Stamina as the Kill Signal
Projects die of exhausted stamina long before they die of exhausted money — so watch the team, not the runway.
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 4
- Confidence
- 88%
The standard question 'how long do I persist on something that isn't working?' has no clean metric answer in high-ambiguity zero-to-one work, because the signal you'd measure against doesn't exist yet. Lovinsky's answer is to instrument the team's stamina instead. Ambiguous projects require an ungodly level of faith and endurance; the real failure mode is not being wrong, it's the team no longer bringing its best self. When the steam runs out, that is the kill signal — and it almost always arrives before the money does.
Origin
Noam Lovinsky, drawing on YouTube, Thumbtack and the Facebook NPE incubator. The 'running uphill and chewing glass' line is credited to Thumbtack CEO Marco Zappacosta.
Core principles
- 01What kills most projects and early companies is stamina, not capital or market.
- 02High-ambiguity work has no reliable external signal, so it runs on faith — which is a finite resource.
- 03Some suffering is the job; resilience is a muscle to be deliberately practiced.
- 04A depleted team is not a signal the idea was wrong — sometimes it just needs a change of scene.
How to run it
- 1
Accept the glass-chewing baseline
Establish upfront that the work will feel like running uphill and chewing glass, and that this is expected rather than diagnostic. Teams that read normal difficulty as evidence of failure quit at the wrong moment.
Pro tip Name the discomfort explicitly to the team ('we'll take our medicine') so nobody mistakes hardship for a verdict.
- 2
Deliberately build resilience while the tank is full
Treat stamina as a trainable capacity, not a fixed one. The teams that survive long ambiguity are the ones that practiced enduring it before they needed to.
- 3
Monitor stamina, not just milestones
Watch for the qualitative shift: are people still bringing their best selves, or have they been hitting their head against the wall so long that the inertia and quality have visibly changed? Lovinsky describes this as an entirely different game between a motivated team and a down-and-out one.
Pro tip Compare the current team against how they looked when excited about the same problem — the delta in energy is the metric, not the absolute level.
Watch out This is a subjective read and can be gamed by a team afraid of losing headcount. Look at output quality and inertia, not stated morale.
- 4
When steam runs out, change the scene rather than grind
Once the stamina is gone, more time on the same problem produces worse work, not more signal. Move the team — a change of scene or pace often lands them somewhere much better than the grind would have.
Pro tip A stamina-driven kill is not a judgment about the person or the idea, and framing it that way preserves both.
Watch out In the startup case this moment arrives well before you run out of money, so waiting for the bank balance to make the decision for you means burning capital on work nobody has the faith to do well.
In the wild
Thumbtack CEO Marco Zappacosta used to describe the work as running uphill and chewing glass — and the team embraced it as a badge rather than a warning sign, deliberately practicing resilience while rebuilding the product, the monetization model and the growth engine simultaneously under negative growth.
→ The team held its stamina through the rebuild long enough to turn the business around — the 'prettiest smile graph' a Sequoia board member had seen.
Lovinsky contrasts two teams: one that is motivated and excited, where the inertia and quality of everything is visibly a different game, versus one that has been hitting its head against the wall for a long time and is down and out.
→ Rather than being pushed to persist, the depleted team is given a change of scene or pace — and often lands in a much better situation.
Common mistakes
Using runway as the kill signal
Money is a lagging indicator. Stamina runs out first, and every dollar spent after that point buys work from a team no longer capable of finding the faint signal the project needs.
Quitting the moment the work gets painful
The difficulty is the baseline, not the verdict. Ambiguous zero-to-one work requires an ungodly level of faith by design; a team that treats normal hardship as evidence of a bad idea will never finish anything.
Reading a stamina failure as an idea failure
The two are different. Killing the idea when the team is simply exhausted throws away a possibly-good bet, and moving a team without acknowledging the ambiguity was the cause teaches them nothing.
Is it for you?
Best for
Founders and product leaders running zero-to-one or high-ambiguity projects who need a decision rule for persistence when no metric will settle it.
Not ideal for
Well-instrumented, low-ambiguity work where real signal exists and the persist/kill decision should be made on data, not on team energy.
From the transcript
“that what kills most projects most early companies is is stamina.”
“they require like just a in in ungodly level of faith and”
“sometimes they just need like a change of scene a change change of pace”
“so my my honest answer is is yeah it's it's the when do you run out of steam is usually the question”
From the episode
The happiness and pain of product management
Noam Lovinsky (Grammarly, Facebook, YouTube, Thumbtack)