Pixar's Braintrust: How Brutal Honesty Saved Toy Story 2 (And Every Film Since)

Toy Story 2 was a disaster.

Four months before release, Pixar screened an early cut for the leadership team.

It was bad. The story didn’t work. Characters felt flat. Jokes landed with thuds.

John Lasseter, Pixar’s creative chief, said: “This isn’t working.”

Most studios would have panicked, blamed the director, or shipped it anyway.

Pixar did something different.

They convened the Braintrust.

A group of Pixar’s best storytellers sat in a room with the director. For hours, they picked apart every scene. Brutally honest. No sugarcoating.

Not: “It’s great, just needs a little polish.”

But: “This character has no motivation. This scene is boring. This joke doesn’t land.”

Ruthlessly candid.

But here’s the key: The director wasn’t required to take any of the advice.

The Braintrust had no authority. They could only observe and suggest.

The director retained final creative control.

Four months later, Toy Story 2 was released. It got 100% on Rotten Tomatoes and made $511 million.

The Braintrust saved it.

And it’s been Pixar’s secret weapon for every film since.

What Is the Braintrust?

The Braintrust is Pixar’s feedback process for films in development.

The structure:

• Frequency: Every few months during production
• Duration: 4-8 hours
• Attendees:
  • The director and core team
  • 8-10 of Pixar’s most experienced directors and storytellers
• Format: Watch the current cut, then discuss

The goal:

Surface problems the team is too close to see.

The Rules

1. Candor is mandatory

Don’t hold back to be polite.

If something isn’t working, say it clearly.

Bad feedback: “Maybe the third act could be stronger?”

Good feedback: “The third act is boring. The villain has no stakes. I don’t care what happens.”

2. The director has no obligation to follow the advice

The Braintrust has no power.

They can’t force changes. They can’t override the director.

They can only shine a light on problems.

What the director does with that information is up to them.

Why this matters:

• Removes defensiveness (they’re not forcing anything)
• Preserves creative ownership
• Director feels safe hearing hard truths

3. Focus on the work, not the person

Not: “You’re failing as a director.”

But: “This scene isn’t working.”

Not: “You don’t understand story.”

But: “The character’s motivation isn’t clear.”

Attack the problem, not the person.

4. Everyone is equally candid

This isn’t about rank.

Junior directors give feedback to senior directors.

Ed Catmull (Pixar President) said:

“In the Braintrust, there is no hierarchy. Everyone’s opinion matters equally if it serves the story.”

5. The goal is to make the film great, not to be right

Not: “I told you that wouldn’t work.”

But: “Let’s figure out how to fix this together.”

No points for being right. Only points for making the film better.

How It Works in Practice

Here’s what a Braintrust meeting looks like:

Phase 1: Watch the Film (2 hours)

Everyone watches the current cut in silence.

No interruptions. No commentary.

Taking notes:

• Moments that feel off
• Story holes
• Pacing issues
• Character inconsistencies

Phase 2: Discuss (4-6 hours)

The director presents: “Here are the problems I’m aware of. Here’s what I’m struggling with.”

Then the Braintrust responds:

Typical conversation:

Braintrust: “The villain isn’t threatening. Why should we fear them?”

Director: “We show them defeating the hero in Act 1.”

Braintrust: “But we don’t feel it. The stakes are abstract. What do they personally lose?”

Director: “Maybe we need to show the villain hurting someone the hero cares about.”

Braintrust: “Yes. Make it visceral.”

Director takes notes. Doesn’t commit. Just listens.

Phase 3: The Director Decides

After the meeting, the director goes away and decides:

• What feedback resonates?
• What should we change?
• What should we ignore?

Key: The Braintrust doesn’t mandate solutions. The director owns the decisions.

Why This Works

1. Removes Politeness Filters

Most feedback sessions are too polite.

“It’s good! Just a few minor tweaks.”

Result: Real problems don’t get surfaced.

The Braintrust removes the filter:

“This isn’t good enough. Here’s why.”

Because everyone knows:

• The goal is making the film great
• No one gets fired for giving hard feedback
• The director wants to hear it

2. Collective Intelligence

The director is too close to the work.

They’re thinking:

• We spent 6 months on this scene
• The voice actor nailed this take
• The animation is gorgeous

They can’t see:

• The scene doesn’t serve the story
• The joke doesn’t land
• The pacing drags

The Braintrust sees with fresh eyes.

3. Psychological Safety (Because There’s No Power)

Traditional executive reviews:

• Executives have power
• “Fix this or you’re fired”
• Director becomes defensive
• Feedback is a threat

The Braintrust:

• No power
• “Here’s what we see”
• Director stays open
• Feedback is a gift

Ed Catmull: “The Braintrust has no authority. That’s the key. Remove the power dynamic, and you get honest conversation.”

4. No Politics

Traditional studios:

• Director pitches to executives
• Executives want something different
• Political maneuvering
• The loudest voice wins

The Braintrust:

• Everyone cares about the story
• No agendas
• No pet features
• Just “Is this working?”

5. Iteration, Not Perfection

The Braintrust doesn’t expect the first cut to be good.

They expect it to be bad. That’s normal.

The philosophy:

“Every film sucks at first. The Braintrust is how we make it not suck.”

This removes shame from early drafts.

Real-World Examples

Toy Story 2: Saved from Disaster

As mentioned, Toy Story 2 was failing.

Braintrust diagnosis:

• The villain (Al) wasn’t threatening
• Woody’s motivation was unclear
• The emotional stakes were low

What they changed:

• Made Woody’s fear of obsolescence the core theme
• Deepened the relationship between Woody and Andy
• Added the “When She Loved Me” sequence (Jessie’s backstory)

Result: One of Pixar’s best films.

Ratatouille: Restructured in the Middle

Midway through production, the Braintrust told director Brad Bird:

“The pacing is off. Act 2 drags.”

Bird’s response:

He restructured the entire second act. Moved scenes around. Cut 20 minutes.

Braintrust: “Better. Now the villain needs more screen time.”

Bird added more scenes with Ego (the food critic).

Result: Ratatouille became critically acclaimed.

Up: The Problem with Act 3

Braintrust feedback:

“The first 10 minutes are heartbreaking. The rest is just okay. The villain feels tacked on.”

Director Pete Docter:

“We built the entire movie around this villain. We can’t cut him.”

Braintrust:

“We’re not saying cut him. But the emotional core is the relationship between Carl and Russell. The villain is a distraction.”

What changed:

They refocused Act 3 on Carl letting go of the past (the house) and embracing Russell.

The villain stayed, but became secondary.

Result: Up’s emotional depth is what people remember, not the villain.

The Anti-Patterns (What Doesn’t Work)

Pixar tried variations that failed:

Attempt 1: Add More People

Thinking: “More perspectives = better feedback.”

Reality:

• Too many voices
• Meetings became unfocused
• Signal-to-noise ratio dropped

Fix: Keep it small. 8-10 trusted people max.

Attempt 2: Make It Mandatory to Follow Advice

Thinking: “If the Braintrust spots a problem, the director must fix it.”

Reality:

• Directors became defensive
• Feedback sessions became battles
• Creative ownership eroded

Fix: Restore director autonomy. Braintrust can only suggest.

Attempt 3: Have It Too Early

Thinking: “Catch problems early!”

Reality:

• Film is too rough to evaluate
• Feedback is premature
• Director already knows the problems

Fix: Wait until the director has a coherent cut, even if rough.

Attempt 4: Have It Too Late

Thinking: “Wait until we’re almost done.”

Reality:

• Too expensive to make big changes
• Team is exhausted
• Feedback feels like criticism, not help

Fix: Regular cadence. Every 3-4 months during production.

How to Adapt the Braintrust for Software Teams

You don’t make movies. But you make products. The principles transfer.

1. Design/Product Reviews

Traditional:

• Designer presents mockups
• Stakeholders say “looks good” or nitpick colors
• No real critique

Braintrust approach:

• Show the design
• Ask: “Is this solving the user’s problem?”
• Give candid feedback: “This flow is confusing” “The value prop isn’t clear”
• Designer decides what to change

Rules:

• No politeness filter
• Focus on the user, not the designer
• Designer has final call

2. Architecture Reviews

Traditional:

• Engineer proposes architecture
• Architects approve or reject
• Political dynamics

Braintrust approach:

• Engineer presents design
• Team asks hard questions: “How does this scale?” “What’s the failure mode?”
• Candid feedback: “This creates a single point of failure”
• Engineer owns the decision

3. Retrospectives

Traditional:

• “What went well, what didn’t?”
• Polite, superficial
• No real problems addressed

Braintrust approach:

• Watch a “replay” of the sprint (demo, metrics, incidents)
• Brutal honesty: “Our velocity dropped because we didn’t plan dependencies”
• Focus on system, not people
• Team decides what to change

4. Post-Mortems

Traditional:

• Production incident
• Blame game
• CYA mode

Braintrust approach:

• No blame
• Candid: “The deploy process is broken” “We have no rollback mechanism”
• Focus on fixing the system
• Team decides improvements

How to Build Your Own Braintrust

Step 1: Pick the Right People

Not:

• Most senior people
• People with authority

But:

• People with good judgment
• People who care about the outcome
• People who can give candid feedback

5-8 people max.

Step 2: Set the Rules

Before the first meeting, be explicit:

1. Candor is expected
2. No authority (we can only suggest)
3. Focus on the work, not the person
4. Goal is to make the project better, not to be right

Write them down. Remind people every meeting.

Step 3: Create Psychological Safety

The first meeting will be awkward.

People will be too polite.

Your job as facilitator:

Model candor. Say hard truths. Show that it’s safe.

Example:

“I’m going to be blunt: This feature doesn’t solve the user’s problem. Here’s why…”

Others will follow.

Step 4: Remove Power Dynamics

Don’t:

• Have the CEO in the Braintrust
• Let execs override decisions
• Make attendance mandatory for junior people

Do:

• Keep it peer-level
• Ban rank-pulling
• Make it about the work

Step 5: Iterate the Format

Your first Braintrust will be messy.

• Too long
• Too shallow
• Too polite

That’s okay. Improve it.

After each session:

• What worked?
• What didn’t?
• How do we improve?

Meta-Braintrust the Braintrust.

Common Mistakes

Mistake 1: Confusing Candor with Cruelty

Candor: “This user flow is confusing. The user doesn’t understand what to do next.”

Cruelty: “This design is terrible. Did you even think about the user?”

Candor focuses on the work. Cruelty attacks the person.

Mistake 2: Giving Solutions Instead of Identifying Problems

Bad: “Change the button to blue and move it to the left.”

Good: “I didn’t notice the button. It doesn’t draw my eye.”

The Braintrust identifies problems. The creator solves them.

Mistake 3: Being Too Polite

The biggest failure mode:

Everyone wants to be nice. Real problems don’t get surfaced.

Fix:

Reward candor. Thank people for hard feedback.

Mistake 4: Not Giving the Creator Control

If the Braintrust can overrule the creator:

• Creator becomes defensive
• Feedback feels like an attack
• Psychological safety breaks

Fix:

Make it explicit: “You own the decision. We’re just here to help you see clearly.”

Mistake 5: Doing It Once

The Braintrust isn’t a one-time event.

It’s a recurring process throughout the project.

Pixar: Every 3-4 months.

Software: Maybe every sprint or milestone.

The Deeper Lesson

The Braintrust works because it separates insight from authority.

Traditional feedback:

• Senior person gives feedback
• Feedback comes with implicit authority
• Receiver must comply or face consequences
• Result: Defensiveness, politics, CYA

The Braintrust:

• Peer gives feedback
• Feedback has no authority
• Receiver chooses what to do with it
• Result: Openness, honesty, better work

The insight:

Good feedback doesn’t need authority. It just needs clarity.

If the feedback is right, the creator will see it and act on it.

If the feedback is wrong, forcing compliance makes things worse.

Trust the creator to make good decisions when they have good information.

Key Takeaways

• The Braintrust is candid feedback without authority
• Focus on the work, not the person
• The creator retains final decision-making power
• Small group of trusted peers (not executives)
• Regular cadence throughout the project
• Psychological safety comes from removing power dynamics
• Candor is not cruelty - be honest, not harsh
• Creative teams love this because it makes the work better

Pixar has released 27 feature films.

All 27 have been profitable.

Most have been critically acclaimed.

Many are considered masterpieces.

This is unprecedented in film history.

The secret?

Not technology. Not talent. (Though they have both.)

The secret is the Braintrust.

A process that surfaces problems early, gives brutal honesty, and preserves creative ownership.

You don’t need to make films to use this.

You just need:

• A team that cares about the work
• A culture that values candor
• A willingness to separate feedback from authority

Build that, and you have your own Braintrust.

And your work will be better for it.