ExpertsFMThe airline CEO putting Starlink on every plane
Cheeky Pint · October 27, 2026
5 Selected Lessons:
heuristic · 2 framework · 2 case · 1
Invest in modernization with no ROI case to unlock future optionality
Source
Cheeky Pint — The airline CEO putting Starlink on every plane (October 27, 2026) · at 6:35 ▸ · at 10:28 ▸
When this applies
when considering infrastructure investment whose value is optionality for future speed rather than a calculable near-term return
The takeaway
Some investments should be justified by the option value of future agility, not by a projectable ROI — and that's precisely why most companies underinvest in them.
Why it works
Modern infrastructure removes constraints on future speed of execution; the payoff shows up later as the ability to ship things competitors structurally cannot, which doesn't fit a standard ROI calculation at the time of investment.
One level up
The hardest-to-justify investments (no clear ROI) are sometimes the ones that create the largest future competitive gaps, because that very difficulty keeps competitors from making them.
Where it stops holding
This stops working when the organization cannot sustain the spend long enough to realize the optionality, or when the infrastructure becomes obsolete before being leveraged.
Detail
procedure:
- Identify infrastructure that structurally limits future speed (e.g., legacy mainframes)
- Fund the migration even without calculable ROI, anticipating that it enables future capabilities
- Use downturns/crises as periods competitors won't match, since they're cutting while you invest
rationale: No ROI, no payback at the time — the value is the capacity to move fast and do things other airlines can't because they lack the tech stack.
In their words
There was no ROI, there was no payback, which is why companies don't do it. Like it's really hard to do the ROI on something like that. But what it does is it enables you to move fast and do things.
Hub concentration over duplicative point-to-point flying
Source
Cheeky Pint — The airline CEO putting Starlink on every plane (October 27, 2026) · at 25:32 ▸ · at 53:09 ▸
When this applies
when allocating a scarce, shared capacity resource between many small redundant uses versus fewer high-value consolidated uses
The takeaway
When two players duplicate feeder/connector capacity into a shared constrained resource, consolidating that feeder capacity under one network increases the system's output of higher-value uses.
Why it works
Each unit of scarce slot capacity used for redundant short connector flights (duplicated by a competitor) is a unit not used for unique long-haul routes; consolidating connector flying onto one carrier at a hub frees capacity for higher-value routes and ultimately benefits the whole city's economic activity.
One level up
Scarce shared infrastructure performs better when redundant low-value uses are consolidated, freeing capacity for differentiated high-value uses.
Where it stops holding
This stops working when the 'duplicative' capacity is actually providing valuable competitive choice/pricing pressure that customers value more than the gained high-value routes.
Detail
procedure:
- Identify catchment-area flights that exist only to feed a hub, not serve point-to-point demand
- Recognize that duplicating these flights across two competing hub carriers wastes slot capacity
- Consolidating connector flying to one carrier frees slots for more long-haul/international flights
- Result: more total economic activity for the city despite fewer total flights
rationale: Atlanta and Dallas, smaller cities than Chicago, have more international service and drive more economic activity because they're single-hub cities that don't waste slots on duplicative catchment flying.
In their words
Single hub airports drive more economic activity. This is the reason, because they don't have to waste several 100 of their flight possibilities on catchment area
Ignore surveys for future decisions, use them only for retrospective feedback
Source
Cheeky Pint — The airline CEO putting Starlink on every plane (October 27, 2026) · at 2:08 ▸
When this applies
when trying to predict future behavior or demand for something that doesn't yet exist
The takeaway
Self-reported predictions about hypothetical future states are unreliable; direct observation of behavior after exposure is reliable.
Why it works
Surveys capture people's stated feelings under current conditions (fear, unfamiliarity with a not-yet-experienced thing) and those stated preferences don't predict behavior once circumstances or the product itself changes.
One level up
Trust revealed behavior over stated intention, especially for anything novel or conditioned on a temporary emotional state.
Where it stops holding
This stops working when the thing being tested is well-understood and low-novelty, where stated preference actually does correlate with future choice.
Detail
rule: Use surveys only to learn how people felt about something already done; never use them to decide what to do next — just build it, test it live, and iterate.
threshold: If people say 'no' now but the underlying condition (e.g. fear of COVID) is temporary, disregard the survey.
rationale: People surveyed during COVID said they'd never travel again, but once the crisis passed, demand returned — the stated answer was conditioned on a transient state, not a stable preference.
In their words
I like surveys for how did people feel about stuff we've already done. I ignore surveys for what we could do in the future.
Kirby believes airlines should stick to their strategic identity rather than chase a competitor's model
Source
Cheeky Pint — The airline CEO putting Starlink on every plane (October 27, 2026) · at 57:00 ▸ · at 59:21 ▸
When this applies
when a company with one structural advantage is tempted to copy a competitor's different structural model instead of doubling down on its own
The takeaway
Know your structural strengths and limitations, and resist the temptation to copy a competitor's different model — discipline means accepting what you may not be able to match.
Why it works
Structural advantages (cost structure, network depth, hub economics) are not easily transplantable; attempting to mimic a fundamentally different business model usually produces a worse version of both models instead of a hybrid winner.
Where it stops holding
This stops working when the company's original structural advantage has eroded and genuinely needs to pivot to a new model to survive.
Detail
rule: In Kirby's view, don't try to out-compete a specialist competitor by copying their model — he's observed that every major airline that tried an internal low-cost sub-brand to beat Southwest failed at it, and he doubts Southwest's attempt to be more like full-service carriers will work either.
rationale: 'Know who you are, know what your strengths are, and stick to your knitting.' Allegiant thrives in its niche (small-city-to-Vegas) precisely because it doesn't try to be United, and United doesn't try to be Allegiant.
In their words
What your strengths are, and stick to your knitting. Companies are not good at trying to be something they're not.
United's contrarian COVID fleet order
Source
Cheeky Pint — The airline CEO putting Starlink on every plane (October 27, 2026) · at 0:00 ▸ · at 2:08 ▸
When this applies
when a crisis causes near-universal consensus to extrapolate the bad present into the future, and competitors are all retrenching in the same direction
The takeaway
When the whole field is moving one direction based on a shared, falsifiable assumption, betting against that assumption asymmetrically rewards being merely 'not totally wrong,' not being exactly right.
Why it works
When nearly everyone is de-risking the same way, taking the opposite position costs little if wrong (since you're 'not totally wrong' relative to the field) but pays enormously if the consensus reverses, because pricing/terms and competitive position are both favorable.
One level up
Look for situations where the downside of being wrong is capped by the herd's own extremity, and the upside of being right is uncapped.
Where it stops holding
This stops working when the consensus view turns out to be correct and the underlying asset/commitment can't be unwound or absorbed if the bet fails.
Detail
situation: In 2020 every airline expert predicted permanent 50% demand destruction; every global long-haul airline except United retired part of its fleet.
decision: United did the opposite: placed the biggest aircraft order in industry history (200 wide-bodies, several hundred narrow-bodies, listing near $400M each).
outcome: Kirby told the board 'we don't have to be right, we just have to be not totally wrong' because the field was so one-sided; the bet paid off as travel demand recovered.
lesson: Find situations where the odds are stacked in your favor because everyone else is wrong in the same direction.
In their words
this is one of those cases because everyone else is shrinking, we don't have to be right. We just have to be not totally wrong