The Exchanges

Q This is a take home task. Okay. And so that's what good looks like. What would some red flags be? We, we can have that as green flags. Are there anything where you'd be like, yeesh, didn't like that?

A The biggest red flag to me is when I hear a heavy bias towards experimentation. So maybe requiring every change made in the product to be an experiment or over focusing on optimizing, adding friction, removing friction, uh, raising prices, reducing prices, um, versus making substantial changes to the core product experience itself. And the reason why this is a red flag to me is it suggests an over-focus on needing experiments to validate judgment. Where in early companies actually having the right judgment before you make an investment is important so that you get return on how you spend your time. And also it would suggest to me an over-focus on optimization rather than an ability to think more critically about how the product serves the user and how value is derived from that. So ideally things like new features might even be something that a growth leader could suggest or contribute because they have a thesis On how that feature will help the business rather than just saying, you know, let's add a couple tool tips here or, um, a couple buttons over there.

AI assessment note: “The biggest red flag to me is when I hear a heavy bias towards experimentation.”

Q I love that kind of separation of the two. How often is it the first? How often is it the second?

A I would say rule of thumb is for growth teams, maybe a 20% batting average. So you'll see metric wins or hit that goal 20% of the time. Um, now that's a little different than, uh, how to set a goal that's sufficiently ambitious. So, uh, when my teams set goals, we usually aim for a 70% hit rate, meaning that if there were 10 teams in the growth org, seven of them Probabilistically would hit the goal that they set. Um, and you know, the other 30% wouldn't, and that's by design to make sure that goals are both ambitious and pragmatic. And it also means that then the team needs to have so many ideas that even if only 20% of them contribute to that goal, That will be enough to get to that metric.

AI assessment note: “for growth teams, maybe a 20% batting average... aim for a 70% hit rate”

Q I love that. Thank you for answering what was the very base and really terrible question. Um, what are some other very common mistakes you see startups make when it comes to user onboarding? I'm sure you go through onboarding flows and you're like, oh no. What are some that you see very often?

A There are a couple that come to mind. I mentioned briefly that I generally don't like tooltips because they're basically a quick wall of text, usually defining a bunch of terms that the user doesn't understand and will continue to not understand. The other pattern that I don't love is a checklist or a more passive opt-in. So you could imagine you sign up for a product tomorrow and something pops up that says do these 10 things. And if you choose to do them, you know, you will be all set up to use this product. Um, the reality is that something passive like that, where you can choose to engage with educational material, is probably only going to have about a 10% opt-in rate, so 90% of your customers will never actually learn any of those things. And this means that actually if you want effective onboarding, you have to be more assertive, more opinionated, and actually Ensure that 90% of your customers are completing that educational material. Now, there's a balance there, because you have to pick the right material that's relevant to most people, but in general, a checklist is just not going to get you the breadth of engagement that is needed to really move the needle on activation.

AI assessment note: “I generally don't like tooltips because they're basically a quick wall of text”

Q What are the biggest differences? Because in my mind, I would be like, oh, Great. They bring a consumer lens, they bring the creativity and virality of consumer growth to B to B, which is more stodgy and sticky. Talk to me about the differences and why that may be not the case.

A So if you think about Facebook or Instagram. A growth tactic that may be effective in either of those apps would be making a critical action more prominent, making a button bigger, making stories appear at the top of your Instagram feed instead of in the middle, which they used to, if you may remember, um, reducing friction. Maybe there's only two steps to complete signup instead of three. And generally those optimizations will yield Relatively meaningful changes in behavior for a consumer user base for whom there's more of a choice to be using that product today or tomorrow, and little differences in effort required might actually make the difference on if they take an action. In SaaS, often a tool is used at work. So if my boss asks me to send them A notion doc to use notion to deliver something. I will do it. And that means that a slightly bigger button or a slightly shorter signup flow is actually not going to move the needle on if I use a tool. Um, there's of course some incredible delight and value that comes from consumer grade product experiences in SAS. But I find that in consumer growth, there tends to be an over-focus on very small detail changes. Whereas in SaaS, it's more important to make bigger or bolder changes that really react to the user's intent so that their behavior will change.

AI assessment note: “In SaaS, often a tool is used at work... slightly bigger button is actually not going to move”

Q You mentioned this quite big change, really, to how Notion really organizes content and delivers it, and what the content is that it provides. Postmortems are fascinating to me. Fascinating. How do you do postmortems? Who's invited? Who sets the agenda? How often are they? Can you just take me into the world of postmortems and how you do them?

A Yes. We, we do them all the time. We call them retros. Postmortem is also great. And, uh, they're recursive. So it could be that a small team, maybe even a pair, a PM and an engineer who worked on something will retro together. It could be a team. It could be an org. It could be at the company level. Uh, but generally for something substantial, like a new launch, I would allow for time to see results and then Hold of meeting where everybody can share their Rosebud Thorne equivalent, you know, what went well, what we could do better, what didn't go well, and also assess and understand the performance of that launch and feel and follow through on accountability for if the original hypothesis in launching that feature delivered, and also what surprised us about the behavior we saw from customers and the business impact that we saw on the other side. This is something that can be pretty second nature in growth because we are so close to the data, but I find that that focus on metrics, on results, on impact can sometimes be less central in the rest of product doors and in places that are further away. Um, and it's a shame because there's so much you can learn from saying, Hey, we launched this feature. What did we expect to see in terms of usage, in terms of engagement, in terms of conversion, behavior changes?

AI assessment note: “We call them retros... they're recursive. So it could be that a small team”

Q I love that kind of separation of the two. How often is it the first? How often is it the second?

A I would say rule of thumb is for growth teams, maybe a 20% batting average. So you'll see metric wins or hit that goal 20% of the time. Um, now that's a little different than, uh, how to set a goal that's sufficiently ambitious. So, uh, when my teams set goals, we usually aim for a 70% hit rate, meaning that if there were 10 teams in the growth org, seven of them Probabilistically would hit the goal that they set. Um, and you know, the other 30% wouldn't, and that's by design to make sure that goals are both ambitious and pragmatic. And it also means that then the team needs to have so many ideas that even if only 20% of them contribute to that goal, That will be enough to get to that metric.

AI assessment note: “rule of thumb is for growth teams, maybe a 20% batting average.”

Q fire round at the end, but there's like several amazing companies here. That I want to understand what you took from them in a relatively condensed timeframe. And so we're going to do a quick fire on a lesson from each company and how it impacted your mindset. And so I'm going to start in chronological order. Dropbox, what was the biggest lesson and how did it impact your mindset?

A Hmm. My biggest lesson from Dropbox was probably that growing a business is a game of inches. We worked on trial conversion, helping folks convert to premium plans after trying Dropbox with their team for the first time. And I understood just how much impact could compound from ensuring that different markets, different payment methods, different customer profiles could all experience Dropbox and get value out of that experience and going really, really deep in the data to figure out where Different pockets of customers weren't converting as well as they could be. Uh, I don't think I really understood how much return you can get simply from being precise and rigorous and detail oriented in ensuring every customer is getting maximum value. And I really learned that there.

AI assessment note: “My biggest lesson from Dropbox was probably that growing a business is a game of inches.”

Q I love that. Thank you for answering what was the very base and really terrible question. Um, what are some other very common mistakes you see startups make when it comes to user onboarding? I'm sure you go through onboarding flows and you're like, oh no. What are some that you see very often?

A There are a couple that come to mind. I mentioned briefly that I generally don't like tooltips because they're basically a quick wall of text, usually defining a bunch of terms that the user doesn't understand and will continue to not understand. The other pattern that I don't love is a checklist or a more passive opt-in. So you could imagine you sign up for a product tomorrow and something pops up that says do these 10 things. And if you choose to do them, you know, you will be all set up to use this product. Um, the reality is that something passive like that, where you can choose to engage with educational material, is probably only going to have about a 10% opt-in rate, so 90% of your customers will never actually learn any of those things. And this means that actually if you want effective onboarding, you have to be more assertive, more opinionated, and actually Ensure that 90% of your customers are completing that educational material. Now, there's a balance there, because you have to pick the right material that's relevant to most people, but in general, a checklist is just not going to get you the breadth of engagement that is needed to really move the needle on activation.

AI assessment note: “I generally don't like tooltips because they're basically a quick wall of text”

Q in terms of, kind of, that user experience. I know that this is, kind of, one area where you're world class, to be precise, from one of your colleagues, I can't remember who was telling me, yes, truly world class at great user onboarding. Um, wonderful references, by the way. Um, my question to you is, what makes truly great user onboarding, and how do you think about that today?

A User onboarding is my specialty. I'm glad you asked. Um, Onboarding is a term that's used very broadly, and there are many tools that you can buy to help you, or that claim they can help you with onboarding, where you help orient a user to your product with the hopes that they are more likely to be more successful. The metric that many companies use to represent this is activation rate. So the percentage of folks who sign up for your product, who then ultimately are expected to retain long term. And usually there's two dimensions on which activation is defined. The first is either engagement or retention. So did Harry sign up today and keep coming back tomorrow, next week, and next month? But then also sophistication or early habit building. So did Harry also take an action that was high value? Did he create a document in Notion that was meaningful? Did he build something in Airtable that was useful? And those together indicate that someone has found value. They keep coming back because they see something that they want to come back for. And also that they're doing the right things to experience that recurring value over time and uncover and unlock the magic of what your product offering is. So to me, onboarding is one of the most important things you can invest in. Interestingly, it can be very hard to move that activation rate metric. So unlike a lot of growth experiments whe…

AI assessment note: “you're in the business of helping people learn more effectively”

Q That is hilarious. I did hear from many of your colleagues about your data centricity. So we're going to get to that later. Um, but sticking on the theme of the quick fire, Blue Bottle. I mean, this is a cool addition to the roster. What was the biggest lesson from Blue Bottle, and how did it impact your mindset?

A So at Blue Bottle, I focused on our e-commerce business, and the primary product we offered was a subscription where we mailed fresh whole coffee beans to your home, recurring, so that you could make Blue Bottle coffee at home instead of traditionally going to the cafe. And what I learned with my team in endeavoring on this journey was actually the importance of knowing your customer. We felt at first that The customers most likely to want blue bottle coffee at home would be the same customers who go to our cafes. And what we found in actually visiting Many of our online customers is that they were completely different and that someone who's making coffee at home every morning isn't going to the cafe that day. They actually have a fully different routine. They tend to be more of a hobbyist, uh, enjoy the craft of the morning ritual of brewing coffee at home. And that unlock really helped us focus our marketing, our product offerings, what we were building for Those of those people who were interested in the e-commerce business rather than retail. Um, it's actually something that I've carried with me that in general, you should always have evidence and insights to inform what you work on and what you prioritize on a growth team or any team. And in talking to customers, we understood the why behind some of the trends we were seeing.

AI assessment note: “what I learned with my team in endeavoring on this journey was actually the importance”

Q What would an example of that be when you said about the correlation and causation? Just so I understand what that actually means in real life.

A I think a great example to call on Airtable would be if a creator of a new base, which is the spreadsheet interface that's central to the Airtable product, were to build something of substance, so there's some number of entries in that spreadsheet, and bring a collaborator into that base to work with them. That, to me, would be a signal, and I've done the math, so I can tell you that's correct, um, that they're much more likely to keep using Airtable six months from now, because they've done substantial work, they've brought in a collaborator in a product that's designed for teams, And together, those two people are working on something that they can ultimately grow into a more sophisticated deployment of the product over time.

AI assessment note: “I think a great example to call on Airtable would be”

Q So what should I have? Cause like, I'm just thinking Google Analytics. You're gonna be upset with me for that. So like, what should I have in place where you're like, good, we've got the, we've got the foundations. I can do my job now.

A Well, it depends a little what you're trying to do. If the goal of bringing someone in to work on growth is really to help with acquisition at the top of the funnel, that's more growth marketing, uh, then maybe Google analytics is enough to get started to work on SEO and SEM. Um, if more it, you know, if more is needed or, or if you're interested in doing more, including stuff in the product, maybe, uh, experimenting with your pricing model, onboarding, Uh, in general, some degree of data engineering is necessary. So actually having somebody in seat who can help set up the initial data pipelines and logging for you to be able to understand if changes made to the product are good for users and good for the business is necessary. Um, some sort of, uh, analyst facing tool could be amplitude, could be mode, could be hex, um, is, You know, necessary to make sure that what you are learning can also be digested by and acted on by whoever's working on growth. Um, and then ideally an experimentation tool like Eppo, uh, to round that out, uh, and make sure that, you know, you're able to actually try stuff and learn from it.

AI assessment note: “having somebody in seat who can help set up the initial data pipelines and logging”

Q What are the biggest differences? Because in my mind, I would be like, oh, Great. They bring a consumer lens, they bring the creativity and virality of consumer growth to B to B, which is more stodgy and sticky. Talk to me about the differences and why that may be not the case.

A So if you think about Facebook or Instagram. A growth tactic that may be effective in either of those apps would be making a critical action more prominent, making a button bigger, making stories appear at the top of your Instagram feed instead of in the middle, which they used to, if you may remember, um, reducing friction. Maybe there's only two steps to complete signup instead of three. And generally those optimizations will yield Relatively meaningful changes in behavior for a consumer user base for whom there's more of a choice to be using that product today or tomorrow, and little differences in effort required might actually make the difference on if they take an action. In SaaS, often a tool is used at work. So if my boss asks me to send them A notion doc to use notion to deliver something. I will do it. And that means that a slightly bigger button or a slightly shorter signup flow is actually not going to move the needle on if I use a tool. Um, there's of course some incredible delight and value that comes from consumer grade product experiences in SAS. But I find that in consumer growth, there tends to be an over-focus on very small detail changes. Whereas in SaaS, it's more important to make bigger or bolder changes that really react to the user's intent so that their behavior will change.

AI assessment note: “in SaaS, it's more important to make bigger or bolder changes”

Q What's the biggest and boldest change you've made, do you think?

A Well, just last week, my growth team at Notion launched a template gallery, relaunched our template gallery, and released more than 4500 new templates for what you can do on Notion. This is a pretty big change for us. We had only about 500 before the launch, and in addition to that, we also feature the creators who are making them. So in the case where Notion has not made the template, anybody else can, and you can actually present it and share it right from Notion's website, which is a celebration for us of our creators, and also a bet on leaning into the horizontal nature of the product and celebrating Student use, personal use, hobbyist use, and also professional use. So as one example, we have more than 100 habit trackers in our template gallery now, which is pretty cool that, you know, if you want to track your habits in Notion, you can actually shop through many different ways to do that and decide which one you like best.

AI assessment note: “my growth team at Notion launched a template gallery, relaunched our template gallery”

Q How does that change the person that you have?

A So in the first camp where you're really focused on acquisition, the right archetype of a hire is growth marketing. This is someone who's worked in data-driven marketing, who understands the dynamics of key data-driven campaigns. In some cases, maybe has experience with community-led growth or content marketing, depending on exactly, you know, what you think might work well, given what you know or what you don't. Um, in the second case, You're looking for someone who is more comfortable interfacing with or is an engineer. And the reason for that is that actually meaningfully changing the code base of the core product is necessary to To have the change that you're seeking. Um, in general, it's hard in growth because LinkedIn might just say growth and it's not clear if that candidate is someone who's got more of a marketing background or more of a product and engineering background, but the work they're experienced in will vary.

AI assessment note: “in the first camp where you're really focused on acquisition, the right archetype”

Q That is hilarious. I did hear from many of your colleagues about your data centricity. So we're going to get to that later. Um, but sticking on the theme of the quick fire, Blue Bottle. I mean, this is a cool addition to the roster. What was the biggest lesson from Blue Bottle, and how did it impact your mindset?

A So at Blue Bottle, I focused on our e-commerce business, and the primary product we offered was a subscription where we mailed fresh whole coffee beans to your home, recurring, so that you could make Blue Bottle coffee at home instead of traditionally going to the cafe. And what I learned with my team in endeavoring on this journey was actually the importance of knowing your customer. We felt at first that The customers most likely to want blue bottle coffee at home would be the same customers who go to our cafes. And what we found in actually visiting Many of our online customers is that they were completely different and that someone who's making coffee at home every morning isn't going to the cafe that day. They actually have a fully different routine. They tend to be more of a hobbyist, uh, enjoy the craft of the morning ritual of brewing coffee at home. And that unlock really helped us focus our marketing, our product offerings, what we were building for Those of those people who were interested in the e-commerce business rather than retail. Um, it's actually something that I've carried with me that in general, you should always have evidence and insights to inform what you work on and what you prioritize on a growth team or any team. And in talking to customers, we understood the why behind some of the trends we were seeing.

AI assessment note: “what I learned with my team in endeavoring on this journey was actually the importance”

Q for you, the challenges with notion with air table, these are horizontal tools that are incredibly powerful, but to many different users with many different needs, the way that I as a podcaster turned investor is very different. And even in my job compared to how someone else uses it in their job, how do you Have a horizontal user onboarding when the needs and value derived is very different?

A Yes, very, very good question and hard to do right. Something that we do at Notion that I think generally, uh, is tried and true is to ask someone what they're here for. Many businesses will actually cast or assume or segment that a user is a certain kind of user. Maybe they sign up with a Gmail account, and one might assume that they're there for a personal use case, but it's possible that They used their Gmail, but are actually evaluating your tool or your product to, um, deploy at their business. And just starting with that simple question of, are you here to do work with your team? Are you here as a student? Are you here for personal projects? Can really help you as the builder of the product, understand the value the customer is seeking and reflect it back to them. And then Demonstrating value is more of an exchange. So when you ask a customer to tell you something about them, you can meaningfully change the product experience and reflect that back and help them see that this product can be relevant and specific to their use case and that, in fact, the product will help them get there. So there might be templates or examples or, um, different customizations that the product can make for them to help them Get there along the way. As, um, as one example, if I signed up for Notion today and said that I was a student, it wouldn't really make sense for me to see tasks and sprin…

AI assessment note: “Something that we do at Notion... is to ask someone what they're here for.”

Q What would an example of that be when you said about the correlation and causation? Just so I understand what that actually means in real life.

A I think a great example to call on Airtable would be if a creator of a new base, which is the spreadsheet interface that's central to the Airtable product, were to build something of substance, so there's some number of entries in that spreadsheet, and bring a collaborator into that base to work with them. That, to me, would be a signal, and I've done the math, so I can tell you that's correct, um, that they're much more likely to keep using Airtable six months from now, because they've done substantial work, they've brought in a collaborator in a product that's designed for teams, And together, those two people are working on something that they can ultimately grow into a more sophisticated deployment of the product over time.

AI assessment note: “I think a great example to call on Airtable would be if a creator”

Q In terms of learning from other people, for a lot of founders listening, they're thinking, great, I'd also have to run, you know, a sales team, a CS team, a product team, and actually this is a growth function in a way, um, I should probably need to hire for growth. When's the right time to hire for growth, do you think?

A It does depend on your go-to-market. So if a business decides that a product-led go-to-market is the right approach in the early days, I would have somebody, usually an engineer, though it could be a technical PM, focused on growth earlier than in other businesses. And the reason for this is that you can actually deliver more scaled impact early, As you're developing that initial motion. Um, if, you know, I, I guess if PLG is not necessarily the crux of go to market or you're still figuring it out, um, the rule of thumb is to introduce someone working on growth after there's product market fit and to invest more if you see good return on that initial investment of bringing someone into work on growth. But it does depend and I've seen some Companies bring in one person, usually an engineer who's passionate about the area, pretty early to set up the scaffolding, make sure experimentation is possible, do some of the initial data logging, try out initial onboarding, and that has good return.

AI assessment note: “the rule of thumb is to introduce someone working on growth after there's product market fit”

Q So what should I have? Cause like, I'm just thinking Google Analytics. You're gonna be upset with me for that. So like, what should I have in place where you're like, good, we've got the, we've got the foundations. I can do my job now.

A Well, it depends a little what you're trying to do. If the goal of bringing someone in to work on growth is really to help with acquisition at the top of the funnel, that's more growth marketing, uh, then maybe Google analytics is enough to get started to work on SEO and SEM. Um, if more it, you know, if more is needed or, or if you're interested in doing more, including stuff in the product, maybe, uh, experimenting with your pricing model, onboarding, Uh, in general, some degree of data engineering is necessary. So actually having somebody in seat who can help set up the initial data pipelines and logging for you to be able to understand if changes made to the product are good for users and good for the business is necessary. Um, some sort of, uh, analyst facing tool could be amplitude, could be mode, could be hex, um, is, You know, necessary to make sure that what you are learning can also be digested by and acted on by whoever's working on growth. Um, and then ideally an experimentation tool like Eppo, uh, to round that out, uh, and make sure that, you know, you're able to actually try stuff and learn from it.

AI assessment note: “analyst facing tool could be amplitude... ideally an experimentation tool like Eppo”

Q This is a take home task. Okay. And so that's what good looks like. What would some red flags be? We, we can have that as green flags. Are there anything where you'd be like, yeesh, didn't like that?

A The biggest red flag to me is when I hear a heavy bias towards experimentation. So maybe requiring every change made in the product to be an experiment or over focusing on optimizing, adding friction, removing friction, uh, raising prices, reducing prices, um, versus making substantial changes to the core product experience itself. And the reason why this is a red flag to me is it suggests an over-focus on needing experiments to validate judgment. Where in early companies actually having the right judgment before you make an investment is important so that you get return on how you spend your time. And also it would suggest to me an over-focus on optimization rather than an ability to think more critically about how the product serves the user and how value is derived from that. So ideally things like new features might even be something that a growth leader could suggest or contribute because they have a thesis On how that feature will help the business rather than just saying, you know, let's add a couple tool tips here or, um, a couple buttons over there.

AI assessment note: “The biggest red flag to me is when I hear a heavy bias towards experimentation.”

Q What's the biggest and boldest change you've made, do you think?

A Well, just last week, my growth team at Notion launched a template gallery, relaunched our template gallery, and released more than 4500 new templates for what you can do on Notion. This is a pretty big change for us. We had only about 500 before the launch, and in addition to that, we also feature the creators who are making them. So in the case where Notion has not made the template, anybody else can, and you can actually present it and share it right from Notion's website, which is a celebration for us of our creators, and also a bet on leaning into the horizontal nature of the product and celebrating Student use, personal use, hobbyist use, and also professional use. So as one example, we have more than 100 habit trackers in our template gallery now, which is pretty cool that, you know, if you want to track your habits in Notion, you can actually shop through many different ways to do that and decide which one you like best.

AI assessment note: “relaunched our template gallery, and released more than 4500 new templates”

Q How does that change the person that you have?

A So in the first camp where you're really focused on acquisition, the right archetype of a hire is growth marketing. This is someone who's worked in data-driven marketing, who understands the dynamics of key data-driven campaigns. In some cases, maybe has experience with community-led growth or content marketing, depending on exactly, you know, what you think might work well, given what you know or what you don't. Um, in the second case, You're looking for someone who is more comfortable interfacing with or is an engineer. And the reason for that is that actually meaningfully changing the code base of the core product is necessary to To have the change that you're seeking. Um, in general, it's hard in growth because LinkedIn might just say growth and it's not clear if that candidate is someone who's got more of a marketing background or more of a product and engineering background, but the work they're experienced in will vary.

AI assessment note: “in the first camp where you're really focused on acquisition, the right archetype”

Q Have there been any really tangible takeaways, either from companies that've really worked, that really haven't worked, where you're like, ooh, that was an interesting insight from that expect, like from that moment, or not experiment, but that journey that I didn't think about before?

A One of the things that I think I've become much crisper on is if you're focused on top of funnel. So if you're focused on bringing more people into your product, being thoughtful about which strategies will yield the right customers that fit the profile of your ICP or your ideal customer profile. And, uh, an example of this would be if you have an amazing YouTube strategy or community strategy, That will bring in lots of passionate folks who want to give your product a try, but it may not actually be the way to get a large enterprise to discover who you are or to learn what you offer. So being smart and thoughtful about that, where actually your first touch and getting to know a customer will, you know, bring in a certain profile of customer that ideally you should then be able to serve all the way down funnel and your go-to-market should match. Um, that, that's something that I feel I've gotten a lot sharper on and working with companies that are earlier and are trying many different things to, to see what works.

AI assessment note: “One of the things that I think I've become much crisper on is”

Q Having invested in emerging markets, I completely agree with you on that statement. So I feel, I feel the pain there. Final one, and then we're going to dig into some more of the nitty gritty, but Airtable most recently, how did that impact your mindset, and what are the one or two biggest lessons?

A Yes, so, Airtable is an area where I started to really focus on product-led growth for SaaS. I had seen a bit of it at Dropbox much earlier, but this was more of a recent chapter, and PLG, product-led growth, has really taken off in the past few years. Something that I think we did really well at Airtable is thinking of PLG, or that product-led motion, and the sales-led motion or sales-assisted motion as being part of one engine. For the business. I find that commonly businesses will say, well, I need to pick one. Either I'm a product led business or a sales led business, and we were actually able to work the two in concert and make sure that when folks gave Airtable a try, We were actually thinking through the full journey of that customer over time and when it would actually be the right moment for them to graduate from a free plan or small team use into something more sophisticated or a more complex deployment. This is something that to me feels under understood, if you will, um, that we did pretty well at Airtable.

AI assessment note: “thinking of PLG, or that product-led motion, and the sales-led motion... as being part of one engine”

Q for you, the challenges with notion with air table, these are horizontal tools that are incredibly powerful, but to many different users with many different needs, the way that I as a podcaster turned investor is very different. And even in my job compared to how someone else uses it in their job, how do you Have a horizontal user onboarding when the needs and value derived is very different?

A Yes, very, very good question and hard to do right. Something that we do at Notion that I think generally, uh, is tried and true is to ask someone what they're here for. Many businesses will actually cast or assume or segment that a user is a certain kind of user. Maybe they sign up with a Gmail account, and one might assume that they're there for a personal use case, but it's possible that They used their Gmail, but are actually evaluating your tool or your product to, um, deploy at their business. And just starting with that simple question of, are you here to do work with your team? Are you here as a student? Are you here for personal projects? Can really help you as the builder of the product, understand the value the customer is seeking and reflect it back to them. And then Demonstrating value is more of an exchange. So when you ask a customer to tell you something about them, you can meaningfully change the product experience and reflect that back and help them see that this product can be relevant and specific to their use case and that, in fact, the product will help them get there. So there might be templates or examples or, um, different customizations that the product can make for them to help them Get there along the way. As, um, as one example, if I signed up for Notion today and said that I was a student, it wouldn't really make sense for me to see tasks and sprin…

AI assessment note: “ask someone what they're here for”

Q I'm seeing a recurring theme here. But, uh, I want to move to Facebook. You spent an incredible, several years at Facebook, three or four years I think it was at Facebook. What was the biggest lesson from Facebook and how did that impact your mindset?

A So at Facebook, I was lucky to work on several of the initiatives that were related to the launch of the Geophone in India and generally growth in international markets and developing markets. One area was internet.org. Another was the launch of the Facebook app that was retrofitted to the Geophone when it launched in India. And overall, I think my biggest learning from that was that other markets around the world are very different from the U S and I know that sounds obvious on its face, but I find that especially in technology, many businesses build with a Western bias and it's important to Really understand if you're serving all customer needs, especially when you have a big international customer base. Um, in working on those projects, I came to learn a lot about the nuances across markets that I didn't understand. And I think we built much better products by being global in our thinking rather than building what we thought would be the best product for the market.

AI assessment note: “my biggest learning from that was that other markets around the world are very different”

Q In terms of learning from other people, for a lot of founders listening, they're thinking, great, I'd also have to run, you know, a sales team, a CS team, a product team, and actually this is a growth function in a way, um, I should probably need to hire for growth. When's the right time to hire for growth, do you think?

A It does depend on your go-to-market. So if a business decides that a product-led go-to-market is the right approach in the early days, I would have somebody, usually an engineer, though it could be a technical PM, focused on growth earlier than in other businesses. And the reason for this is that you can actually deliver more scaled impact early, As you're developing that initial motion. Um, if, you know, I, I guess if PLG is not necessarily the crux of go to market or you're still figuring it out, um, the rule of thumb is to introduce someone working on growth after there's product market fit and to invest more if you see good return on that initial investment of bringing someone into work on growth. But it does depend and I've seen some Companies bring in one person, usually an engineer who's passionate about the area, pretty early to set up the scaffolding, make sure experimentation is possible, do some of the initial data logging, try out initial onboarding, and that has good return.

AI assessment note: “rule of thumb is to introduce someone working on growth after there's product market fit”

Q Talk to me about that. Does that not mean you don't hit your goals, but you're still happy?

A There's a difference between missing a goal because your hypothesis was disproven and missing a goal because you didn't execute well. So in the first case, you've actually learned something really valuable that a hypothesis that you held in building a product for your customers was incorrect. And ideally you understand the why behind that and you fix forward and build differently and better for them. In the case of missing a goal because execution didn't go well, that is where really getting into the details in a post-mortem or a retro and saying, What didn't we do well and what can we do differently next time to make sure we do give it our best shot on goal is really important. But in the first case, you actually been blessed with learnings that you didn't have before. So I think that's actually a great outcome.

AI assessment note: “in the first case, you actually been blessed with learnings... a great outcome.”

Q You mentioned this quite big change, really, to how Notion really organizes content and delivers it, and what the content is that it provides. Postmortems are fascinating to me. Fascinating. How do you do postmortems? Who's invited? Who sets the agenda? How often are they? Can you just take me into the world of postmortems and how you do them?

A Yes. We, we do them all the time. We call them retros. Postmortem is also great. And, uh, they're recursive. So it could be that a small team, maybe even a pair, a PM and an engineer who worked on something will retro together. It could be a team. It could be an org. It could be at the company level. Uh, but generally for something substantial, like a new launch, I would allow for time to see results and then Hold of meeting where everybody can share their Rosebud Thorne equivalent, you know, what went well, what we could do better, what didn't go well, and also assess and understand the performance of that launch and feel and follow through on accountability for if the original hypothesis in launching that feature delivered, and also what surprised us about the behavior we saw from customers and the business impact that we saw on the other side. This is something that can be pretty second nature in growth because we are so close to the data, but I find that that focus on metrics, on results, on impact can sometimes be less central in the rest of product doors and in places that are further away. Um, and it's a shame because there's so much you can learn from saying, Hey, we launched this feature. What did we expect to see in terms of usage, in terms of engagement, in terms of conversion, behavior changes?

AI assessment note: “We do them all the time. We call them retros.”