EO StudioHow to Build a Product that Hits PMF on Day 1 | Granola, Christopher Pedregal
CHAPTERS
- 0:00 – 1:00
PMF on day one: why fast user feedback beats founder certainty
Christopher Pedregal opens with the claim that Granola had product-market fit on launch day, and uses it to frame a core belief: user conversations are the lifeblood of building product. He stresses that feedback must arrive quickly enough that the team still remembers why they made the original decisions.
- •Granola effectively had PMF at launch, but the team didn’t recognize it immediately
- •User testing reveals misunderstandings you can’t predict from the inside
- •Long feedback cycles degrade learning because decision context is forgotten
- •User feedback is essential—but blindly building requested features is a trap
- 1:00 – 2:01
What Granola is and the founder’s motivation for building tools
Chris introduces himself and Granola: an AI meeting notepad that turns rough notes into polished outputs after meetings. He shares the company’s rapid growth and why building tools that improve people’s lives is personally motivating.
- •Granola: AI notepad that listens in meetings and refines notes afterward
- •Launched ~18 months ago; grew from 4 people to ~35
- •Raised Series B at a $250M valuation; fast growth trajectory
- •Intrinsic motivation: building tools that help people do things better
- 2:01 – 3:01
From curiosity to HCI: products succeed or fail at the human layer
Chris traces his path from early fascination with the internet to studying computer science and discovering human-computer interaction. The central lesson: even deeply technical systems often fail because they don’t match how people behave.
- •Early inspiration from Google and Silicon Valley’s “how do they do it?” mystique
- •Security lesson: most vulnerabilities come from human/system design mismatch
- •Discovery of human-computer interaction as a guiding discipline
- •Great products require understanding users in the moment of use
- 3:01 – 5:54
Designing “products with soul” (and why org structure shouldn’t show)
He explains Granola’s internal ideal that the product should feel like it has “soul”—coherent, consistent, and value-driven, like interacting with a person. As teams grow, alignment becomes harder, and inconsistencies can reveal organizational seams.
- •“Soul” = coherence, consistency, and shared values expressed through the product
- •Users form mental models of products similar to mental models of people
- •Misalignment across contributors makes experiences feel disjointed
- •If you can infer org structure from the UI, that’s a warning sign
- 5:54 – 7:26
Explore vs. exploit: the pre-launch advantage and the 50% cut
Chris outlines an explore/exploit operating model: first search broadly for the right shape, then polish relentlessly once you’ve found it. Granola delayed launch for a year, explored heavily, and then cut about half the product to focus on what mattered most.
- •Explore phase: try many shapes, add/experiment rapidly
- •Exploit phase: polish details only after the shape is right
- •Granola stayed unlaunched ~1 year, enabling drastic changes
- •The “50% rule”: they cut roughly half the product before launch
- 7:26 – 8:26
Knowing when to launch—and why PMF recognition is so hard
He describes the decision point: continue small-scale onboarding or launch to learn from many more users and unexpected use cases. Despite strong traction, Granola didn’t realize for six months that PMF had been present on day one, illustrating how ambiguous the signal can be.
- •Launch timing framed as a learning-rate decision: small cohort vs. broad exposure
- •They launched when it felt “good enough” to learn faster at scale
- •PMF is hard to identify in real time—even for famous companies (Facebook anecdote)
- •Founder mistake: not noticing PMF for six months despite day-one fit
- 8:26 – 9:27
The “fog-of-war” metaphor: product solutions are unknowable upfront
Chris pushes back on the idea that founders can design the perfect solution in advance. He compares early product building to exploring a video game map: you must ship probes into the world and learn from what comes back.
- •Execution isn’t the whole job; discovering the terrain is unavoidable
- •Early-stage understanding is blurred; clarity comes from movement and contact
- •You can’t fully reason your way to PMF without real-world interaction
- •Iterate by probing, observing responses, and refining direction
- 9:27 – 10:58
Systemize user conversations by lowering activation energy
He argues that talking to users must be a repeatable company system, not a heroic effort. Lessons from his prior company Socratic show that creating reliable access to users dramatically increases product iteration speed.
- •User feedback must be continuous to build something great
- •Make user contact easy; high “activation energy” kills frequency
- •Socratic’s challenge: talking to high schoolers was logistically hard
- •Creating a standing cadence of user access accelerated improvement
- 10:58 – 11:28
Granola’s operating cadence: standing interviews and remote-native feedback
Granola benefits from remote user interviews that mirror the product’s meeting context. Chris describes a structured program—standing interviews multiple days per week—open to anyone in the company to join and learn firsthand.
- •Remote interviews (Zoom) fit naturally because Granola is used in meetings
- •Standing user interviews four days a week
- •Company-wide access: anyone can join, builders can ask directly
- •Institutionalizing proximity keeps learning fast and shared
- 11:28 – 11:59
Stop redundant feedback: fix obvious issues immediately to learn the next thing
He explains the compounding value of fast iteration: if a problem is clearly understood from one session, don’t wait for 10 confirmations. Ship the fix so the next user session reveals the next bottleneck instead of repeating the same one.
- •You don’t need 10 users to validate an obvious usability issue
- •Change immediately once you understand the confusion
- •Avoid big-company patterns that over-rely on repetition and process
- •Speed of iteration determines the rate of insight accumulation
- 11:59 – 12:29
Interviewing with skepticism: probe negatively and ignore flattery
Chris emphasizes rigor in user research: people are polite, so positive comments are unreliable. He advises never trusting stated intent (“Would you use this?”) and instead testing behavior with probing, sometimes adversarial, follow-ups to surface truth.
- •Never trust “would you use this?” answers; focus on what they’d do next
- •Probe with skeptical questions to stress-test claims
- •Assume users will be nice; treat praise as noise
- •Honesty and critical thinking prevent self-delusion
- 12:29 – 12:59
Use conservative metrics to avoid kidding yourself about engagement
He describes Granola’s intentionally strict definition of an active user to prevent inflated engagement narratives. Only users who run a new meeting with meaningful transcription count, which forces clarity about real usage versus superficial activity.
- •Define ‘user’ conservatively: new meeting that day + >5 minutes transcription
- •Opening the app or reviewing old meetings doesn’t count as active usage
- •Strict metrics reduce rationalizations and false confidence
- •Conservatism keeps the team anchored to real value delivered
- 12:59 – 13:59
Intuition over feature requests: build context, then design the right product
Chris rejects the idea of turning user requests into a prioritized backlog. Instead, user conversations are for building deep context; product direction should be guided by vision and intuition, informed by a vivid mental model of real users.
- •Don’t build whatever users ask for; requests conflict and mis-specify needs
- •Use interviews to internalize context, then apply product intuition/vision
- •Granola’s vision: simple, minimal, pleasant, non-distracting design
- •Teams should be able to “predict” reactions of specific users they know well
- 13:59 – 15:30
Shorten feedback loops to sharpen intuition—and keep doing it forever
He returns to the practical discipline: intuition improves only through frequent exposure to real users, and it decays when you stop. Granola’s iterative approach—ship the minimum, observe failures, fix quickly—depends on tight loops to stay aligned with reality.
- •User misunderstanding is common even for “great” internal ideas
- •If feedback takes a month, learning quality collapses
- •Iterative method: build minimum, find failure reasons, fix, repeat
- •Intuition improves with practice but degrades when user contact stops
- 15:30 – 16:01
AI product strategy: augment humans and expand the workflow beyond notes
Chris closes by positioning Granola as an AI assistant that enhances rather than replaces human work. Starting with meeting notes, the roadmap expands into follow-ups, memos, preparation, and cross-meeting analysis to improve daily productivity.
- •Two AI paths: replace the human vs. augment the human
- •Granola chooses augmentation: help with post-meeting work
- •Future capabilities: follow-up emails, memos, meeting prep, analytics across meetings
- •Goal: incremental daily improvement in users’ work lives
