🧭

Idea Maze

How to explore the space of ideas and find unique opportunities by studying the history of similar products

10 minFramework

Idea Maze is a concept proposed by Balaji Srinivasan (former a16z partner). The core idea: good startup ideas are not eureka moments but the result of deeply understanding all paths already taken in a space.

Key Metaphor

Imagine your business idea as the entrance to a maze. At every turn, you make decisions: which business model to choose, which market to enter, how to acquire users. The best founders know this maze better than anyone - they've studied all the dead ends and know where the exits are.

Why Understanding the Maze Matters

Most "new" ideas have already been tried. If you don't know the history of your space, you risk:

  • Repeating mistakes that could have been avoided
  • Not understanding why previous attempts failed
  • Missing changes that make an old idea viable now
  • Overlooking a unique approach nobody has tried

How to Explore the Idea Maze

1. Study the History of Similar Products

Find every company that tried to solve a similar problem. Not just successful ones - failed ones are especially important to study.

Research Questions
  • What companies attempted this in the last 10-20 years?
  • Why did some succeed while others failed?
  • What has changed since then (technology, market, regulations)?
  • What are competitors doing now and what are they missing?

2. Analyze the Types of Failure

Failures come in different types, and understanding the difference is crucial:

  • Timing: The idea was good, but the market or technology wasn't ready
  • Execution: The idea was right, but the implementation was weak
  • Market: The market simply doesn't exist (the problem wasn't painful enough)
  • Business Model: Users liked the product, but the economics didn't work
Example: Webvan vs Instacart

Webvan (1999) spent $1B on grocery delivery infrastructure and went bankrupt. Instacart (2012) used existing stores and gig economy couriers. Same idea, but right timing + different execution approach.

3. Find What Has Changed

If an idea failed before, what has changed now? Look for:

  • Technology shifts: New technologies (AI, mobile, cloud) that make the idea feasible, cheap, or convenient now
  • Behavioral changes: Shifts in people's habits (e.g., pandemic changed attitudes toward remote work)
  • Regulatory changes: New laws that opened or closed markets
  • Economic factors: Significantly reduced implementation costs or changed funding availability

4. Formulate Your Unique Insight

After research, you should have a non-obvious perspective on the problem - something you understand better than others thanks to your research.

Peter Thiel's Test

"What important truth do very few people agree with you on?" Your insight should be counterintuitive for most but obvious to you thanks to your deep understanding of the maze.

Research Sources

  1. Crunchbase / PitchBook - funding history and failures
  2. Product Hunt - what products launched and how they were received
  3. Hacker News - discussions and startup post-mortems
  4. Reddit / specialized forums - real user pain points
  5. Founder podcasts - first-hand stories
  6. SEC filings, annual reports - for public companies

Example: Travel Apps Idea Maze

Suppose you want to build a travel app:

Maze Map
  • 1990s: Expedia, Booking.com - hotel aggregators
  • 2000s: TripAdvisor - reviews, TripIt - trip organization
  • 2010s: Airbnb - alternative lodging, Hopper - price prediction
  • Failures: Dopplr (social tracking), Gogobot (social recommendations)
  • What changed: AI for personalization, post-pandemic travel patterns

Tools for the Idea Maze

In the course, we use AI assistants to accelerate research:

  • Idea Maze Coach - a prompt for Claude/GPT that helps structure your research
  • Competitor Analysis Template - a template for analyzing competitors
  • Timeline Builder - industry history visualization

This article is a brief overview of the Idea Maze concept from the "AI Founder" course. In the course, you'll get access to AI prompts for research and apply the method to your own idea with mentor guidance.

Want to put this into practice?

In the "AI Founder" course, you'll not only learn these frameworks but apply them to your own idea with guidance from a mentor with 3+ years of AI startup experience.