先抄对、再做更好、最后才加新：Mark Pincus 的爆款产品公式

Pincus describes Proven, Better, New as the framework Zynga adopted early on and treated almost like a religion in product management. It rests on his belief that human instincts are right about 95% of the time while the ideas layered on top are wrong about 75% of the time. The framework isolates the instinct in your gut and tests many ideas around it so that you fail for the right reasons rather than the wrong ones.

He illustrates with Sid Meier's Civilization on Facebook. Within ten minutes of launch, Zynga's product managers emailed that it was dead on arrival, because the first-time user experience took so many clicks that nobody would ever reach the great game design. Even a famous designer tripped over best-practice onboarding, and his innovation was never seen.

Proven means copying what works, legally and with taste, until you are, as he puts it, a PhD in it. For an AI camera app, that means first studying the best existing mobile cameras. Better is usually a small, polished increment, such as being free or needing no download, that 10 out of 10 existing users would say yes to, and it is what power users notice most.

The new idea is the reason to try the product, such as Words With Friends attaching Scrabble to the Facebook graph so friends were already there. Pincus insists we should assume the new idea will probably fail and have four others ready to test. He says he wishes he could tell his younger self that one of his instincts was a year ahead of a company now worth $1.6 trillion, and that his Tribes idea resembled what became Reddit.

Lenny summarizes the framework and points to the iPod and iPhone as examples. Pincus says founders often misuse "proven" to justify a bad idea, pointing to a game that was popular in the 90s. Proven has to be precise: proven on this platform, for this audience, for this experience. Slack, he suggests, may have been proven and better with little new at all, which can be even better because people do not like change.

Copying feels like cheating because we were taught that in school, and entrepreneurs want to be innovators, so Pincus calls it a moral arbitrage for people with less ego. His advice to product makers is to burn your resume and define ambition in the eyes of your consumer, such as nurses in Indiana for FarmVille, rather than in the eyes of peers. If you only copy, there is no reason to choose you, but one inch better can beat something she has never seen.

He tells the story of Craig Newmark spending two years adding photos to Craigslist listings. Pincus sees it as the mark of a world-class product maker, because people feel ownership of products they rely on and get angry when they change, even for the better. A junior maker might turn the whole page into pictures and push the text people used to compare prices below the fold.

Not every product follows this path, he says. Rovio made 45 games before Angry Birds was a hit, which he likens to wildcat drilling, while OMG Pop, on its last dollar, copied the turn-based system from Words With Friends for Draw Something. That became the number one app for 60 days in 2012, and Zynga bought the company.

Pincus says he is still relearning this lesson. If you are too visionary at the outset you will probably miss product-market fit, because the starting points of massive hits are embarrassingly small, like Facebook starting as a way to check out people at Harvard. His first two companies, Freeloader and Support.com, began with humble premises and did better than he expected.

After success, he says, we tend to feel we must do something bigger, and with Tribe he tried to do everything instead of picking one use case. People loved the idea, but it failed, and he was humbled enough that at 41, a multi-time founder, he built a Facebook poker game. People thought he had no dignity, but he brought his ambition down to what he calls a thousand-foot altitude, which was the key to success.

He thinks new founders even have an advantage over experienced ones, who have too much rope to hang themselves because it is easy to raise money against a big vision before finding product-market fit. Startups also benefit from not needing a multi-billion-dollar opportunity from the outset, unlike a large company. He cites Bolt.new, which toiled in obscurity before combining its web-stack virtual machine with an AI coding assistant, and Slack's origins in a team's internal tool after a game attempt.

Pincus adds that founders face a tension between showing confidence and staying honest. He admits that being transparent to a fault has burned out people at Zynga and since, and he is still learning. The real founder mode, he argues, is having the courage to tell your team and investors that the current B+ is not the north star.

Pincus distinguishes belief from hope: hope is confidence without basis, while belief is grounded in your experience, user behavior and data. The best product makers, he says, are collecting winnings rather than making bets, and he cites Airbnb's Brian Chesky as someone who already knows a launch is a hit. He contrasts a minimum viable product with a maximum launchable product and says the word "viable" is where hope comes from.

AI, he says, is so powerful it is dangerous, because getting to a viable product in three months instead of three years makes it easy to spend that time on a single idea. He expected people to use AI to build testing machines that evaluate more ideas in a week than an industry does in a year. His advice is to build it wrong before you know it is right and be willing to waste a day or a week, because the point is to learn.

He gives a Zynga example from a FarmVille expansion pack. The team had a $10 million ad budget, but Pincus suggested putting teaser art on the game board instead, where tens of millions of players were already active. The clicks showed which marketing worked, and offering early-access keys for a friend and oneself brought in $19 million, turning an afterthought into signal, direction and revenue.

Lenny asks about people who are put off by games like FarmVille and CityVille. Pincus says those were Zynga's biggest hits by installs, engagement, retention and revenue, while Words With Friends and Poker were more enduring because they moved to mobile. He admits Zynga was partly a victim of its own success because viral games took over friends' feeds, but says eight of ten major launches became hits because of more than spamminess.

The first reason was a mission of connecting the world through games and adding what he calls a bold beat, a new dimension to something that had gotten boring. Games let people, especially middle-aged women who loved FarmVille, share a hobby with friends through co-op play and gifting. Zynga described its games as invest, express and connect: people do not want to be creative, they want to feel creative.

The second was the core metric, retention. Pincus thinks Zynga may have been the only consumer company tracking day-365 retention, and argues the most valuable companies have the highest of it, though early indicators like day 1 and day 30 predict it. He compares viral-driven companies to sinking speedboats that try to go faster than they sink.

Zynga also built a metric called Active Social Network, counting round trips between players, such as a turn and a turn back or a gift and a gift back. Going from zero to one gave an 80% chance of seeing a player the next month, and four gave an 80% chance of seeing them 22 of the next 30 days. The team could build and innovate directly against that number.

Asked what would make a consumer social app work today, Pincus starts with latent demand: a category's absence does not mean people do not want it. In 2007 he had stopped playing games even though gaming was a $23 billion business, because it was too much effort, so he made games free, familiar, three clicks and five to 15 minutes. Gaming is now a $280 billion industry, and he thinks there is still more latent demand there.

For social, he says the demand is there but the adrenaline is gone: using Instagram can feel like eating potato chips. He cites NPS data showing that people who quit Facebook or Instagram went from plus 35 to minus 35, as if they had quit smoking. Someone, he believes, will reinvent social for the agentic AI age, probably by offering real productivity, like the lead generation that Napster, Friendster, Facebook and LinkedIn offered.

His central metaphor is the cocktail party, a place you are glad to be that produces great leads. Zynga dropped games into the cocktail party of Facebook, but today, he says, everyone is hanging out in Claude or GPT and there is no party, only a quiet, lonely web as before social networking. His challenge to listeners is to figure out how to make it rowdy.

Answering how to know a product is a B+, Pincus compares it to dating: when you are with the right person, you do not ask whether this is the one. Quoting Osho, he says that if you are torn between two paths, both are wrong. When you have lightning in a bottle you love and are addicted to the product, friends love it, and the metrics show it, and nobody had to ask whether GPT was it.

The power, he says, is in being intellectually honest that it is a B+, and then deciding whether to kill it or use it to learn. As an example, he pulled the plug two weeks before the interview on his Earth project, his fourth attempt at a version of the metaverse he has been building for 20 years. It cost four years and $25 million, but since then he has been more inspired by ideas than he has been in four years.

On distribution, Pincus argues AI is not yet a new platform: it is an important technology with a new portal in chat apps, but not a hardware or interface platform for other developers. We are in the mobile and web era, and since average app installs per user per month are around zero, discovery is not happening. He says roughly 40,000 new games launched last year and none became a top-ten hit, which makes categories like social and games hard to invest in.

Consumer distribution therefore cannot be an afterthought, because hoping for virality or word of mouth is a hope strategy rather than a belief strategy. He expects the cost of getting into the market to rise a hundredfold within a year, and sees success coming from prosumer approaches that target power users willing to pay up front. He also repeats an idea from Gary Tan: design consumer services assuming tokens become nearly free, since their price will fall in a way pig ears and delivery never did in the dot-com era.

Pincus also floats ideas for agents. One is a social membrane, an intelligent layer of trust where agents broker plans between people without hurting feelings, like a great chief of staff. Another is an agentic travel agent, a free 24/7 service that rebooks you mid-trip, which he thinks has latent demand but never made economic sense for human agents.

He is somewhere between belief and hope that a consumer-facing platform for AI apps will emerge, likely driven by the major model makers as they look to differentiate beyond coding. Lenny wonders whether the labs will simply build everything themselves. Pincus points to Microsoft's history of absorbing what surrounded it, and Facebook, which skipped only music and games.

Pincus says all his management principles came from desperation because he dislikes managing. The core of management, he found, is getting people to do the right thing when you are not in the room, so he gave leaders a hill to take and made them real CEOs, with operating control, a plan and a budget. Those he hires are often frustrated expert witnesses: know-it-alls with pent-up demand to prove they were right.

The expert witness syndrome is that people in the trenches are closest to the data and the right answer but furthest from decisions. He advises founders to stay close to the metal, citing Steve Jobs choosing conference-room carpeting and the founders of Discord, who decided the founders should be the first and last mile of the UX instead of delegating it to the least experienced people. The best player should be on the field, not spending the time on investors and scaling.

He says micromanagement is beautiful and that he did it to a fault, with two-hour standups through 50 employees, and that it made the team more effective. The first principle is to be in the room as much as you can, and only delegate when you cannot be everywhere. To spread your passion, he used a teaching-hospital approach of bringing people into your meetings and a tech assistant who follows you for six to 12 months, as people did for Bezos.

Another idea borrowed from Bezos is that a CEO's number one job is to be right. Being in the right body of water matters more than having the right boat, Pincus says, and he would take misfits who are right over people who fit well. He looks for people who have been right about something, and for intellectual honesty.

Pincus, a father of five with children ranging from a one-year-old to teenagers, says his greatest job is growing good humans. His first principle is meeting kids where they are, neither talking down to them nor treating them like adults, and engaging human to human. He taught his twins and their friends what he calls daddy math during the pandemic, and later found they had covered material through eighth grade.

He thinks we are at the end of a hundred-year cycle of mass-produced education aimed at factory work and then knowledge work, and that knowledge work is changing. He tells his older kids he does not care if they go to college, but that they develop critical thinking and find a way to be useful. He teaches them to ask why someone holds a view and to ask better questions rather than know more answers, and to be generative, not consumptive, online and offline.

His daughter Carmen, who has ADHD and dyslexia, created a sweatshirt brand called Comfy Fancy and a group for neurodivergent middle schoolers. He kept a running Google doc of life philosophies for his older girls, including that nothing is personal and not to be a victim, and he admits his attempt to hold off on smartphones until 16 ended at 14.

Asked for a closing thought, Pincus says it took him until he started Zynga at 41 to articulate his why: he wants to create an internet treasure, a service people cannot remember life before or imagine life without. He calls it the greatest ambition for product makers. He hopes readers will steal from the playbook in Life at the Speed of Play and carry it further.