别从第六章讲起：Casey Winters 谈向上沟通、守住已有增长，以及先点燃再烧起来

Casey Winters traces his career from an analyst role at apartments.com, where he measured marketing channels and then moved into product, through GrubHub, where he was the fifteenth employee and was told to just grow how many people ordered food. That was growth before the term existed. He then led growth product at Pinterest from 40 million to 150 million monthly users, advised companies like Airbnb, worked at Greylock, and has been Eventbrite's chief product officer for three years.

At Eventbrite during the pandemic, Winters noticed managers trying to handle hard situations alone instead of escalating them. Later they would say he did not understand how impossible things were, and he would reply that of course he did not, because no one had told him. Escalating lets a leader change the circumstances, or at least understand the trade-offs made, communicate them across the company and evaluate results fairly.

He thinks people generally under-communicate upward and then complain that executives are out of touch. When they do communicate, they are often so deep in the weeds that the executive feels like Matt Damon in *Ocean's Twelve*, being spoken to in code. His framing is that anyone below the executive level is telling a story, and it must start at chapter one: which part of company strategy you are working on, which metrics, and which assumptions.

The opposite mistake is to spend 20 minutes re-explaining strategy that everyone already knows and run out of time before the new information. Winters coaches his team to find the last point in the story that would be completely obvious to the listener and continue from there. Telling that story differently to executives, designers and engineers takes real work.

Winters says the best coaching happens before a meeting, not after. Product people like a big reveal and an "aha", but he wants to de-risk the meeting instead. He runs through important presentations with his team while playing different audience members, such as the CFO asking about numbers or the CEO asking why, so those answers are woven in before anyone asks.

He adds that executive communication is really communication with individual executives, who have different styles and concerns. Pre-meetings with key people help surface concerns early so nobody is surprised. He is now trying to scale this with shared frameworks across his team, since confidence is part of how these meetings go.

How much prep is needed varies by person. Winters himself does better free-forming from things he knows deeply, because he writes a lot of notes and documents, while others need extensive rehearsal. What matters is how well you know the material and your audience.

At Pinterest, product reviews with the CEO Ben and head of product Jack required knowing Jack's early data questions, because failing to answer them cast doubt on the rest of the meeting. His test is whether you have considered the whole universe of questions the document could provoke. That might take three hours or dozens of hours, but it is a high-leverage use of time because key decisions get made in these forums.

Winters cites Adobe CPO Scott Belsky's product life cycle: users flock to a simple product, the product adds features for power users, and users then flock to the next simple product. Companies try several design hacks to avoid that: un-bundling, as Facebook Messenger and Uber Eats did, progressive disclosure, which Pinterest invested in heavily, proactive training, or segmented experiences for different user types.

None of these fit Eventbrite well, because event creators range from someone expecting five guests to people running a hundred events a year, and creators move between those levels. Segmentation and progressive disclosure fail because some users should never see the advanced material. Instead the team aims for "perceived simplicity", where advanced features are easy to find if you look, but effectively hidden if you do not.

WhatsApp is his best example. It is a chat app at its core, yet voice messages, video calls and phone calls each took him less than a second to learn when he needed them. An Eventbrite example is automated Facebook advertising for creators, where the default is simple and handles targeting and creative, with an on-ramp for those who want more control.

Winters argues that some of the most impactful projects at scale, such as performance, user experience and developer velocity, are the hardest to measure and so are chronically underfunded. His tactics include building custom metrics to show their value, running small tests, setting team principles, and using experiments to build wider buy-in. First, though, you need peers on board: two or three people, such as your engineering manager and design lead, aligned on why the work matters.

The second argument is about mindset. Early-stage companies are always trying to add upside, but at scale you can lose what you have built, so you need to help managers see what a high conversion rate or strong engagement would look like if it went away. Pinterest learned this when its growth stalled after changes in the market related to Facebook, and later when a search algorithm change hit its SEO-driven growth.

Winters links this to his chart of product-market fit over time: user expectations and the competitive landscape keep rising, so if you stop improving experience and latency you may fall out of fit entirely, maybe in a year or five. Recovering takes longer, and you also have more technical debt to clean up. Lenny jokes it will give founders nightmares, and Winters says that is not the intention.

Winters says he got the idea for his ops post after being invited to speak to marketers at a marketing-operations conference, having just written that martech should target engineers more than marketers. His provocation is that needing marketing ops can signal you are bad at marketing. He ran a double-digit-million marketing budget at GrubHub without it because the team invested in automation and process, and he is wary when the first tactic is adding people to scale.

He is not against ops, since Eventbrite has a product ops team, but he objects to ops as a permanent function doing manual process. At Eventbrite their explicit job is to find inefficiencies and build process or software to remove them, and then move on. He agrees with an Intercom post that the goal of business operations is to not exist, and calls growing ops teams to patch a functional problem a form of empire building.

Asked what a CPO actually does, he says he leads and facilitates products that create value for customers and translate to value for the business. At Eventbrite he leads product management, design, research and growth marketing. His job includes choosing and adjusting strategy, improving the process for prioritizing and executing, removing bottlenecks, training the team and hiring.

For his own path, he focused on leverage, learned the whole business and refused to specialize, which slowed visible progress but gave him a faster route to being able to talk with CEOs about many topics. He says the CPO role is a company leadership role, so you must optimize for the company even at the expense of your team. He also advises assuming executives' questions are attempts to learn, not tests.

Inspired by Omar, who runs product at Cambly and previously ran core product at Pinterest, Winters describes a spectrum of product people. On one end are innovators with many ideas, most of them bad and one in ten a game changer, who struggle to turn ideas into action. On the other end are executors who deliver on a strong strategy but do not track the industry or generate new product ideas.

Every CPO wants people in the middle, but there are not many of them. If forced to lean, Winters would take executors over idea generators, because a company always has too many ideas, whereas a VC should lean toward innovators because only one in ten investments needs to work. The problem is that executors who want to reach director level or beyond need to become more strategic, and the industry has no good way of making that transition.

He is trying several things, including Reforge strategy programs, mentoring, and bringing in speakers to show the team what great looks like. Early in a PM career, he says, the most valuable thing is shipping real things customers like. But if you want to run a business unit or a pillar, he expects you to write the strategy doc without him, and many people cannot, which caps their careers at senior PM or group PM level. He calls it "the great filter".

On new channels, Winters says there are not new ones as such, unless you count Web3 tokens, which he does not. Instead companies are getting more leverage from their existing channels. At Eventbrite, the team unified separate direct-response and lead-generation flows for acquiring creators, which lowered effort and improved cost per acquisition, and lets sales pick up product-qualified leads who need help.

He calls this product-led sales: joining self-service product loops and sales loops in a B2B business into one more efficient system that breaks down silos. He expects an explosion of B2B companies that get sales, product and marketing working as one cross-functional team. A second trend is founders building growth loops into products before product-market fit, not to scale early but so distribution is ready when fit arrives, since an acquisition loop is arguably a requirement for fit.

On timing, he separates growth into two phases. "Kindle" strategies are non-scalable hacks to get early users, usually done by founders. "Fire" strategies, such as content, sales, viral loops or paid acquisition, drive scale. Kindle strategies exist only to unlock fire strategies, and once you unlock one, you hire a full-time person, whether a salesperson, growth PM or performance marketer, to make it ten times better.

The tactic he thinks is most underrated is data network effects, by which he means using product usage data to make the product better, not selling data. Examples are personalized results at Pinterest and better ad targeting at Eventbrite. With Facebook and Apple platform changes, he says a product that generates its own data has a real edge.