capability) outweigh the benefits of competition.
Where it stops holding
This stops working when no single buyer has enough market power to coordinate consolidation, or when preserving competition matters more than preserving capability.
Detail
situation: Post-Cold War defense budgets were shrinking massively, threatening to put multiple major contractors out of business and erode America's industrial base/capability.
decision: Perry explicitly told defense CEOs the government could no longer 'feed all the metaphorical mouths' and instructed them to merge.
outcome: This directly triggered a wave of mergers (Lockheed+Martin Marietta, Boeing+McDonnell Douglas, etc.), consolidating from many contractors to five primes, while a proposed Lockheed Martin-Northrop Grumman merger was blocked by the DOJ to preserve minimal competition.
lesson: A government acting as dominant customer can directly dictate industry structure when it judges market-driven outcomes would destroy a capability it deems essential, while still trying to preserve a minimal competitive floor.
In their words
In July of 1993, the then Deputy Defense Secretary William Perry calls the CEOs of all the major crime defense contractors to a dinner in Washington, at which he explicitly tells them defense spending is going to shrink massively.