EP 311 - Duolingo Fmr. Director of Product on Decision Science in Product Management

本播客由Mighty Capital赞助，这是一家位于硅谷的风险投资公司，投资于既出色又具有商业潜力的产品，如Airbnb、Digital Ocean和Amplitude。

This podcast is brought to you by Mighty Capital, a Silicon Valley venture capital firm that invests in great products that are also great businesses, like Airbnb, Digital Ocean, and Amplitude.

像Amplitude这样的高速增长公司称我们为投资回报率最高的合作伙伴。

Hypergrowth companies like Amplitude call us the best value for the dollar invested.

了解更多，请访问mighty.capital。

Learn more at mighty.capital.

你好。

Hi.

我是Essi Moadi，Products That Count的创始人。这是一个具有影响力的加速产品创新的平台，覆盖全球超过30万名产品经理，并帮助思科、Capital One等领先企业将产品团队转化为竞争优势。

I'm Essi Moadi, founder of Products That Count, an influential product acceleration platform that reaches over 300,000 product managers globally and helps leading brands such as Cisco and Capital One turn their product teams into a competitive advantage.

我是本播客的主持人，这档节目位列苹果播客前500名和福布斯前15名科技播客。

I'm the host of this podcast show, an Apple top 500 and Forbes top 15 tech podcast.

本节目的唯一目标是定义什么是优秀的产品。

This program's sole focus is to define what makes a great product.

虽然这听起来像是一个简单的问题，但答案却十分复杂。

While it seems like a simple question, the answers are complex.

于是我转向了CVP级别的产品高管，邀请他们主持一系列围绕产品构建特定主题的访谈，涵盖从提升用户参与度、功能优先级排序到促进多样性等多个方面。

So I turned to CVP level product executives and invited them to host a series of interviews that cover a specific theme of building products, from maximizing engagement to prioritizing features to encouraging diversity and more.

你将听到这些产品领袖与同行的对话，比如Netflix的产品副总裁、Coinbase的首席产品官和Box的首席产品官。

You'll hear these product leaders in conversation with peers, such as Netflix product VP, Coinbase CPO, and Box CPO.

我希望，通过我们的共同努力，能够一次打造一个更好的产品，从而逐步构建一个更美好的世界。

My hope is that together, we'll build a better world one product at a time.

你准备好了吗？

Are you ready?

让我们一起来聆听今天的节目，认识我们的主持人和他们的嘉宾。

Let's hear today's episode and meet our hosts and their guests.

你好，我是海蒂·吉布森，欢迎收听《产品对话》。

Hello, this is Heidi Gibson with Product Talk.

今天我非常兴奋地为大家带来豪尔赫·马扎尔，我的嘉宾将为我们详细介绍决策科学以及如何将其应用于产品管理。

I am very excited today to bring you Jorge Mazal, who my guest today, who is going to be telling us all about decision science and how to apply that to product management.

豪尔赫，首先请你自我介绍一下，跟我们分享一下你的背景吧？

Jorge, first off, would you introduce yourself and tell us all a little bit about your background?

当然。

Sure.

谢谢你邀请我，海蒂。

Well, thank you for having me, Heidi.

今天能和你聊天真的很愉快。

Pleasure to be chatting with you today.

不久前，我还是多邻国的首席产品官，在那里工作了大约五年，负责产品、数据科学、用户体验研究，并协助制定产品战略和公司战略，审查产品变更，基本上涵盖了产品负责人所做的所有工作。

Until recently, I was the Chief Product Officer at Duolingo, where I was there for about five years leading product, data science, UXR, and just helping generally with product strategy, company strategy, reviewing product changes, basically everything that head of product does.

在加入多邻国之前，我担任过几个其他产品职位，我的产品生涯始于Zynga，之后作为产品研究员加入Feiner Perkins，然后去了MyFitnessPal。

And before Duolingo, I did a few other product roles, started my product career at Zynga, Feiner Perkins as a product fellow, then went to my fitness pal.

最终，我成为了一款名为Lucidchart的B2B生产力应用的首席产品经理，之后加入了多邻国。

Eventually, I was the lead PM for a product called Lucidchart, which is a B2B productivity app, and then went to Duolingo.

这就是我的大致经历。

So that's basically me.

所以UXR是指用户体验研究吗？

So UXR is UX research?

只是想为大家澄清一下。

Just want to clarify for everybody.

没错。

Correct.

是的。

Yes.

对。

Yeah.

有意思。

Interesting.

所以你的背景几乎全是面向消费者的，但那次进入了B2B领域。

So your background's almost all consumer, but but that one foray into b to b.

我们稍后再回过头来聊聊，这对你的影响可能有什么不同。

Well, We'll pick that back up a little later and talk about how that might have been different for you.

你最初是什么促使你进入产品管理的？

What got you into product management in the first place?

我当时在读研究生。

I was in grad school.

我修的是MBA和MPA联合项目，学制三年。

I did a combined MBA and MPA, and it was a three year program.

那段时间是我探索人生下一步方向的重要时期。

And it was really a time of exploration of what I wanted to do next in life.

我原本以为读研后会创业，成为一名企业家。

And I thought going into grad school, was going to probably start a business, be an entrepreneur.

但我意识到，当时我的生活状况并不适合创业。

But I realized that where I was in life wasn't really conducive to that.

我已婚，刚有了我们的第一个孩子，这非常美好，但同时也极其耗费时间和精力——既要当父母，又要上学，还要应付其他事情。

I was married, had our first child, and it was delightful, but also very time consuming, very exhausting to be a parent and have school and everything else.

所以我意识到，我需要一份时间投入更可控的工作。

So I realized I needed something that was a little more measured in terms of time commitment.

我通过一次雇主做的演讲了解了产品管理。

And I learned about product management from a presentation that some employer was doing.

这感觉像是一个完美的平衡点，既有创业带来的乐趣和刺激，又有稳定的工作和职业保障。

And it felt like a sweet spot where we have a lot of fun and exciting things of being an entrepreneur, but also the stability of a job and a career.

所以我觉得这是正确的选择。

So it felt like the right thing.

幸运的是，我申请了Zynga的产品经理实习，并且非常享受产品经理的工作。

And luckily, I applied for a PM internship at Zynga and really enjoyed the work of a PM.

因此，当我毕业时，我决定回到这个领域。

And so when I graduated, I decided to go back to that.

那MPA是什么？

So what's an MPA?

MPA是公共管理硕士。

An MPA is a Master's in Public Administration.

在我从事产品工作和读研之前，我曾在非营利组织工作，担任一家专注于通过培训、辅导和部分资金支持帮助移民和少数族裔成为小企业主的组织的教育主管。

So before my product career and before grad school, I worked in nonprofit, where I was director of education at an organization that focuses on helping immigrants and minorities to become small business owners through training, coaching and some financing.

我曾在类似的非营利组织做过志愿者。

I volunteered at a similar nonprofit.

真有趣，我们在这方面有共同点。

That's fun that we have that in common.

我也有个刚满四岁的孩子，所以我猜我们可以做个完整的播客，聊聊在科技行业工作和当父母的事。

I also have a kid that just turned four, so I suspect we could have a whole podcast with working in tech and being a parent.

是的，

Yes,

这确实有很多独特的挑战。

there's a lot of unique challenges to that.

但没错，MPA的经历非常有趣。

But yeah, the MPA experience was super interesting.

你知道，他们对这个项目的理解方式是在哈佛肯尼迪学院。

You know, the way they they think about this particular program, it was at the Harvard Kennedy School.

他们深刻认识到政府渗透在生活的方方面面，因此在学习内容上给了你很大的灵活性。

They really understand that government is involved in everything in life, so they give you a lot of flexibility of what you're gonna study when you're there.

所以我专注于你今天早些时候提到的一个领域，也就是决策科学，这基本上是研究人们如何做决定。

And so I focused on something you mentioned earlier today, which is decision science, which is basically researching how people make decisions.

研究或思考这个问题的方式有很多很多种。

And there's many, many ways to to study that or to think about that.

所以我专注于这个领域，它非常迷人。

So that that's what I focused on, and it was fascinating.

那么，这和心理学有什么不同呢？

So how how is that different than psychology?

因为听起来，为什么这个领域会在公共管理学院，而不是心理学系呢？

Because it sounds why would that be in public administration versus, like, a psychology school?

是的。

Yeah.

他们的理念是，会尝试应用一些心理学概念，以及其他如神经科学或战略学的概念，来帮助人们做决定，并以有利于良好政策、公共政策或人类整体进步的方式促进决策。

The idea is that they will try to apply some psychological concepts and other things as well, like some neuroscience concepts or even strategy concepts to help people make decisions and also help people make decisions in a way that's conducive to good policy or good public policy or general improvement of humanity.

因此，这些方法在政府或教育等领域中的应用是独特的。

So the application of those things was unique government or education, for example.

好的，我能想象到，可以设计方法或活动来减少吸烟、提高疫苗接种率，或者类似的事情。

Okay, so I could see crafting methodologies or campaigns to reduce smoking or increase adoption of vaccines or things like that.

没错。

Exactly.

对。

Yeah.

关于这方面的标志性研究，或者说典范案例，是过去人们刚加入公司时，公司会问你是否愿意加入401(k)计划，那时你必须主动选择加入。

The hallmark research on this or the poster boy of this is back in the day, when you were joining a company, they would ask you if you wanted to join the four zero one ks, you kind of have to opt in.

当时大约只有30%的人会选择加入。

That was about 30% of people would opt in.

后来他们做了一项研究，决定默认让所有人自动加入，但允许他们在文件中选择退出。

And then they did this research where they said, we're going to opt everyone in and give them the chance to opt out in the paperwork.

结果，这个比例完全颠倒了。

And then they basically reversed the statistic.

绝大多数人只是保持了默认的加入状态，真正主动选择退出的人非常少。

The vast majority of people just stayed opted in, and very few actually practically opted out.

他们认为，从纯粹理性的角度来看，是需要主动选择加入还是默认加入后可以选择退出，其实不应该有什么区别。

And they felt like it's it's an interesting finding that from a purely rational point of view, whether you opted in or have to opt out or or vice versa, it shouldn't really matter.

但因为人们并非完全理性，你可以通过一些小的干预来影响他们的选择。

But because people are not 100 rational, you can make small interventions that will impact their choices.

如果你是一个有道德、有伦理的人，你会把这些干预手段用于人们的福祉。

And if you are a moral, ethical person, you will use these interventions for their well-being.

是的，你提到了默认选项，需要主动退出。

Yeah, you mentioned a default, having to opt out.

我立刻想到，在我的领域里，这属于一种黑暗模式，对吧？

And my head immediately went to, well, in my world, that's a dark pattern, right?

你必须非常小心地使用这种黑暗模式。

And you have to be real careful about using dark patterns like that.

但在这个案例中，这恰恰是一个用黑暗模式做好事的绝佳例子，对吧？

But in this case, like it's a great example of using a dark pattern, I'll say for good, right?

为了用户自身的利益。

For the kind of the user's own good.

我认为你职业生涯的很大一部分，不是围绕黑暗模式，我并非想暗示这一点，而是帮助你的客户实现他们的目标。

And I think you've made a bit of a career, not around dark patterns, don't mean to imply that, but kind of helping your customers can achieve their goals.

是的。

Yeah.

你能跟我们详细说说吗？你是怎么看待这个问题的？

Would you tell us all a little bit about that and how you think about it.

当然。

Sure.

实际上，当我还在上学、读研究生的时候，我最喜欢的一门课就是将教育游戏化。

Actually, back when I was in school, still in grad school, one of my favorite courses that I was in was gamifying education.

我记得上这门课的时候，我想想，那甚至还是Duolingo公司成立之前。

And I remember taking this course, it was, I think if I remember correctly, this was even before Duolingo was even a company.

所以当时现实中几乎没有这方面的例子。

And so there wasn't a lot of examples of that in the real world.

非常理论化。

Was very theoretical.

我对这个概念着迷不已。

And I was just fascinated by this concept.

而且我从未想过这会成为我的职业。

And didn't imagine there was something I would do for a living.

但当我毕业之后，去了Zynga公司，游戏化以及将游戏作为决策实验场的理念让我非常着迷。

But when I finished school, went to Zynga, and this idea of gamification and games as a laboratory for decision making was very fascinating.

在我做了大约十个月之后，我记得有机会去了解你会怎么称呼它？

And after I did that for about ten months, I remember I had the chance to see what would you call it?

我正在做的这款游戏中的一个高价值客户。

A high value customer from the game I was working on.

这群人的平均消费金额，我不知道，大约是8万美元在游戏上。

And this group of people had an average spend, I don't know, like $80,000 on the game.

天啊

Oh, my

上帝啊

God.

8万美元花在一个游戏上。

$80,000 on a game.

好的。

Okay.

是的。

Yeah.

这是一款免费游戏。

And it's a free to play game.

所以如果你愿意在游戏上花这么多钱，你可能整天都泡在游戏里。

So if you're gonna spend this much money on the game, you probably spend also your whole day on it.

这对我来说是一个巨大的警醒。

And it was a huge wake up call for me.

我当时就想，哇，这用了这么多手段，却带来了我不确定是否认同的效果。

I was like, wow, this is using all these tools and is leading to an effect that I don't know that I feel good about.

我不确定这是我想与之关联的事情。

I don't know that this is something I want to be associated with.

但那些人却乐在其中。

And those people are happy doing that.

没有人强迫他们这么做，你可以说。

No one's making them, you could argue.

但我只是觉得我不想把时间花在产生这类结果上。

But I just didn't feel like I wanted to dedicate my time to generating outcomes of that sort.

所以我离开了。

So I left.

从那以后，我一直在思考：我们如何才能以更透明的方式，利用这些有趣而强大的工具，实现用户明确承诺并有意识追求的积极成果？

And then since then, I've been thinking, well, how can we use these really interesting, powerful tools in ways that are more explicit to users and also for good outcomes that they commit to, that they consciously commit to.

之后我去了MyFitnessPal，那是一家健康与健身公司，我在那里也应用了一些决策科学。

So I went to MyFitnessPal after that, which is a health and fitness company, and had some of that decision science being applied there.

后来我最终来到了Duolingo，在这里，这种理念已成为产品战略和公司战略的核心部分——用户使用产品时，清楚地知道它能为你完成两项任务。

And then eventually ended up at Duolingo, where it's a key part of the product strategy, of the company's strategy, and that customers or users come into the product knowing that it does two jobs for you.

它教你一门语言，同时也能让你保持动力。

It teaches you a language, but it also keeps you motivated.

所以，如果你不想要一个能让你保持动力的产品，市面上有成百上千个这样的产品可以选择。

So if if you don't want to have a product that keeps you motivated, there's hundreds of products that won't.

但Duolingo的独特之处在于，它能同时为你完成这两项任务。

But what makes Duolingo unique is that it does both of those jobs for you.

很有趣，正如你描述你在Zynga的经历时提到的，作为产品经理，我们的产品对人们生活的影响，往往对我们来说可能显得有些抽象。

It's interesting, as you described your experience at Zynga, one of the largest gaming companies out there, interesting reminder that the influence that we as product managers and our products can have on people's lives, you know, I think it's often our customers and the impact that we have on them can be a little abstract.

尤其是我的专长是小型企业软件，对吧？

Especially, my specialty is really small business software, right?

比如一些消费类健康应用，但它们同样对人们的生活产生影响。

Kind of business software, things like consumer health apps, But it matters in their lives too.

所以这很好地提醒了我们这些工具有多么强大。

So it's a really good reminder of how powerful these tools can be.

同时也提醒我们，它们终究只是工具，你可以用它们行善，也可以用它们作恶。

And also that they're just tools, and you can use your powers for good or evil.

所以，请做出你的选择。

So make your choice.

是的，我很感谢你这么说。

Yeah, I appreciate that.

所以，让我们看看，抱歉，我刚才稍微走神了。

So can you let's see, sorry, I've lost my train of thought here for a second.

你能具体分解一下，在Duolingo中，你协助推出过哪个基于决策科学原理的功能或用户体验吗？

Could you maybe break down like a feature or some user experience that you helped bring to market at Duolingo that was based on the principles of decision science?

为所有听众阐明一下这个过程和方法论。

Kind of illuminate the process and methodology for everyone listening.

是的，我首先会考虑的是，明确你希望在用户身上培养或强化的具体行为。

Yeah, so one of the first things I like to think about is specifically identifying what behavior you want to instill in the user or reinforce in the user.

你必须非常明确地定义这种行为，然后找出可能有助于培养或促进这种行为的各类游戏化机制或功能。

And you have to be very specific about what that is and then identify different gamification mechanics or different features that you think might help instill that behavior or foster that behavior.

通常有效的方法是那些非常直接、一目了然的方案。

And then usually what works is something that is very much on the nose, very direct.

例如，我协助在Duolingo中引入的一个功能就是排行榜。

So for example, one feature that I helped introduce into Duolingo is the feature of leaderboard.

这个功能的作用是，每周你都会被匹配到与你上周参与度相似的用户。

And what this does is that every week, you get matched with people who have a similar level of engagement as you in the last week, in the prior week.

然后你们互相竞争。

And then you compete with each other.

排名前十的用户会被提升到更高的联赛。

And the top 10 users get promoted to the next league higher up.

排名后十的用户会被降级到前一个联赛，而中间的用户则保持原位。

The bottom 10 users get demoted to a league before, and the middle users basically stay there.

这是一种经典的游戏机制。

And a classic gaming mechanic.

人们对此非常熟悉，这在你试图进行干预时非常有帮助，因为人们自然理解正在发生什么，因为他们曾在其他地方见过。

People are very familiar with it, which is very helpful when you're trying to make an intervention if people naturally understand what's going on because they've seen it somewhere else.

它所驱动的行为本质上激发了人们的竞争精神。

And the behavior that it drives basically taps into people's competitive spirit.

当人们竞争时，往往会更加努力地去做某件事。

And when people compete, they tend to push themselves harder at something.

这意味着几件事。

And that means a couple of things.

它可能意味着更高的持续性，也意味着在你使用产品的每一天都能获得更深的参与度。

It can mean greater consistency, but also greater depth of engagement on any given day that you use the product.

并且行为模式非常可预测：通常在一周开始时强度很高，到了周中，人们的投入度往往会下降。

And behaves in a very predictable way, Usually high intensity at the beginning of the week, then middle of the week, you tend to lose the commitment.

而到了周末，你会意识到如果错过某些东西，就会产生一定的错失恐惧感。

And then towards the end of the week, you realize that there's some certain FOMO impact if you're missing out.

等等，我离第三名这么近了。

Oh, wait, I'm so close to third place.

如果我再努力一点，就能冲到第一名。

If I push myself a little harder, I can get to first place.

我处在最后十名。

I'm at the bottom 10.

我必须确保自己不会被降级。

I need to make sure I don't get demoted.

因此到了周末，每个人的参与度都会重新燃起，竞争性也随之增强。

So towards the end of the week, everyone's engagement reignites and it increases competitiveness as well.

这基本上完全达到了预期效果。

What that did basically worked exactly as intended.

它提升了用户参与度。

It increased engagement.

我们有一个我觉得相当有趣的指标，那就是我们所说的高度活跃用户。

And we had a metric that I thought was pretty interesting, which was what we call highly engaged users.

这些用户每天至少花一小时、每周五天学习语言，我们认为这在时间投入上相当于一门不错的课程。

These are folks who spend at least one hour a day, five days a week learning the language, which we thought was equivalent to basically a coolish course in terms of time commitment.

我们看到这一指标增长了三倍，即有更多用户以如此高的频率使用学习内容，这让我们感到非常惊讶，因为这些原本就有学习意愿的人突然开始学得更多了。

And we saw that metric increase increase three x the number of users who were engaging that much with the with the learning content, which we thought was pretty amazing that all these folks who already had the intent to learn, suddenly they were learning a lot more.

你能否也将这个所谓的领先指标——即核心用户的参与度指标——直接与业务成果联系起来，比如留存率或公司的其他财务指标？

Were you able also to take that, I'll call it a leading indicator, that engagement metric for power users, and tie that directly to business outcomes, like retention rates or other kind of financial outcomes for the company?

是的。

Yeah.

是的。

Yeah.

我们还看到了留存率的提升。

We saw retention increase as well.

不是三倍，但增幅显著。

Not 3x, but a significant amount.

所有这些通常还会对变现和病毒传播等产生下游影响。

And all of that usually has downstream impact as well for monetization and virality and so forth.

还有一些无形的影响，比如人们开始在社交媒体上分享他们在排行榜上的成就，当不使用产品或已变得不活跃的人看到朋友在推特上分享时，这提高了他们对产品的认知度。

And there were some intangible things as well, like people are starting sharing on social media about their leaderboards' accomplishments, which increase kind of top of mind awareness for people who are not using the product or who became inactive using the product as they see their friends share on Twitter, for example.

是的。

Yeah.

我能理解这些行为影响的光环效应。

I could see those kind of halo effects of influencing people's behavior.

这很有趣。

That's interesting.

你提到过，人们在一周开始时参与度较高，然后出现下滑，之后又重新活跃。

So you mentioned this sort of weekly pattern of people engaging at the beginning of the week and then kind of a dip and then reengaging.

这从逻辑上讲是说得通的，但我很好奇你们是怎么发现这一点的。

Like, that makes sense logically, but I'm curious how you discovered that.

你们是通过观察产品中核心用户的现有行为，然后尝试引导更多人效仿这种行为来获得这一洞察的吗？

Did you gather that insight about the behavior based on looking at how power users in the product already behaved and then trying to influence more people to act like that?

还是基于你们在学校学到的关于人类行为的第一性原理？

Or was it based on first principles around things you learned in school about people's behavior?

理想情况下，你不会在中间看到下滑。

Ideally, you would not see a dip in the middle.

所以你们并不是刻意设计成在中间制造一个低谷。

So it wasn't like you were designing it to create a dip in the middle.

但我早就知道会是这样，因为我有游戏行业的经验。

But I knew it would happen that way because of my experience in gaming.

排行榜在游戏里很常见，它们总是呈现出这种参与模式。

Leaderboards are fairly common in games, and they always have that type of engagement pattern.

因此，当你引入排行榜时，你就会预期出现这样的结果。

So by introducing a leaderboard, you expect that type of result.

然后你可能会采取一些措施来提升一周中段的参与度。

And then you might do things to try to increase engagement in the middle of the week.

比如，你可以在周三设置一个欢乐时光，如果在指定的几小时内登录游戏，就能获得双倍积分，这样就能再次提升中段的参与度。

So you might have a happy hour on Wednesdays where, you get double the points if you play within x hours, you know, and that will drive engagement in the middle again.

好的。

Okay.

至于你用来与客户沟通、引导他们参与排行榜和产品的具体策略，我知道Duolingo是个应用，这些主要是推送通知和邮件吗？

And then as far as, like, the the actual, like, tactics that you used to communicate with your customers and drive them to interact with the leaderboard and the product, I know Duolingo is an app, so are these primarily push notifications, emails?

在这方面，你们尝试过哪些方法？

What did you experiment with on that front?

是的，以上所有方式都用过。

Yeah, all of the above.

此外，产品中还可能出现弹窗提示，比如‘你排名上升了’或‘你即将跌落’。

Also, there might be pop ups in the product saying, Hey, you moved up or you're about to go down.

这很有帮助。

That helps.

在你完成一节课后，如果你在排行榜上名次上升或下降，屏幕上会显示提示告诉你发生了什么。

At some point, after you did a lesson, if you moved up or down on the leaderboard, a screen would show up telling you what happened.

是的，有多种方式。

Yeah, different ways.

产品内部有多个沟通渠道，还有我们自己的渠道，比如邮件或推送通知，我们会对所有这些渠道进行实验。

There's several communication channels that the product has internally and also own channels like email or push notification, and we would experiment with all of them.

是的，因此你们将这一过程拆解为一系列实验，以优化与用户沟通的频率，避免过度打扰而耗尽他们的注意力。

Yeah, so then you broke it down as a series of experiments to try and optimize that balance of communicating with users but not over communicating and exhausting their attention span.

这样说对吗？

Does that sound right?

是的。

Yeah.

对，没错。

Yeah, exactly.

关于如何将事情拆分为实验，这是个很好的问题。

And it's a good question about how to break things up in experiments.

必须构建一个最小可行产品，使其能够以认知科学的方式运作。

And there's a minimum viable product that has to be put together to work in in a cognitive science type of way.

有时候，你拥有这些知识，但很难判断界限在哪里。

And sometimes, you have that knowledge, it's hard to know what where the lines are.

让我给你举个例子。

So let me give you an example.

你可以认为，作为最小可行产品，创建一个排行榜会容易得多——不是寻找同样活跃的用户并与之匹配，而是直接将你与你的朋友匹配。

You could say it would be much easier to create a leaderboard just as an MVP, where instead of finding users who are similarly engaged and matching you with them, we just match you with your friends.

如果我们一开始不设置每周竞争并重置的晋升系统，那会简单得多。

And it would be much easier if we didn't have a promotion system at first, which is competing every week and it just kind of resets.

这会是一个简单的最小可行产品，但它行不通，还会让你对排行榜的潜力产生错误的信号。

And that would be a simple MVP, but it wouldn't work, and it would give you the wrong signal about the potential for leaderboards.

原因在于，你的朋友可能活跃程度不同，因此你不会感到真正的挑战。

And the reason why is your friends might have different levels of engagement, so you don't feel really challenged.

所以，这种竞争元素实际上并不存在。

So that competition aspect is not really there.

如果你没有得到晋升，那就没有奖励，所以你会更快感到无聊。

If you're not being promoted, then you're not having a reward, so you'll get bored faster.

所以，如果你做了一个简化太多、无法产生你期望的认知效果的最小可行产品，你就无法获得关于更大构想的正确反馈。

So if you did this much, much more reduced MVP that is not having the cognitive effects that you were hoping for, you're not going to get the signal about the bigger idea either.

理解用户心理有助于你判断哪些内容必须包含在第一个版本中，哪些可以延后。

Understanding how user psychology works helps you identify where you draw the lines of what has to be in the first version versus what can wait.

对我们大多数人来说，我没有哈佛大学公共管理硕士学位，也没有机会学习决策科学。

For most of us, I will say, do not have masters in public administration from Harvard and have not had the opportunity to study decision science.

因此，对于希望基于决策科学原理思考自身影响力和用户行为的产品经理，你建议他们如何开始？

So for product managers who may want to start thinking about their influence and user behavior based on principles of decision science, how do you advise they get started?

他们可以从哪里进一步学习？

Where can they learn more?

这是个好问题。

It's a good question.

市面上有很多关于用户心理的书籍。

There's a lot of user psychology books out there.

有很多关于游戏机制的文章，但通常不是应用于普通产品时写的。

There's a lot of articles about gaming mechanics, usually not from applying it to a regular product.

它们实际上都来自游戏行业。

They're just actually from the gaming industry.

所以如果你从游戏化的角度思考决策科学，我建议你先广泛了解游戏，然后再尝试弄清楚如何最好地将其应用到非游戏场景中，这一点目前还没有明确的定义。

So if you're thinking about decision science from a gamification standpoint, there's I would just start learning about gaming generally and and then trying to figure out how to best apply it outside of a gaming setting, which is is not very well defined, I would say.

我认为目前还没有一套清晰的科学方法来指导这一点。

I don't think there's, like, a clear science on it yet.

所以保持好奇心。

So be curious.

多提问，多阅读各种资料。

Ask a lot of questions and and read a bunch of stuff.

是的。

Yeah.

并且多做实验。

And experiment.

是的。

Yeah.

对。

Yeah.

所以如何回到将这些应用到现有产品中的问题上。

So how kind of back into, you know, into trying to apply this into into existing products.

你提到为多邻国设计排行榜是一个影响行为的绝佳例子。

You mentioned like designing a leaderboard for Duolingo as a great example of a feature that influence behavior.

我从事产品管理很久了，学到了一些关于如何优先处理机会的经验。

I've been doing product management for a long time and I've learned a few things about where, how to prioritize opportunities.

对。

Right.

所以我总是希望优先考虑那些能带来广度或深度的机会——要么影响大量用户，哪怕程度有限；要么影响较少用户，但影响深远。

So I always want to prioritize things that will either have breadth or depth, affect a lot of users to some extent or, you know, a smaller number of users, but affect them deeply.

我很想知道，你是如何用决策科学的框架来寻找机会的？

And I'm curious how you look for opportunities with the framework of decision science.

这和一般优先考虑产品改进的机会有什么不同吗？

Is it any different than kind of prioritizing opportunities for product improvements in general?

我认为基本原理是一样的。

I think the fundamentals are the same.

你需要找到量化影响的方法，但它能为你提供更多方式来细分用户，并思考机会格局。

You want to find ways to quantify the impact, but it does give you more ways to segment your users and think about the opportunity landscape.

比如这个排行榜功能，它特别吸引某些用户，比如那些有竞争心的人，以及那些潜藏竞争欲望的人。

So for example, this leaderboard thing, it really appeals to certain users, like people who are competitive and people who have latent competitiveness.

尽管我推出了这个功能，但我在生活中大多数时候并不具有竞争性。

And even though I introduced this feature, I'm very not competitive in most of my life.

所以，登上排行榜榜首、进入下一个段位，对我来说完全没有动力。

So getting to the top of a leaderboard and getting to the next league has zero motivation for me.

这对我没有任何作用。

It doesn't do anything.

即使我是首席产品官，我也毫不介意自己处于倒数一两个段位。

Don't mind even though I was the CPO, I didn't mind being in the bottom one or two leagues.

这对我来说根本不是什么大事。

It's just not a thing that ticks for me.

但还有其他能激励人们的事情。

But there are other things that motivate people.

有些人希望获得更强的掌控感，这就是他们的动力。

Some people want to have a stronger sense of mastery, and that's what motivates them.

有些人非常注重协作和共同解决问题。

Some people are really into collaboration and solving problems together.

有些人喜欢社交，也就是随意聊天，不一定有明确目标，只是为了感到连接。

Some people are into socializing, meaning just casual conversation, not necessarily with a goal, other than just feeling connected.

如果你发现了一个不错的竞争机制，那就可以思考：我们能否设计出一个出色的掌控机制、社交机制或协作机制？

Then if you find that you have a good competition mechanic, well, then you know that could we create a good mastery mechanic, could we create a good socialization mechanic or a good collaboration mechanic.

如果你做得好，可能会看到类似的效果。

And if you do it well, you might see a similar impact from that.

好的。

Okay.

至于机会的规模，这和我们衡量其他所有事情的方式差不多。

So as far as sizing opportunities, it's kind of the same way we size everything else.

比如，目标受众有多大？

Like how big is the audience?

如果我们让某个指标提升了某个幅度，这会如何与我们关心的业务成果相关联呢？

And if we impact some metric by some amount, how does that correlate maybe to business outcomes that we care about?

比如留存率和转化率之类的东西。

Retention and conversion and things like that.

是的。

Yes.

我想说的是，有些规模估算更偏向理论性。

What I would say is some of the sizing is more theoretical.

我无法直接查看我产品的使用数据，来确定有多少人是喜欢协作的。

It's not like I can look at the data, usage data for my product, and know how many people are collaborative.

或者，如果有了协作功能，他们会变得愿意协作吗？

Or would it be collaborative if there was a collaborative feature?

但你可以从文献和科学研究中了解到，大致有多少人更倾向于合作而非竞争。

But you can know from the literature, from the science, that there is roughly this amount of people who prefer to be collaborative versus competitive.

因此，基于这一洞察，你可以选择优先开发这类功能。

So based on that insight, you can choose to prioritize features of that sort.

很有趣。

Interesting.

你们如何进行用户研究？

And how do you do user research?

你提到你在多邻国也做过用户体验研究。

So you mentioned that you ran UXR, user experience research at Duolingo as well.

在这个背景下，用户研究有什么不同？

How is user research different in this context?

这是个非常好的问题。

Yeah, it's a really good question.

幸运的是，我们聘请了一位非常优秀的领导者来带领这个团队。

Luckily, we were able to hire a really good leader to lead that team.

所以我不得不自己弄清楚所有这些。

So I then had to figure all this out by myself.

整个团队其实非常强大。

The team overall was a really strong team.

当你从整体上看待Duolingo时，并不是每个问题、我们构建的每项功能都是游戏化功能。

When you think about Duolingo as a whole, not every problem, not everything we built was a gamification feature.

还有一些非常传统的学习功能或用于变现的转化功能。

There were very traditional learning features or conversion features for monetization.

因此，部分用户体验研究相当传统或标准，和其他产品中看到的差不多。

So some of the user experience research was fairly traditional or standard, what you would see in other products.

在涉及游戏化内容时，我认为不同问题的难易程度存在一个梯度。

When it came to gamification things, I would say there's a scale of how hard it was to answer certain questions.

最容易回答的是关于可用性的问题。

And the easiest thing was around usability.

你设计好后，把它展示给用户，他们是否能理解？是否喜欢？

You design it, you put it in front of users, do they understand it, Do they like it?

他们有什么顾虑？

What concerns do they have?

他们能顺利完成这些流程吗？

Can they work through the flows?

所以，这同样是比较标准的。

So again, that was fairly standard.

然后还有一些更难的问题，比如先验性问题。

And then there are other things that will be much harder, like a priori.

你能说，这会让我感觉更有参与感、更有竞争性吗？

Could you say, will this make me feel more engaged, more competitive?

人们可能会这么说，但很难知道其中有多少是真实的，又有多少能在数周的竞争中持续存在，这基于用户研究。

People can say that, but it's hard to know how much of that is true and how much of that persists over time, over weeks and weeks of competition based on UXR.

意识到有些问题可能无法通过这种方式得到解答，这是我必须面对的现实。

Understanding that some of those questions might not be able to be answered that way was just a reality that I've had to deal with.

我确实有过这样的用户研究经验：用户希望如何看待自己，或希望被如何看待，可能与他们的实际行为并不一致。

I've certainly had the experience with user research where the way that the user wants to think about themselves or wants to see themselves may not actually match with how they behave.

所以它可能会

So it can

以这种方式带来一点危险。

be a little dangerous that way.

我想知道你是如何看待这个问题的，或者如何加以控制。

I'm curious how you think about that or control for it, maybe.

是的，确实就是这样。

Yeah, that is definitely how it is.

对吧？

Right?

而且，根据你所探讨的问题或主题不同，他们所说的自我认知与实际行为之间的差距可能会变小或变大。

And Depending on the question or the topic that you're addressing, that gap between what they say about themselves and how they behave might become smaller or larger.

每当你思考任何可能受到这种游戏化机制或认知偏差影响的事情时，这种差距都会扩大。

Whenever you're thinking about anything that can be influenced by this type of gamification mechanics or cognitive biases, that gap just gets larger.

这时你就必须更多地依赖自己的经验、直觉，寻找低成本实验的方法来获取真实数据，同时参考行业内的文献或通用研究。

That's where you have to rely a lot on your own experience, your instinct, finding ways to make low cost experiments to get real data and literature or general research from the industry as well.

另外，看看其他产品以及他们所做的实验，了解它们的成功程度。

Also, looking at other products and experiments that they've done and see how successful they are.

对于一位可能希望基于这些第一性原理明确影响行为的产品经理来说，尤其是那些在极度数据驱动的公司工作的人，对吧？

For a product manager who may be interested in explicitly trying to influence behavior in this way based on these first principles, and might be working at a company that's extremely data driven, right?

因为这是我们常挂在嘴边的一句话，对吧？

Because that's one of our catch words, right?

你们是数据驱动的吗？

Are you data driven?

每件事都需要有数据支持。

And everything needs to have data behind it.

如果你有一个关于如何改变体验以影响行为的假设，但你并没有确凿的数据来支持它，因为你提到过应用研究、第一性原理或你自己的直觉。

And if you've got a hypothesis around changing the experience in a way that'll influence behavior, but you don't necessarily have hard data to back it up, because you've mentioned applying research or first principles or your own instincts.

你有什么建议，可以如何向利益相关者或团队清晰表达并说服他们接受这种观点？

Any advice on how to articulate and sell that maybe to stakeholders or the team and get people bought in?

是的，这是个好问题。

Yeah, it's a good question.

正如你所说，这非常依赖于文化。

Like you said, it's so dependent on the culture.

我认为归根结底，如果你有机会进行A/B测试，A/B测试会告诉你它是否有效。

And I think that at the end of the day, if you have the chance to AB test, the AB test will tell you whether it worked or not.

所以最终你还是会做出一个数据驱动的决策。

So you will make a data driven decision at the end of the day.

问题是，你该如何获得进行A/B测试的许可呢？

The question is, how can you even get the permission to make the AB test, I guess?

我认为通常很有效的一种方法是，你能指出其他正在这样做并且取得成效的公司。

I think something that usually works pretty well is when you can point at someone else who's doing it and it's working for them.

你可以说，比如看看动态面板或打卡记录之类的例子。

You can say, I don't know, take the example of either boards or streaks or whatever.

可以说，有四家公司采用了这种方法，这是来自这四家公司五位用户对它们使用频率的评价。

Say, These are four companies that have it, and this is commentary from five different users from those companies about how much they use it.

这是来自Twitter上关于他们对此评价的截图。

And this is screenshots from Twitter about what they're saying about it.

而且有足够的证据表明，投资我们自己的测试是合理的。

And there's enough evidence that it makes sense to invest in a test of our own.

这背后的主题是如何培养一种以证据为导向而非仅仅以数据为导向的文化。

And the the underlying theme of this is how can you develop a culture that's evidence driven, not just data driven?

我所说的‘数据’是指来自你们自己产品的使用数据，因为来自你们自己产品的使用数据只是获取证据的一种方式。

And what I mean by data is usage data from your own product, because usage data from your own product is just one way to get evidence.

但这并不是唯一的方式。

But it's not the only way.

一家仅依赖自身数据做决策的公司，会错过太多其他获取证据以改进决策过程的方式。

And a company that only uses their own data to make decisions is missing out on so many more ways to get evidence into your decision making process.

通常来说，你希望帮助公司拓展其信息来源的广度。

Generally, you want to help your company expand its horizons of where it gathers information from.

我认为在这里一个富有成效的做法是，研究那些拥有良好用户体验但不一定是直接竞争对手的成功公司。

I could see where this would be a fruitful activity here would be to look at successful companies that have good user experiences that aren't necessarily direct competitors.

我确实在很多公司都见过这种情况，即竞争分析仅限于最直接的竞争对手。

This is something I've certainly seen at a lot of companies where competitive analysis is limited to the most direct competitors.

我们却局限于那个范围之内。

And we look outside of that and limit yourself there.

但其他行业可能有效的方法，也可能适用于你的产品。

But techniques that might be working in other sectors could work for your product too.

创造性地思考如何将这一原则应用到你的受众身上，可能会带来丰厚回报，也相当有趣。

And thinking creatively about how to apply that principle to your audience could be rewarding and also kind of fun.

是的。

Yeah.

对。

Yeah.

当然。

Absolutely.

我完全同意你的说法。

Totally agree with what you're saying.

在多邻国，我们一直坚持这样做。

And at Duolingo, we did that consistently.

对吧？

Right?

你会发现，针对不同类型的问题，你会参考不同的公司。

And you'll find that for different types of issues you're trying to resolve, you will look at different companies.

例如，关于变现，我们经常研究约会应用和Spotify，尽管它们与语言学习完全无关，但我们认为它们在变现方面做得很好。

So, for example, for monetization, we often looked at dating apps and Spotify, for example, which were completely unrelated to learning languages, but we felt did a good job monetizing.

它们有付费产品，因此这足够接近，能帮助我们获得变现的灵感。

And they have premium products, so that was close enough to help us get ideas for monetization.

至于增长，我们经常研究那些完全没有学习价值的游戏，但它们非常擅长吸引用户。

And for growth, we often looked at games that have no learning value whatsoever, but they knew how to engage users.

因此，我们可以借鉴这些想法，并将其应用到学习场景中。

So we could borrow ideas from that and translate it into a learning setting.

所以，是的，以这种方式拓展视野也很有趣，当然。

So yeah, and it's fun to expand your horizons that way too, for sure.

正如我提到的，我负责中小企业产品。

I work on, as I mentioned, I work on small business products.

当然，我们的受众中有许多人从事企业级或B2B产品，您也提到您曾经在B2B领域工作过。

Certainly, our audience has a lot of people that work on enterprise or B2B, you've mentioned you had a stint in B2B as well.

是的。

Yeah.

您能否帮助那些不从事您所说的高频消费类产品的人，思考一下如何应用这种思路？

Can you help those of us who aren't working in your kind of frequent use consumer products think a little bit about how we could apply this thinking?

当然可以。

For sure.

对。

Yeah.

对。

Yeah.

我有几点想法。

So a few thoughts.

首先，现在很多B2B产品都是产品驱动增长的模式。

One is a lot of b to b products now are product led growth shops.