Products with legs show in dilated pupils, not in DAU metrics
Google Labs director says early products are measured by whether people's eyes light up—a signal that comes sooner and more reliably than any retention or DAU dashboard.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Good ideas don't come from brainstorms
Josh Woodward says good ideas almost never come from scheduled brainstorms or innovation sprints, but from people tinkering relentlessly on weekends or in hallways. The team maintains an ‘almost possible’ list—whenever an experience crosses the threshold from impossible to possible, they immediately assemble a team to pursue it. The Labs internal docs contain 82 predictions starting with ‘We believe the future is’—Josh admits most of them will be wrong, but maintaining a concrete vision of the future is what matters.
— Josh WoodwardTest products by watching dilated pupils, not retention charts
When judging whether an early prototype has potential, Josh looks not at DAU, not at daily-to-monthly active ratio, not at 7-day retention, but at whether someone's eyes light up during a demo and whether their body leans forward. He believes teams should ‘fall in love with the problem, not with any specific solution’—products worth shipping usually need three to five pivots, and those who fall in love with their first solution tend to crash hard.
— Josh WoodwardTeams know when to kill a project before leadership does
Whether to kill a project, Josh says, the team usually knows first—ebbing enthusiasm is the signal. He gave an example: a Gemini feature about to ship, but after internal testing and eye-contact feedback, the PM sent a company-wide email saying ‘I think we shouldn't ship this,’ and Josh publicly thanked her. Leadership's real job, he thinks, is not to decide unilaterally but to build an environment where teams can say ‘this isn't working."
— Josh WoodwardModel benchmarks are overrated; values conversations are underrated
Josh argues that the industry overestimates model benchmarks—most people don't care about Elo ratings and don't even know how to pronounce them. What's underrated is having principled debates about ‘what kind of product do we want to build and what future do we want?’ Products express the company's values; this conversation deserves more airtime than chasing speed and proxy metrics.
— Josh WoodwardHiring depends on unlearning speed and explosive endurance
One signal Josh now watches when hiring is ‘unlearning rate’—how fast someone can learn a conclusion and then unlearn it, pivoting to a new method. Another comes from a book on Federer: ‘explosive endurance’—sprinting all-out while also taking care of yourself and your teammates for the long haul. He also stresses the ability to build trust collaborating with AI and with people; as execution costs drop, these human skills grow more valuable.
— Josh WoodwardEveryone becoming a builder is a misdiagnosis
Some think all roles will merge into one ‘builder’ type. Josh disagrees: PMs adapt fastest to new roles because their work is more horizontal, but if everyone becomes builder, specialized depth gets diluted—he compares it to majors and minors; vibe coding is the minor, don't drop the PM major. Meanwhile, Labs projects now run on 2–3 people, sometimes 4, down from the old 5–7. The team has optimized for speed over headcount.
— Josh WoodwardLabs lifecycle splits into three stages to protect new experiments
As projects like Notebook LM and Flow graduated from Labs into large-scale products, Josh spotted a risk: mature projects crowd out the effort to run zero-to-one experiments. The team now divides Labs lifecycle into ‘zero to one,’ ‘one to ten,’ and ‘ten to 100,’ each phase tracking completely different metrics—eye-contact feedback for zero to one, conversion funnels and CAC for ten to 100. This carved-out space keeps new experiments alive.
— Josh WoodwardDoubling team size is a vanity metric, not success
Scaling the experimental org, Josh points out, often trips up on hiring too fast: doubling or tripling headcount is intoxicating and often a vanity metric. He recalls an early project he led that reached 30 engineers and made him feel great—but the product-market fit was zero, buoyed only by an inspiring vision. That story predicts its own ending.
— Josh WoodwardIn their own words · checked verbatim
they don't come in conventional places. You can't microwave good ideas.
Josh Woodward6:24
what I'm always doing in those is you're looking at people's eyes. And like that is the metric. It is not a Dow. It's not a Dow over a Mao. It's not a D7. It's none of that.
Josh Woodward15:50
The team usually knows before the leader knows and the signs are like, does the passion start to run out?
Josh Woodward17:52
most people in the world are never, ever, ever going to care about an ELO rating. They may not even know how to pronounce ELO.
Josh Woodward28:13
I'm actually looking a lot right now for people. What's their unlearning rate? Like, how fast can they learn something and then walk away from it?
Josh Woodward32:21
I read a book a year or so ago about Roger Federer, the amazing tennis player. And there was this phrase in there about Federer had explosive endurance.
Josh Woodward32:21
there's zero to one labs there's one to ten labs and there's like ten to a hundred
Josh Woodward46:41
I remember I was on a project years ago and we got to like 30 engineers and I was so happy and we had zero product market fit, but a very inspiring vision.
Josh Woodward52:58
Figures
| Future-belief statements in Labs docs | 82 | 10:39 |
| Labs project team size change | 5–7 down to 2–3, sometimes 4 | 34:23 |
| Product milestone cycle | ~50–100 days | 42:35 |
| Planning window | ~6 months rolling | 42:35 |
| Avg lifespan of internal labs in large companies | 3–4 years before degradation | 44:39 |
| Early-stage project team size and fit | 30 engineers, zero product-market fit | 52:58 |
Glossary
- almost possible
- A shortlist the team maintains tracking technologies or experiences on the verge of crossing from impossible to possible.
- unlearning rate
- How quickly someone can learn a conclusion and then unlearn it, pivoting to a different approach.
- explosive endurance
- The ability to sprint at full intensity while sustaining yourself and your teammates over time.
- good trouble
- Constructive friction the lab deliberately creates by challenging the company's existing processes and ways of working.
How to listen
Executives building innovation or lab teams inside their companies; founders and PMs who need to decide whether to keep shipping a product.
The WorkOS and DX sponsorship intros at the start can be skipped.