Idea-Volume Learning Flywheel
Generate, critique, and test many ideas to increase the learning rate
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 99%
Snap's small, flat design team operates on the premise that good ideas emerge from many attempts. New designers present something on their first day, when limited context makes an excellent proposal unlikely. That deliberately makes the feared event—showing weak work—happen immediately and signals that an imperfect idea is acceptable. The team then holds weekly critiques, examines new designs, and collaborates with product and engineering to develop promising ones. Spiegel estimates that most ideas are not good and emphasizes that the goal is not to predict perfectly but to maximize learning. The flywheel is therefore volume, critique, selection, implementation, and feedback. Volume alone creates noise; the recurring review and connection to builders turn attempts into organizational learning.
Origin
Spiegel describes this as the operating method of Snap's nine-person design team.
Core principles
- 01Most ideas can fail without the process failing
- 02A higher volume of thoughtful attempts creates more chances to find a strong idea
- 03Early exposure to imperfect work reduces fear and invites participation
How to run it
- 1
Lower the fear threshold
Ask each person to present work early, before they can polish away all uncertainty.
Pro tip Make the first contribution small enough to complete quickly.
Watch out Do not use early presentation as a disguised performance trap.
- 2
Generate alternatives
Produce many possible solutions to a defined problem instead of defending the first one.
Watch out Unfocused brainstorming is not a substitute for a clear problem.
- 3
Critique on cadence
Bring the team together regularly to look at work, question it, and identify promising directions.
Pro tip Keep the review centered on the work and user problem rather than status or titles.
- 4
Bridge to builders
Pair selected ideas with product and engineering so constraints and possibilities improve the design.
Watch out A detached idea team can produce concepts that never become useful products.
- 5
Feed learning back
Use implementation and user response to shape the next round of ideas.
Pro tip Optimize for the speed of this complete loop, not raw idea count.
In the wild
Every new Snap designer must make and present something on the first day. Because they lack company context, the work is unlikely to be their best, but showing it makes imperfect contribution normal from the outset.
→ The ritual lowers fear and opens the door to repeated creative attempts.
Common mistakes
Counting ideas as the outcome
The transcript emphasizes learning and finding strong ideas, not hitting an idea quota.
Punishing visible failure
If weak ideas harm status, people will hide unfinished work and the flywheel will slow.
Is it for you?
Best for
It is best for small design or innovation teams that can review work frequently and test selected ideas.
Not ideal for
It is not ideal when idea volume is rewarded without any selection, user feedback, or delivery capacity.
From the transcript
“The best way to have a good idea is to have lots of ideas.”
“More ideas, more feedback.”
“Ultimately you have to maximize your rate of learning.”
From the episode
Snapchat CEO: Exact Formula Used To Build A $130 Billion Company! I Said No To $3 Billion From Mark Zuckerberg! It’s Time To Quit Your Job When You Feel This!