AgentSkillsCN

Startup Failure Patterns

启动失败模式

SKILL.md

startup-failure-patterns

Common startup failure patterns to recognize and avoid.

When to Use

This skill should be used when:

  • Evaluating a new product idea
  • Planning a startup or side project
  • Reviewing business decisions
  • Analyzing why something isn't working
  • Deciding whether to pivot or persevere

The Fatal Patterns

1. Premature Scaling

What it looks like:

  • Hiring before product-market fit
  • Building for 1M users when you have 100
  • Infrastructure over-engineering
  • Multiple product lines before one works

Warning signs:

  • "We need to prepare for scale"
  • Burn rate increasing while revenue flat
  • Team growing faster than customers
  • Building features nobody asked for

Prevention:

  • Do things that don't scale (YC advice)
  • Constrain infrastructure until it hurts
  • One product, one market, one channel

2. Timing Mismatch

Too Early:

  • Market doesn't understand the problem yet
  • Enabling technology isn't mature
  • Regulatory environment hostile
  • Customer behavior hasn't shifted

Too Late:

  • Market saturated with alternatives
  • Incumbents have locked up distribution
  • Margins compressed by competition
  • Customer acquisition costs prohibitive

How to assess:

  • Are people actively searching for solutions?
  • Are adjacent products succeeding?
  • What changed in the last 2 years that makes this possible now?

3. Tech Overkill

What it looks like:

  • Blockchain for a todo app
  • ML when rules would work
  • Microservices for a prototype
  • Custom everything instead of SaaS

Warning signs:

  • "We're building the platform first"
  • Architecture docs longer than customer interviews
  • Hiring specialists before generalists
  • Build time measured in months, not days

Prevention:

  • What's the simplest thing that could work?
  • Could this be a spreadsheet? A Notion doc?
  • Ship something ugly that works

4. Burn Rate Blindness

What it looks like:

  • Spending as if funded, not profitable
  • No clear path to unit economics
  • CAC > LTV (customer acquisition cost exceeds lifetime value)
  • Hoping growth will fix margins

Warning signs:

  • "We'll figure out monetization later"
  • Discounting to hit growth numbers
  • Runway under 12 months without clear path
  • Revenue growth but not profit growth

Prevention:

  • Know your numbers: CAC, LTV, payback period
  • Charge from day one
  • Default alive vs default dead analysis

5. Single-Channel Dependency

What it looks like:

  • All customers from one source (SEO, ads, partnerships)
  • Platform risk (building on Twitter, Facebook, etc.)
  • Whale customer dependency
  • Viral loop that stopped looping

Warning signs:

  • "If X changes, we're in trouble"
  • No organic/word-of-mouth growth
  • Marketing spend only goes up
  • One partnership makes or breaks you

Prevention:

  • Diversify before you need to
  • Build owned audiences (email, community)
  • Track channel health metrics

6. Founder Market Mismatch

What it looks like:

  • Building for a market you don't understand
  • No unfair advantage (domain, network, tech)
  • Competitors have what you lack
  • Learning curve too steep

Warning signs:

  • Constantly surprised by customer behavior
  • Competitors seem to know something you don't
  • Hiring consultants to understand your market
  • "We'll learn as we go" without a plan

Prevention:

  • Build for yourself or someone you deeply understand
  • Identify your unfair advantage before starting
  • Partner with domain experts if lacking

7. The Feature Factory

What it looks like:

  • More features, same revenue
  • Roadmap driven by loudest customers
  • No coherent product vision
  • Competing on features vs. solving problems

Warning signs:

  • "Users want X, Y, and Z"
  • Feature requests outnumber use cases
  • New features don't improve retention
  • Enterprise customization creep

Prevention:

  • Say no by default
  • Focus on jobs-to-be-done
  • Measure feature impact before building more

Recovery Patterns

Recognizing When to Pivot

Pivot signals:

  • Customers use product differently than intended
  • One segment thriving, others flat
  • Side feature getting more traction
  • Market feedback consistently points elsewhere

Pivot vs. persist decision:

  • Is there a kernel of success to build on?
  • Do you still believe in the problem?
  • Is the team motivated to try again?
  • Do you have runway for another attempt?

Graceful Shutdown

When to consider:

  • No kernel of success after honest effort
  • Team burned out
  • Market fundamentally wrong
  • Opportunity cost too high

How to do it well:

  • Be honest with stakeholders early
  • Return remaining capital if possible
  • Help team find new roles
  • Document learnings for the community

Pattern Recognition Checklist

Before committing to an idea, verify:

  • Problem validated: Real people have this problem (not hypothetical)
  • Timing right: Something changed recently that enables this
  • Market exists: People are already paying for inferior solutions
  • You can reach them: Clear path to first 100 customers
  • You have an edge: Unfair advantage (tech, network, insight)
  • Unit economics work: CAC < LTV with reasonable payback
  • Team fit: You understand this market or can learn fast

Resources