💡

InsightHunt

Hunt the Insights

D

Dalton Caldwell

Managing Director & Group Partner

Y Combinator

🎯 Product Strategy (2)🔍 User Research (1)

Key Takeaways

  • 1.The 'Don't Die' Strategy: Most startups fail not because they run out of money, but because founders lose hope and stop trying; survival is an active, irrational choice.
  • 2.Pivot 'Closer to Home': Successful pivots usually move toward the founder's domain expertise rather than chasing a random market trend.
  • 3.The 20-30% Rule: If you audit your calendar and less than 20% of your time is labeled 'Customer Meeting', you are likely hiding behind code or ads.
  • 4.Beware of Tarpit Ideas: Ideas that receive high social validation ('I'd use that!') but have failed repeatedly for decades (e.g., friend coordination apps) are dangerous traps.
  • 5.The 'Hated Incumbent' Heuristic: A reliable way to find B2B ideas is to identify large, private-equity-owned incumbents with low NPS and build modern software to displace them.
  • 6.Ignore Growth Hacking Early: A/B testing and complex analytics are distractions for pre-PMF startups; focus on manual, unscalable user acquisition.
  • 7.The 'Collison Install': Don't stop at the sale; physically or virtually take the keyboard and implement your software for the customer to ensure adoption.

Methodologies(3)

🎯 Product Strategy

A framework for pivoting that argues the best new direction is usually one that leverages the team's specific pre-existing expertise or insights gained during the failure of the first product. It treats pivoting not as starting over, but as returning to a place of strength.

Core Principles

  • 1.Audit for 'Warmth': Choose a new direction where you have unfair domain knowledge or technical advantage, rather than a 'cold' market you think is profitable.
  • 2.Leverage the Struggle: Use the insights gained from the failed product (e.g., internal tools you built to manage the failed product) as the seed for the pivot.
  • 3.Assess Idea Inventory: Only pivot when you have exhausted all genuine growth ideas for the current product; if you still have 6 marketing experiments, do them first.
  • +1 more...

"A good pivot is like going home. It's warmer, it's closer to something that you're an expert at."

#'closer#home'#pivot
View Deep Dive →
🎯 Product Strategy

A diagnostic method to identify 'Tarpit Ideas'—concepts that seem unsolved and receive widespread verbal encouragement ('I would love an app for that!') but have killed startups for decades due to underlying consumer psychology or market structure issues.

Core Principles

  • 1.Check Historical Failures: If people have been trying to solve this since the 90s (e.g., music discovery, nightlife coordination), it is likely a tarpit.
  • 2.Analyze the Validation Type: Be skeptical of ideas where user feedback is overwhelmingly positive in theory ('We should hang out more!') but low in commitment.
  • 3.The 'Hated Incumbent' Test: Instead of a tarpit, look for industries with large incumbents owned by private equity with terrible NPS (e.g., procurement software).
  • +1 more...

"It is only a tarpit if it seems like it's not... You'll get all this positive feedback from the world and people have been starting that startup since the '90s."

#tarpit#identification#protocol
View Deep Dive →
🔍 User Research

A high-touch onboarding tactic named after the Stripe founders, where the product builder physically or virtually takes control of the customer's environment to ensure the product is implemented immediately, rather than waiting for the customer to do it.

Core Principles

  • 1.Don't Accept 'Yes' as Done: A sale is not closed until the software is running in their production environment.
  • 2.Intervene Physically: If possible, go to the customer's office. If remote, ask to share screen and request control of their keyboard.
  • 3.Force the 'Hello World': Do not leave the interaction until the first transaction or action has successfully processed.
  • +1 more...

"They would just install Stripe into the customer's website... They basically would not go away until you finish the implementation."

#'collison#install'#methodology
View Deep Dive →