Intuition as Hypothesis Generator
Treat product taste as a hypothesis machine, not an oracle — then debate it down to a working hypothesis
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 95%
Dylan Field's answer to the 'founder has magic taste' myth: intuition is not a verdict, it is a generator. Its output is a stream of hypotheses that must be surfaced, argued, and tested against evidence before any of them become the working hypothesis the team ships against. The quality of the generator is improved by constant, wide ingestion of raw user signal — every mention, every support channel, every complaint — plus relentless questioning to reach the root problem rather than the stated ask.
Origin
Dylan Field's own stated framework, given at Figma Config 2024 in response to Lenny Rachitelli asking how he built his product intuition — prompted by a Rick Rubin clip about being paid for 'the confidence that I have in my taste'. Field explicitly reframes taste-as-authority into taste-as-hypothesis-generation.
Core principles
- 01Intuition produces hypotheses, not conclusions — the conclusion comes from debate and data.
- 02Your hypotheses compete with everyone else's; the leader's guess gets no automatic priority.
- 03The generator is only as good as its input — ingest signal from everywhere, continuously.
- 04Users state solutions, not problems: 'I need X' often means they want Y or Z.
- 05Converge to ONE working hypothesis, then move — the point of the debate is to unblock, not to stall.
How to run it
- 1
Feed the generator with raw, unfiltered signal
Continuously ingest what users actually say about the product — social mentions, support channels, forums, anywhere on the internet. Field reads mentions of Figma and reposts them into a Slack channel so the team shares the raw feed, not a summarized digest.
Pro tip Route the raw feed into a shared channel rather than a curated report — summaries strip out the anomalies that generate the best hypotheses.
- 2
Interrogate to the root problem
When a request arrives, ask questions until you understand where the person is coming from and what they are actually trying to solve. Do not build the literal ask; find the underlying need behind it.
Pro tip Ask for concrete examples — Field says he 'asks for examples a lot' because an artifact is debatable in a way an abstraction is not.
Watch out People asking for X frequently want Y or Z. Shipping X literally satisfies the request and misses the problem.
- 3
Emit the hypothesis explicitly
Convert the intuition into a stated, falsifiable hypothesis and put it forward. Collect competing hypotheses from the rest of the team at the same time so yours is one candidate among several, not the default.
Watch out An unstated intuition cannot be argued with — it just becomes an unchallengeable executive preference.
- 4
Debate and seek disconfirming data
Argue each hypothesis and actively hunt for data that supports OR negates it. Field explicitly frames the search as two-sided: evidence to support them or negate them.
Pro tip Understand the thing from first principles rather than accepting the consensus answer — Field treats this as the highest-leverage habit for important decisions.
Watch out First-principles digging must be time-boxed. Field's own caveat: it leads to good outcomes 'as long as you make sure it's not bottlenecking the organization.'
- 5
Narrow to a working hypothesis and move
Winnow the candidates down to one working hypothesis, declare it, and execute from it. The working hypothesis is a commitment, not a truth claim — it holds until new evidence overturns it.
In the wild
Field's intuition told him pages were not the elegant solution inside a coherent product-design system, and he remains skeptical of the implementation to this day. He put the hypothesis forward, the team and the market pushed back with the evidence that users needed pages, and the hypothesis lost.
→ Figma shipped Pages and did not unship them. Field kept his private skepticism but the working hypothesis was set by evidence, not by the CEO's taste.
Rather than relying on periodic research readouts, Field ingests every mention of Figma across Twitter/X, support channels and the wider internet, and pipes them into an internal channel — a permanent stream of unfiltered complaints and praise that seeds new hypotheses.
→ A continuously refreshed intuition, grounded in what users actually say rather than in what a summarized report says they said.
Common mistakes
Treating the leader's intuition as the decision
If the founder's hunch skips the debate-and-data stage, the organization stops generating competing hypotheses and simply guesses at what the boss wants. Intuition is the input to the process, not a substitute for it.
Building the literal request
Users describe solutions, not problems. Shipping the stated feature without digging to the root need produces a product that answers questions nobody actually had.
Letting first-principles inquiry become a bottleneck
Field admits that when he keeps asking questions with no answers, people can reasonably think 'this is obvious, why don't you get it'. The rigor pays off only if it does not stall the organization.
Is it for you?
Best for
Founders and product leaders whose team defers to their 'taste' and who want to convert that deference into a testable, debatable decision process without losing the speed of instinct
Not ideal for
Low-stakes, easily reversible decisions where the cost of running a debate-and-data cycle exceeds the cost of simply being wrong and correcting later
From the transcript
“intuition is like a hypothesis generator”
“generating these hypotheses and others are generating hypothesis as well and you then take these hypothesis and you put them forward and you debate them…”
“I'm trying to constantly ingest information about figma uh and it's not just like Twitter SLX whatever that's called now uh but you know anywhere…”
“sometimes people are are saying hey I need X but they really want y or Z”
“it's it's important though to just really understand something from first principles”
From the episode
Dylan Field live at Config: Intuition, simplicity, and the future of design