Build the Airplane While Flying
Launch the safe useful version, learn from users, and keep upgrading
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 97%
Set a hard floor for safety and core usefulness, then stop treating every possible improvement as a prerequisite. Release the smallest version that clears that floor so it can generate customer behaviour, feedback, and potentially revenue. Compare what users do with what the team assumed, select the most valuable correction, and ship another version. The mechanism imports the technology industry's upgrade mindset into slower-moving categories: a product is not a monument that stays unchanged until sales decline, but a sequence of better versions. This does not excuse recklessness. Goldin separates safe production from optional refinements. Its purpose is to learn from reality sooner while operators continue solving improvements that would otherwise delay all market contact.
Origin
Goldin carried an iterative technology mindset from AOL into beverages. Rather than waiting to understand every part of an unfamiliar industry, she launched a safe product, asked customers questions, read their emails, and treated packaging and formulation as improvable versions.
Core principles
- 01There is rarely a perfect time to begin
- 02Safety and core usefulness are non-negotiable
- 03A shipped version creates revenue and learning
- 04Every version can improve after customer contact
How to run it
- 1
Set the non-negotiable floor
Define the safety, legal, and core-use requirements the first version must meet.
Pro tip Write the floor as pass-or-fail conditions.
Watch out Iteration is never permission to expose customers to avoidable harm.
- 2
Park later-version work
Separate enhancements from requirements needed to solve the basic job. Keep a backlog without blocking launch.
Pro tip Ask whether an item prevents use or merely improves experience.
- 3
Release to reality
Put the first safe useful version before real customers. Seek behaviour and revenue, not only opinions.
Pro tip Start with a bounded channel where feedback is observable.
Watch out Do not scale faster than your ability to support the product.
- 4
Ask and observe
Read messages, ask direct questions, and watch where assumptions diverge from use.
Pro tip Inexperience encourages basic questions incumbents stopped asking.
Watch out Do not treat one loud preference as the whole market.
- 5
Upgrade deliberately
Choose the highest-value evidence-backed improvement and release the next version. Repeat before decline forces change.
Pro tip Describe improvements as normal progress, not failure.
In the wild
Goldin did not know beverage distribution when Hint entered Whole Foods. She protected safety, shipped a working product, asked consumers questions, read emails, and stayed willing to change presentation as evidence arrived.
→ The company learned in-market while improving a product unfamiliar to buyers and consumers.
A software team has secure account creation and one reliable workflow but twelve planned customisations. It releases to twenty customers, watches onboarding abandonment, and prioritises the explanation blocking activation.
→ The next version fixes an observed barrier without months of untested enhancements.
Common mistakes
Confusing iteration with recklessness
A fast release must satisfy safety, legal, and core-use requirements.
Waiting for the final version
Waiting blocks the evidence needed to decide which improvements matter.
Changing only after decline
Normalise evidence-led upgrades instead of reserving change for emergencies.
Is it for you?
Best for
It is best for products that can launch safely in stages and improve through real customer feedback.
Not ideal for
It is not ideal where an incomplete first version could cause serious harm or violate mandatory standards.
From the transcript
“there's never a good time but that doesn't mean that you can't do something”
“i learned in the tech industry to actually get something out the door because we want to not only get it out the door and…”
“yeah it's pretty good but we could you know we can always do better”
From the episode
Kara Goldin, CEO of Hint: Building an Empire
Kara Goldin