Product Funnel Not Tunnel

Filter ideas ruthlessly—lots at the top, few at the bottom—don't ship everything that enters

Bill Carr
Unpacking Amazon's unique ways of working | Bill Carr (author of Working Backwards)

Product Funnel Not Tunnel

"Think of yourself honestly as a venture capitalist. They don't fund every company that they meet with. They actually fund a very, very low percentage of them. And at Amazon, we had lots and lots of PRFAQs that were a great idea, but we didn't ship them because we had other ones that were just a better idea, which had a bigger potential impact." - Bill Carr

What It Is

Product Funnel Not Tunnel is the principle that your product development process should filter ideas aggressively, not pass everything through to execution. A funnel has lots of ideas at the top and few at the bottom; a tunnel has everything that enters also exit.

This mental model applies to Amazon's PR-FAQ process: many ideas get written up, reviewed at lower levels, and discarded. Only the best survive to reach senior leadership. The filter isn't bureaucracy—it's quality control for your most precious resources (engineering, design, time).

Bill Carr frames it as thinking like a venture capitalist: VCs don't fund every pitch. They fund a very small percentage of extremely promising opportunities. Product teams should do the same with their ideas.

How It Works

The Funnel Stages (using PR-FAQ as example):

  1. Author alone - Many ideas get crumpled up before sharing
  2. Manager review - Feedback causes more ideas to be abandoned or revised
  3. Wider team - Discussion reveals fatal flaws in some ideas
  4. Leadership - Comparison to other ideas filters further
  5. CEO/S-Team - Only the very best make it here

Example from Bill Carr: "If you truly run this out and you write 100 PRFAQs in a year, maybe 20 of those make it their way to the CEO."

The Tunnel Problem: When everything that enters also exits, you:

  • Lack a real prioritization method
  • Deploy engineering on less-than-best ideas
  • Make decisions based on politics or loudest voice
  • Burn resources on ideas with fatal flaws

How to Apply It

Step 1: Create explicit filtering stages Define where ideas get evaluated and by whom. Each stage should have criteria for what advances.

Step 2: Encourage high volume at the top The funnel works because you generate many ideas. If you only have 10 ideas and need 8 to ship, you can't filter.

Step 3: Make early filtering cheap Early stages should be low-fidelity. A rough PR-FAQ is cheap to write and abandon. A built product is expensive.

Step 4: Compare, don't evaluate in isolation The question isn't "Is this a good idea?" but "Is this better than the other ideas competing for the same resources?"

Step 5: Celebrate killing ideas Reframe discarding ideas as success, not failure. You found the fatal flaw before spending engineering resources.

Step 6: Protect the funnel from politics Without clear process, what gets built often comes from:

  • Whoever is loudest
  • Whoever has political power
  • Top-down mandates without scrutiny

The funnel creates a meritocracy for ideas.

When to Use It

  • Product planning and roadmapping
  • Innovation programs
  • Any process that allocates scarce resources
  • When you notice every idea "makes it through"
  • When prioritization feels political or arbitrary

Diagnosing Tunnel Syndrome

You might have a tunnel, not a funnel, if:

  • Every idea that gets proposed also gets built
  • You're always "almost done" but never shipping the right things
  • Prioritization meetings devolve into politics
  • Your best engineers work on mediocre ideas
  • No one can explain why one idea beat another

The Fatal Flaw Filter

Bill Carr emphasizes checking for fatal flaws early:

"What would often happen is that once you've spent some time looking at that idea more deeply, you then start to identify several roadblocks or maybe a fatal flaw with this idea. And in fact, no, you shouldn't waste any of your time going into building that thing because it has a fatal flaw."

Common fatal flaws:

  • No clear customer problem being solved
  • Problem isn't big enough to matter
  • Solution doesn't actually solve the problem
  • Economics don't work
  • Can't be built with available resources

Source

  • Guest: Bill Carr
  • Episode: "Unpacking Amazon's unique ways of working | Bill Carr (author of Working Backwards)"
  • Key Discussion: (00:48:00) - Explanation of concentric circle review and funneling
  • YouTube: Watch on YouTube

Related Frameworks

Back to Execution