top of page
  • JJ Rorie

How to Build a Flywheel

Episode 067

Aakash Gupta, the Product Growth guy and VP Product at Apollo, joins to discuss how to build a flywheel, including:

  • What is a flywheel and what does it mean?

  • What are some prerequisites that a business needs to have in place in order to succeed in building a flywheel?

  • The biggest mistakes that are made when trying to build a flywheel across your organization.



Connect with Aakash:


In 2001, Amazon’s stock fell 90%. In a hail-mary attempt, Bezos invited a book author to Amazon. Together, they created the foundation for trillions in market cap:

The author was Jim Collins, who published Built to Last in 1994. By 2001, Bezos had already adapted several of its concepts into Amazon’s principles... READ MORE

How have Uber, Spotify, & Netflix continued to grow while focused on a single market? Pursuing a singular flywheel (genius strategy)... READ MORE

Episode Transcript:


flywheel, product, strategy, amazon, team, spotify, create, company, step, build, business, b2b, scale, selection, llm, organization, prerequisites, focused, failure, leader

Intro (the incomparable Sandra Segrest) 00:03

Welcome to Product Voices a podcast where we share valuable insights and useful resources. To help us all be great in product management. Visit the show's website to access the resources discussed on the show, find more information on our fabulous guests or to submit your product management question to be answered on our special q&a episodes. That's all it product And be sure to subscribe to the podcast on your favorite platform. Now, here's our host, JJ Rorie, CEO of Great Product Management.

JJ 00:37

Hello and welcome to Product voices. I am so incredibly excited about today's episode, I have got Aakash Gupta with me. We're going to be talking about the flywheel how to build a flywheel how that impacts your business and your product. And it's just gonna be a great conversation. So everyone already knows Aakash. But Aakash Gupta is the acclaimed writer of the product growth newsletter, he's got over 150,000 followers on LinkedIn and 55,000 on Twitter. I know you've heard of him already. After 15 years of product leadership in the trenches at places like Google and fortnight now, he's working on a paid newsletter, and we'll make sure to put that link in the show notes as well. So if you're not already a subscriber, which most of you probably are, well, we'll get you there because it's amazing stuff. So, so excited to have Aakash here on the show. Aakash thank you so much for joining me.

Aakash 01:29

Oh, that was way too kind. I'm so excited to be chatting with you. Flywheels are like such a misunderstood concept. So hopefully we can try to peel back the layers of the onion a little bit beyond the surface.

JJ 01:42

Yeah, absolutely. I love it. It's such an important concept. But but a misunderstood one. I think still so loving love, gonna love your, your insights here. So let's start with the basics. Okay, so what is a fly will, you know, give us the kind of foundational understanding and what it can do for product teams.

Aakash 02:00

So let's start with what the flywheel is not. It's not this scattershot strategy. Like in B2B, what you'll often see as a B2B SaaS company launching many new products, that is not a flywheel strategy, or on the B2C side, you'll see them consistently open up new markets or new segments. That's not a flywheel strategy. Yes, of course, it worked for block, like they have the square business, they have the cashapp business good for them. But that's actually not the flywheel strategy. The entire idea behind the flywheel strategy is that everything you do at the company is related to kind of one central motion. And so the whole concept of flywheel is let's focus in on a circle type effort. And so the circle type effort can come in so many different archetypes. And we'll talk about some of those in a little bit. But the high level view that people should be thinking about is like the Amazon e commerce business. And this is the example that really helps drive it home. So what is Amazon's core values that for 20, 30 plus years could drive ecommerce, its price, its selection, these are like basically the two most important things. And so the team actually at Amazon when their stock went down 90%. Post y2k, they came up with understanding that how do we reinforce this cost in the selection in a circular fashion. So they'd realize, like we could get lower cost via lower prices, leading to more scale, leading to lower costs. So they've created a circle around lower cost. That's a reinforcing flywheel. And then there's the selection one, so greater selection, leading to customer experience, leading to traffic, leading to more sellers, which comes back to greater selection. And so with the Amazon e commerce example, you can really see how everything they did in the early years of their life in that part of the business was just reinforcing that flywheel. And this is kind of the central insight of the flywheel strategy is you build something that as it scales, the economics improve. And so if you're a tech company that's venture backed, or like Amazon, they Republic and still losing money, you still have a path to profitability in the future. So that's what the flywheel concept is. It's a path to profitability based on scale.

JJ 04:25

Yeah, I love that. And I love the example. So So, Amazon's a great example. I'm gonna ask you about some other examples in a minute, because I've loved reading some of them in your writing. But I have kind of a another basic question, actually, that I think confuses some people. So I'd love to hear your thoughts on this. The Amazon example, is a good example of the business as a whole, you know, needs to embrace this and create this flywheel. So can product teams drive this effort, or is this something that has to come from the very top of the kind of business, you know, corporate chain.

Aakash 05:07

Yeah. So in the Amazon example, the purpose of that example is to show that everybody in the business needs to be aligned with the flywheel strategy. And so if you're a product person, and you're trying to push your business towards a flywheel, what I recommend is two stages. One, build a really strong product flywheel and then two, as you have some momentum around that create, buy in all the way up to the CEO level to drive the whole business viable strategy. So a company that does this really well, for example, a Spotify, if you think about Spotify, what is a product based flywheel that they could create? So what is data? And so what they did right is they took their word of mouth, they got users and scale that drove engagement, which drove data personalization and an improved experience. So there's this circular growth loop purely within product. And then what that means tactically, if you're a product leader, is that you're going to have a lot of teams focused on your data models, you're gonna have a lot of teams focused on your engagement and personalization at uses of those models. And so as a PM, then, if you're a PM, you're thinking through, let's say, you're a PM sitting on the data team, how can I drive more of a flywheel within my business? Well, a lot of it is going to be kind of evangelizing and understanding how you can bring on other teams. So as Spotify grew their product team, they started to take the flywheel outside of just product. And they created the promoters flywheel, which went from word of mouth to lower customer acquisition costs to better cost structure to the users. So what teams does that involve? That involves marketing on the customer acquisition cost side, and then the cost structure side that involves your music industry contracts, so your key music execs, and so they're slowly expanding the aperture of their flywheel. And so if you're thinking about being a product person, try to think of the Spotify example and how you can slowly prove it out and then expand.

JJ 07:10

Yeah, that's great. Another great example. So are there like, again, folks out there listening, thinking, This is amazing. I'm gonna I'm gonna go tomorrow and start this. What are like? What are some? I'll take this two parts. The first part of the question I'll ask you is like, what are some prerequisites are, are there prerequisites that a business that a product team needs to have in place in order to kind of have a foundation that this will succeed and that that eventual evangelizing? And that sort of thing will succeed? Any prereqs for, you know, building a successful flywheel?

Aakash 07:46

Well, yes, there are definitely some prerequisites. So let's start with the B2B example. And then we'll move to B2C. So these worlds are pretty different as it relates to flywheels in the B2B example, if you are like a primarily linear growing sales type growing B2B org, then the flywheel for you as a product team that you're going to want to extract needs to be related to that growing scale. But the requirement then is how do you break out of this linear loop and create it circular? So do you create a self service motion? Do you create a plg motion, that's kind of the requirement that you need to have on the B2B side to truly leverage the scale that you're getting from your linear motion. So that tends to be one of the prerequisites on the B2B side, some teams do work around that. So they might work around that by creating a data flywheel like Spotify had. And so I would say like, don't let any prerequisite I throw it at you deter you, almost anybody can come up with the flywheel. But it's really good to be honest about what your areas of constraint are. So that's b2b. Now, let's go to B2C. So obviously, b2c In general, like you can sign up so that is there. But what you need to think about is how can I from skill connect scale to a core value that my consumers are choosing me for? And so that's why the Amazon example is so brilliant, like pricing selection, those are kind of timeless values in the shopping category. But for every b2c product, it's not quite as clear. Let's take Robin Hood, for example, like what exactly is data they don't want to share, like investment decisions are data for other people? So that's like getting it to them? What is more money getting to them? Even when they started, they had a very performant quick system. So thinking through like on your business, then as the Robin Hood product team, you would try to think through like, Okay, how could we create, for instance, a social product, or something like that, and that takes advantage of the scale. And so that's where I love your question is, you can kind of think of your limitations and then build your flywheel around it and I truly think like anybody can Look why well,

JJ 10:01

Yeah, I love that. And it leads to my second question. Or second part of the question, if you will, which is, you know, how would you start right there, there are people out there again listening that are like I'm doing this this is. So this is so exciting. Like, what would you do? So So part of it would be kind of, you know, a true kind of self assessment of constraints and what what you have today, what you don't? What else? Like, would you advise someone to get started on doing this?

Aakash 10:27

Yeah. So, I will kind of walk through like, basically the seven steps that I would recommend. And this is what I actually did for the product growth newsletter. This is what I did at my last gig in a firm. So it's kind of just timeless steps. But we'll kind of give you a more concrete example as we walk through them. So the step one, like identify your key company goals. And the question you're asking yourself here is growth in what? So for Airbnb, for instance, they established a goal of more booked nights. That's not a really intuitive goal to establish. Because if you think about it, a book tonight could be a $60, Book Night, or it could be a $1,000 book night. So you have to somehow create a lot of conviction across your org. And you need to usually have that driven from the CEO of this is our Northstar, what is your key company goal? So that's like step one of building your flywheel. Because it's not always going to be revenue or profit, it's going to be something else. And so, think of Spotify. It's more paid subscribers, think of Netflix, it's more subscribers, think of Amazon ecommerce, its e commerce transactions. So those are some examples to inspire you. And then step two is figure out the key elements of your strategy. So the strategy question is a lot of layers to it. What I would say is that there are basically three types of strategies. There's the strategy, which is focused on like one driver. So that's like your stripe strategy. More transactions leads to better payments. There's this strategy where it's focused on two drivers. So that's the Amazon example where we gave pricing selection, or the Spotify example where we gave data and word of mouth promoters. So there's the two driver examples. And then there's also the ones where there's like seven to 10 drivers. So a lot of people have probably read Matthew Bowles work on Epic Games. Remember that chart he had, it was very complex chart, it basically showed like lines between like 10 different products. So I actually worked at Epic Games. And I can double click on this example for everybody a little bit. So Epic Games has the engine, Unreal Engine, it's used by literally every single triple A gaming studio and triple A games are basically that you can think of as like the top tier games, so everybody uses that. So amazing, right? And then they have their own games on top of that fortnight and fortnight is getting the latest tech in Unreal Engine. And whereas like your gaming studio at EA or something, they don't have any incentive to like use the newest beta tech, but fortnight they do. And so that gives this reinforcing loop to Unreal Engine. And then they launched the epic game store to sell games. So then they're learning a ton about what they should be putting into fortnight, so they keep adding in layers, and they've added in three or four more layers, like publishing and different things like that to push towards the metaverse. So what I wanted to say in the step two is that you need to identify how many drivers you have. That's like the next step within building your flywheel. And then the step three is you've identified how many drivers you have start to connect the dots. So that's how I tend to think about it is like, okay, so I have all these drivers. Now, let me start to diagram it out, and use all the examples we've shared today diagram it out. Step four, once you've diagrammed it out is how do you enhance that diagram with quality and cost. So literally, every single flywheel that you look at will have an element in quality and cost because those are critical drivers for your long term business success. And using your first draft, like we talked about more book, tonight's are more subscribers, those are very top line focused. And so you also need to think about some of these bottom line areas. Step five is review it. So you're gonna have to pressure test this a lot. And what I've noticed when building flywheels at places like affirm or Epic Games is that a lot of times the big failure pattern is that there's still not a singular purpose. You've created it very much to accommodate and please everybody and cooked it up in the kitchen with seven cooks. And so you usually have to simplify it step five. And that's also step six, where you're iterating, then with the team, so you're taking it back to the team and saying, We removed this, we added this. And then the final step is like pressure tested in the wild to start making decisions to say no to projects, because they don't fit within your flywheel, because that's kind of the fundamental insight of the flywheel strategy, and then see how it works.

JJ 14:57

Yeah, that's amazing. And one of my questions was going To be like, can you? Or when do you change those strategic elements? And it's actually part of the process to iterate to make sure to, you know, test those out and say, yeah, we've got too many, or we've got the wrong ones. And then you can move, move forward, you know, again, into the wild. I love that because it's all wild, wild west, but into the wild, and really test that out. So I love that you've, you've kind of thought about how to set that up in the first couple of steps. And then really, before you even put it out there really kind of validated. I think that sounds like it's a really important part of the process.

Aakash 15:36

Absolutely, especially if it's new to your company.

JJ 15:39

Yeah, yeah, definitely. So, you know, we talked about kind of being or evangelizing this around the organization, maybe share some thoughts, if you will, about how you've seen that work? Well, you know, you've done this yourself, you've helped others do it. Like how, you know, what are some of the things that, you know, assuming we get through the process, and we're going to, we're going to try this this flywheel? What are some things that product teams, product leaders and others in the organization can do to make sure that, you know, they find these champions and they start to communicate the value of these flywheels?

Aakash 16:20

Yeah, so I would say like, depending on your level, if you're at like a mid to senior level, your aperture really needs to be kind of your C suite, your senior leadership team. And so if you're just implementing the flywheel within your team, you know, it's going to help you a little bit, but that's going to be your major unlock. Now, if you're truly a junior pm, then I think your aperture should be your product leadership. And so that's how I would really think about it is like, I either want my product leadership to get on board. Or if I'm a product leader, I want to get my C suite and full team on board. Because the failure pattern here is we're going super hard on a flywheel. But other teams are complicating that they're focusing on other things, kind of the fundamental insight. The reason the flywheel strategy actually came about in the Amazon example was Jim Collins studied a bunch of s&p 500 companies over 40 years from like, 1955 to 95, or something like that. And he looked at those companies that over those 40 years, are able to stick to that one flywheel across departments, not just a product. Those were the companies that actually survived. Actually, a lot of the companies looked at died, because 40 years of being a public company is very hard. But not only did these companies survive, but they are some of the best performers like Southwest Airlines, you look at how simplified their businesses it's like, just around this core flywheel. And so that would be how I would think about kind of socializing it. Now the second part of your question was like, now you've got some champions, I love that framing and phrasing, and I would say, like, people should triple click on this specific technique, which is, who loves your strategy and the level below the C suite. So we're trying to influence C suite. So in my experience, like VPs of operations, s VPS, of finance, even maybe like VPs of growth, marketing, these can be really great allies. And if they're completely on board with the flywheel, you start to create that chorus of feedback, because you know, product teams are only with the senior leadership team for I don't know, 10% of the time, right. But the marketing team, the operations team, everybody's starting to talk about this concept. That's when it really sinks in that it's kind of like a departmental wide strategy.

JJ 18:46

Yeah, starts to echo around the organization and it's it's no longer just just ours but it's everyone's or not just yours, it's ours. Right. So yeah, I love that seems to make a lot of sense. Now, you mentioned a couple of these kind of, you know, failure points or or you know, mistakes that are made but like what do you see the biggest few mistakes that are made as you're trying to to build this flywheel across your organization?

Aakash 19:13

Number one is just not enduring. So you what you'll see often is that like, you know, some senior product leader comes in they are there at the company for three years they driven some prioritization decisions where we didn't go after flywheel strategies. They've influenced the marketing and sales staffing, so that we went after flywheel strategies, but then they leave and the company abandons the flywheel strategy. And all of a sudden, they're launching all these new product lines. They're serving all these new buyers, and you're like, huh, I thought you were focused on the product leader buyer. Now you're talking to the ID ops buyer, and you're talking to the sales ops buyer, and it's a very interesting and common failure pattern, which is that this flywheel strategy, this is really one of those very long term strategy If you look at an Uber, or a Spotify or an Amazon who's pursued this, they've pursued it for many decades, and so you can't kind of just a one and done strategy. So failure pattern one, failure pattern two that I tend to see way too often is doing the look flywheel strategy and lip service only. And so you have a flywheel strategy. You focused on it, you published it, you talked about it at all healings. But when it comes down to how you're actually staffing your company decisions, you're constantly de risking against your flywheel. So if you look at for instance, I would say like, I can't fully comment on meta, but I would think that like they're rebranding their name to meta when it was delivering Metaverse was delivering no profits, and they're getting 10s of billions of dollars in profits in their core social media businesses is kind of that example, which is to say, like, leaders, at some point, they get worried about a new platform, or they're trying to own the next platform, something happens where they get out of their course sweet zone. And then you see like a Facebook Big Blue app, it's not always growing now, I think it's had several quarters or months where it's been slightly decreasing. And so that's an example where you kind of lose conviction in your flywheel something happens in your market circumstance, you start to invest in other areas, and the flywheel becomes a name only. So that's flywheel kind of mistake number two. And then the third mistake that I would highlight for everybody is believing that, Oh, we didn't get results on the flywheel in one year when we implemented it. And so it doesn't work. And that kind of goes back to that core thing, which is that this is kind of a business wide, long term strategy. Whereas some of the tactics that you hear about as a product leader, you know, I'm going to implement OKRs, like, you'll see some change in how your product managers behave if you implement OKRs, the next quarter. But if you implement flywheel strategy, it might take years, it will require door product directors to organize their teams appropriately that will take time, then it will require the actual PMS, that'll take time then bring on other departments. That'll take time. And so that's the third failure pattern I tend to see.

JJ 22:14

Yeah, you know, and it's, it's so interesting. And I think you've really hit on why this? I don't know that it's an anomaly anymore. In organizations, there are enough organizations that have done it well. But it's certainly not the norm, at least in my experience, are there enough organizations out there still struggling to build their their flywheel? And I think it's, I think, if I were to put, you know, a pin in the reason, it's because it's that long term, right? We want instant gratification, you know, the tenure of employees, leaders, etc, is shorter than decades, right at this point. And so it really has to be a core thing of the entire organization and part of the culture part of everything, you know, throughout all of the employees that come in and out of that, and that's hard to do. Right? You agree with that? That being kind of the the one thing why it's it's obviously there are multiple failure patterns. But do you think that's the reason why not out? Not every organization does this?

Aakash 23:13

Well? Yes, in all the case studies, I walked through 10 of the most important in case studies in my piece, and all of the case studies basically have like a founder led company where somebody's been there for decades. And so if you're a product leader, you really need to think about like, can I actually drive success with the flywheel strategy, if I bring this in if the company doesn't already have it? So if the company does have it, like upgrading the strategy is really great. Or if you think you can kind of convince the CEO to like, carry it on even after you leave. So this is one of those tactics that's like, really interesting business strategy wise, if you're C suite, or you're a founder, you should be thinking about building a flywheel. But if you're a product leader, you're not going to always be implementing a new flywheel strategy, you might be upgrading a flywheel strategy. I bet you a lot of you listeners, like you work at Netflix or Spotify, and you're upgrading the flywheel strategy, but you're not going to like bring in an entirely new one. So I don't want anybody to listen to the seven steps, and then try to bring in the flywheel to a linear company, and hope that that's gonna get them their next promotion. It's actually not great at that specific function.

JJ 24:24

Yeah, yeah. Are there any elements within it? If you're in that circumstance, where a look, we're not going to create a flywheel here, this is just not you know, it didn't come from above. And it's not it's not going to happen here. But are there elements that that those listening can take from the concept and say, we're at least going to do this and that alone is going to add some value? Any any thoughts on that?

Aakash 24:49

That is a great point. And that is like a fantastic place for us to end is like everybody should think about how do I build in some sort of skill based advantage. Yeah, and this is an area that every single product leader, no matter whether you're on a platform team growth team, core product team, you can actually do a little bit of this in this new world that we live in. So the coolest example, of course, the hottest example right now, is generative AI, like how do you train more of your customers to provide input into your generative AI model. So your generative AI model is really good for your specific use case. So that's probably a flywheel that most of you can go implement, like right now, like anthropic just released a 100k, token API based LLM that you could use. So what that 100k token means is you can provide a ton of context, it could read, if you're a CRM, it could read your customers, entire CRM data. And so you could use something like the entropic API, build it into your LLM and create advantages to scale. But you don't need to go with the hype, like maybe you don't believe in LLM 's or generative AI, you can go back to the very core insight behind all of this, like a neural network that's trained on your data and personalize. We've all been using personalization as a tactic as product leaders for the past 15 years. So just do it. In this case, with the newest tech on neural networks. That's another way. Or a third way is like, look at all the examples we've gone through. We've talked about cost as it relates to scale. We've talked about data as it relates to personalization. We've talked about many other examples, just see which of those analogize to you and start to dig up one or two of those.

JJ 26:35

Yeah, I love it. And it's such great kind of parting words for you know, someone listening and you know, what, what they can do? Or did I can, you know, embrace the entire fly will, or just elements of it and really bring this in. So amazing conversation.

JJ 26:52

Aakash Gupta, thank you so, so much for joining me and for sharing your awesome wisdom on this subject. Thanks so much.

Aakash 26:59

Thanks for having me. It was pleasure.

JJ 27:01

Awesome. And thank you all for joining us on product voices. Hope to see you on the next episode.

Outro (the incomparable Sandra Segrest) 27:07

Thank you for listening to product voices hosted by JJ Rorie. To find more information on our guests resources discussed during the episode or to submit a question for our q&a episodes, visit the show's website product And be sure to subscribe to the podcast on your favorite platform.


bottom of page