把目的地画出来：YouTube 产品总监怎么写、讲、卖一份愿景

Atawodi opens by insisting that every PM, however junior, needs some micro or macro vision. She uses a pilot analogy: a passenger would think twice about a flight if the pilot said only that the company needs to fly 400 flights. A good pilot paints the destination instead, Miami, beaches and 24 degrees, while the mission is simply flying people safely where they are going.

She lists four properties of a good vision. It should be lofty enough to scare you in an exciting way, yet realistic and attainable. It should be set apart from the limitations of today, so that you can imagine yourself five years out and report back what must be fixed now. And it must be grounded in a clear, potent user problem.

Her best example comes from Uber, where the mission was reliable transportation everywhere for everyone, and the vision was a city without parking, where continuous trips free up the 25% of urban space used for it. That vision led to ideas like Uber Pool. Elon Musk's push toward Mars is her example of one that sits at the edge of attainable, and Lenny adds that a confident leader can make the seemingly impossible feel possible.

She and Lenny then pin down the distinction with a mountaineering analogy. The vision is describing what you will see from the top of Everest, and the mission is why you are climbing and the principles for getting there together. Lenny cites Stripe's and TED's missions and Microsoft's vision of a computer on every desk as parallels.

Atawodi calls storytelling one of a PM's most powerful skills and says pure stories travel further than numbers blended with stories. Her simplest tool is an "once upon a time" template: once upon a time there was a problem, then one day something happened, and as a result the world is different, ending with what you finally leave people with. She illustrates it with YouTube Shorts, where polished long-form content made people feel they could not create until 60-second videos let anyone express themselves again.

The second format is an Amazon-style working-backwards article. She writes the headline and subtitle that would run when the vision comes true, pastes it into a TechCrunch or New York Times page so it looks real, and may write the full article later. At Uber she did this for a multimodal trip, with a headline about Uber replacing your Clipper card, and while her team did not build that product, it informed the payments and commerce platform they did build.

The third format is a picture. At Uber a design partner drew, in pencil, a world where you could top up your Uber balance with cash at any corner shop, and it took four years to build a version of it. At a YouTube strategy session she handed out sheets with blank app-store panels and asked what screenshots would sell the solved problems, which forces teams to choose three or four big ideas.

She rejects "I don't have a designer" as an excuse, since sketches and low-fidelity mocks work, and she cites Steve Jobs's advice to tell the story without slides. In her own case, a designer who saw her post-it sketch turned it into a low-fidelity mock and then a polished one, even though she was only an L4 PM at the time.

Atawodi's process for developing a vision has three steps: empathize, create and evangelize. Vision work is slow at senior levels, she notes, because her "Vision 2026" for YouTube only reached the planning cycle a year and a half later. Patience matters since naysayers will always say engineering cannot do it, which is why she wants the vision set apart from today's technical limits.

To empathize, she wants PMs to do "understand work", including going through their own onboarding flow, with multiple accounts on YouTube, Instagram and TikTok. She calls using your own product "dog food" and using competitors' products "cat food". Research is valuable for foundational problems a few cycles ahead, but she says PMs should also trust the intuition built by exposing themselves to great products.

Her signature tool is a living "top 10 things you should know" document in which every PM lists up to ten problems, qualitative, quantitative or tech debt, and revises them quarterly. At Uber she called it "more money, more problems" and paired PMs with data scientists, researchers and engineers. The test is whether her direct reports and her engineering and design counterparts all give the same top five problems.

On research, Lenny cites a Patrick Collison view that research updates your mental model of the customer, and that model then informs what to build. Atawodi agrees and says that if you could drop research into ChatGPT and get a PRD out, you have not done your job. She sees four pillars of a PM, product sense, leadership, execution and technical ability, and stresses that it is product sense, a feeling, not product logic.

For a more strategic view, Atawodi asks stakeholders such as marketing, support, research and content strategy to bring their own "10 things you should know" and stack-rank them, instead of presenting 50 slides. One surprising entry in a recent session was that the average American reading age is 11, which prompted thoughts about using more images and video. She then sandpapers the lists down to a final ten.

Her strategy sessions run three days, with two days as the absolute minimum so that there is white space for the work. Day one is insights, day two is strategy, and day three is the big rocks. The room usually has four leaders from product, engineering, design and research, plus data science where needed, and people who were consulted find the evangelizing stage much easier.

To evangelize, she thinks in three concentric circles. The core is the team, starting with the studio leads and PM weekly, with the workshop output written up in a doc open for comments only, because friction polishes ideas like rocks in a washing machine. Next come stakeholders and adjacent teams, who bring objections, and finally leadership, where she advises going as big as possible and letting managers pull you back.

She notes this effort is worth it for long-term visions, and for near-term ones she suggests simply writing the tweet. Her team's mission and vision show up in hiring conversations and job descriptions, where candidates' eyes twinkle. Short links make the vision easy to find, and she says it ripples broadly through the company.

If you have to rewrite the vision every year, Atawodi says, you have not created a good vision. A macro-vision should last at least three years, like a computer on every desk or going to Mars, while teams add fidelity over time, from low-fidelity mocks to a sizzle reel at Uber. PMs from L4 to L7 on her team each have a variation, and smaller problems get micro-visions, such as a mock of next year's shipped experience.

Lenny highlights the "how might we" phrasing, which one PM he worked with called a fertile question because it leads to productive brainstorms. When asked what to work on first, Atawodi says she is baffled by how many PMs ramble when asked for the top problems that keep them up at night. Start with ten, she says, and she usually ends up with three or four, like the four numbers and four insights she showed YouTube's new chief product officer.

She also argues that infrastructure is the product and tech debt is product debt, because you cannot build a skyscraper on a shaky foundation. After listing problems, she suggests sketching on a post-it what the world looks like if you solved them. Even convincing yourself with that picture is already evangelizing the vision.

Atawodi says product management is "clarity and conviction". Clarity means removing what pollutes the core problem, and she gives small examples, such as emails that leave readers unsure whether they are an FYI or a request. It also saves cycles, since a team that is clear on a known problem can skip foundational research and go straight to UX validation.

Her tool for clarity is a two-page document, four at most, with insights, a strategy or approach, and big rocks. The big rocks should be three to five memorable things, because an endless roadmap is like pouring a cocktail before the ice, and everything else is the sand around the rocks. She frames it as the pitch a startup would give investors, and she points inquirers to it by link so that many intro calls fall away.

This narrative is not the vision, she says, but an evergreen document refreshed every quarter or six months and written before planning. On her team, each group took the overarching vision, wrote a two-pager and gathered feedback from partners, then built the roadmap from it. A rule of thumb was to consolidate any problem with fewer than three engineers, and the vision itself can be shaped in a day in a room, as she did on a no-meeting day.

Atawodi defines conviction as a feeling of how the world should be. It is not certainty or perfection, and it is what people mean by product sense. Everything discussed so far, she says, is about converting the conviction in your head into something others can consume, so that the vision is the conviction and the narrative is the clarity.

She says that if you have conviction but it is not clear, you do not really have it. She stress-tests teams by asking what they would build with only five resources, and she warns against "peanut buttering" resources too thinly. If a team comes with two options and pros and cons, she expects them to have done the work of choosing and to ask for help mitigating specific risks.

Lenny summarizes conviction as picking the right thing, and Atawodi adds the battle metaphor. Clarity shows you are committed to fighting a particular war, and conviction is how you see the world if you win that battle.

Atawodi started at Uber as a general manager in Nigeria, where there were seven cars on the road, and describes a hands-on boardroom-to-streets culture with high autonomy. She says culture will evolve whether or not you steer it, so leaders should ask which behaviors they reward and which they condone. She links the lack of that revisiting to the "Delete Uber" moment, and cites the value of principled confrontation, under which Travis Kalanick, who opposed cash, still let her team test it, which is why cash exists on Uber.

At Netflix she saw a very intentional evolution of culture through a structured debate process in which people write their arguments and the decider owns the decision. She says that because the buck stops with you, people work harder to move from 95% to 99% conviction. Debates about access, in a world where many people lack a TV or credit card, eventually led to the introduction of advertising.

At Google she uses a story of two young fish asking what water is to describe a culture that is felt more than defined, built on respecting the user, the opportunity and each other. Within it, microcultures in YouTube, Cloud, Maps and Search differ, like neighborhoods in a city, under a looser macro culture. Lenny contrasts aggressive and friendly cultures at Uber versus Lyft and Airbnb versus Wimdu, and Atawodi adds that Uber's value of celebrating cities let local teams adapt the product.

Atawodi and her engineering and design partners deliberately designed their team culture, borrowing from the Netflix memo and Uber. She cites freedom and responsibility, being an owner and not a renter, and an informed captain who is the single person on the hook for a decision, not a committee. The model resembles RAPID, where one decider gathers the context.

She says she does not optimize for being liked. She defines love as the choice to extend yourself for the spiritual growth of yourself or another, which sometimes means hard conversations, and teams that trust that care can take raw feedback. On her team people send PRDs to an email alias with a plus tag so peers can help shape them.

She suggests testing a relationship with your engineering manager by asking if you know their birthday, work anniversary and career goals. The idea came from an early rough start with an associate engineering manager that later turned into friendship, and Lenny reveals that manager was Gergely Orosz. She argues that since we spend half our waking life at work, we might as well enjoy it with people we like and love.

Asked about YouTube, Atawodi says her team sees YouTube Studio as a creative partner for creators big and small. She highlights AI inspiration, launched at Made on YouTube, which would suggest ideas based on what a creator's audience watches and is due next year, and a thumbnail test-and-compare feature that tests two or three thumbnails against each other. The latter had just rolled out to a larger set of creators.

For aspiring PMs, she says she is a Nigerian woman with no MBA and no consulting background, and that people from other fields bring empathy for the other side. Her advice is to start doing product management before getting the title, by opening favorite apps, listing the top ten problems and designing the fixed world in your head. That trains product sense so that opportunity meets preparation.

In the lightning round she recommends *48 Laws of Power*, *The God of Small Things* and *Outliers*, and names *The Bear* as a recent favorite. She asks managers about their leadership philosophy and other candidates about a favorite product and what they would improve. Her motto is the closing lines of the poem *Invictus*, "I am the master of my fate, I am the captain of my soul", and she DJs, publishing mixtapes named after sauces on YouTube.