 Hello, everyone. My name is Nikola. I'm a product manager at Waymo, and I'm here today to talk about road mapping and very particularly about growth stage product road mapping. We have a very big agenda, so we're going to dive in very quickly. The idea is to talk first about what is a road map, then what is a growth stage product, and then combine the two and define what a growth road map is. Now we're going to look at some of the key elements of a growth road map. We're going to look about what is a growth lever, what does it mean to prioritize when it comes to growth, how you experiment, the different type of growth, some key examples, and then risks and takeaways. Very simple start. What is a road map? You can think of it, and I think everybody's aware of it. It's just a document outlining what you're going to do, how much is it going to cost you, when you're going to get done, and where you're going to get out of it. You use spreadsheet, you use JIRA, you use any type of product management. In the very essence, it's just a list that is in a specific order and it says, this is what we're doing. Now, a little bit more finicky. Here's the definition of growth stage product. Negative way of defining it is to say that a growth stage product is a product that is not zero to one, and it's not declining either. In a more positive way, what you would say is that a product that has validated its use case, we probably earlier adopters, and it's ready to mass adoption, it's a product that is entering growth stage. Any product where the retention curve has flattened, so you've found a core use case, you've found a core set of users that are interested in your product and are sticking with it, is a product that is entering growth stage, and what it means is that you're probably going to see an increase in process innovation and a decrease in product innovation. Why is that? Because you want to push for optimizations versus innovation because optimization is safer, has less risk, and it's more likely to give you outcomes. And since you have already achieved market fit, you don't want to jeopardize that by taking swings and attempting new things. You want to make sure that you extract the value that you've built at the maximum possibilities. And so when you think about growth roadmap, it's really about putting together these two things. What are the projects that are going to help me maximize my return of investment by quantifying cost, outcomes, and risk in a very clear way? How can I make sure that I'm going to take the projects that are more likely to give me this, and so it means more likely to have a clearly contained cost, a clear contained outcome, and very clear contained risk? Why? Because you don't want to jeopardize your market fit, you want to extract from your position. And so we're going to dive more, but this is a key difference with zero to one. Zero to one roadmaps don't have clearly quantifiable cost outcomes and risk. Why? Because you don't know how long it's going to take to build what you want to take. You don't know what is it going to get out of it, and you don't know what problems you might encounter down the line. And so that's a very clear difference. Once you've built something, now you know what it takes to build it, you know what it takes to improve it as well. And so growth is about smoothing the edges, clearly in the pipes. Everything is much more clear. It means also everything needs to be done in a more precise way. And it's about volumes, because you want to have a lot of small incremental changes. But we'll see a little bit more. Let's dive into key elements now. North Star metric, growth is about measurement, and so it's about North Star metric. Every product that metrics zero to one products have metrics. They have sometimes North Star metrics. Most of the times they have multiple metrics. And the very big difference between zero to one and growth is that growth stage product are always able to roll up any granularity of metrics up to a North Star. That is not always true for zero to one because in zero to one, you still don't know how you're delivering value. And so you don't know which of the two, three, one, five metrics you've picked as like very high level is the one that you need to trade off on, is the one that you want to move the most. And generally zero to one product to have one, two, three metrics at times, because you just have to figure out which one is the right one, which one is the measuring the real success of the product. Some examples, some material, if you want to dive deeper, but I think this fairly clear. The next step is a user funnel. Again, something that like you have as well on zero to one is just breaking down the user journey through your product. The reason why you do that is because you want to identify like each product feature and how it relates to your metric. And some very high level concept when you think about user funnel, it always goes between like, cannulically can be interpreted as like going from awareness to interest to evaluation to conversion, right? We'll take different products, we'll see how this relates, but it's a very high level, that's the idea, right? When you put this, this is an example, I forgot about this slide, this is an example of what user funnel look like, awareness, this is content creation for, so awareness is, I have an entry point here on the bottom right, you don't see very well, but there is like a blue, this is Twitter, formerly known as Twitter. And so you have on the bottom right, you have a plus that says like enter creation funnel. So I'm aware that that exists, then I go to the next step is an interest aspect. I clicked on it, I want to see what is it about, there's an evaluation aspect. I added a picture, so I'm going to make a post and then finally there's the conversion aspect I'm going to post, right? So different types of steps for different levels of this product funnel, right? Now, when you put this two things together, you get into a growth model. So you think about all of this product funnel, you can break them down into specific metrics. Number of creators is the top north star metric when you're thinking about being in charge of content creation. Then that number is a function of different metrics and each of specific dysmetric can break that can be breaking down into more granular metric and more granular metric relates to specific part of the products. So existing creators times posting rate is plus new users and activation rate that's your creator's number. Now posting rate is a function of like user entering the creation funnel and the completion rate. So going through it. Now the number of users entering the creation funnel is a function of like how many users are at the top of the creation funnel and their CTR and so down the line, you can break all of this. And so you can start thinking about like how you go from this north star metric down very granularly to a very tiny metric that is specific to a very specific part of the product. We'll see why you'd wanna do that. But the bottom line is moving north star metrics extremely challenging and you don't want to start a project thinking gonna move that directly because you might not be able to measure it. And the reason for that is that you might not have the time or the experimentation power to just go make a product and make an impact and measuring it in a short term. And so that's why you rely on smaller metrics that are more sensitive to your change and that you know can be rolled up to your key north star eventually. We'll see more what it means. So now you have like a north star, you have user funnel, you have a growth model. Let's talk about how you move the needle, how you come up with projects, right? So this is a good time to introduce growth levers. Growth levers are like some principles here on line five, you can Google, you'll find different, but this is the find that I found be very consistent into the product that I work with. The idea is like you wanna create, how do you move the needle, how do you grow that product funnel, right? Like these are some principle that you apply to each of those steps of the user funnel to move the metrics in the direction you want. First step, you can increase discoverability, right? So for every of those steps that we talk about, you increase the discoverability to go to the next one. You make it more visible for users to increase the awareness. Then you have reducing friction, you wanna make that simpler, less clicks. And so you move the user down the funnel, then you have increasing friction. So you increase how hard it is to actually go back into the funnel. Now this can be seen as like a little bit antagonistic at the beginning, because the idea is like, oh, you're gonna trap the user. It can be done that way and it is done that way, but there is actually a healthy way of doing this as well. And there's mimicry, which is basically the idea of like, I want to help my users just to do what the other user are doing, right? Like so you see something and you just wanna replicate that. So that is to a little bit simplify the process for the users and just like having to do something that someone else already did and they've seen it and they're like, yeah, that works for me. I wanna do the same. And then the last aspect is the reward, right? So increasing the reward of like completing this funnel of going through the process and why you wanna do that? Cause if you increase the reward, they're more likely to do it again, right? Like user are more interested, like they went through it and they was like, this looks good, I like it, I wanna do it again. Now we'll see some example, but let's dive into prioritization. So the main key difference here in prioritization is that growth prioritization is a continuous functions and leverage some of the variables that are available on a growth stage such as like this clearly quantifiable risk and returns, right? And this is the big difference with zero to one where you generally have a prioritization that looks very much as like P zero, P one, P two, right? And that's a very step function is like P zero, all of this projects P one, all of this projects and so on, right? But growth is more continuous. You don't really wanna fall back to P zeroes. You wanna go more into like, this is specific outcome. This is my ROI and here is the continuous function of all of the projects, all of the ROIs. We look a little bit at an example, but let's take a little bit of a closer look. ROI is you expect a movement of your North Star metric or whatever metric you use to pick has to be consistent across different projects, but it's basically like your end goal, right? It's good to have like a lower and upper bound is good to have them so that you can kind of feel where is it gonna move. Then you have like confidence. Confidence is basically how likely the projects that you're talking about is going to deliver what you predict is going to do. And you can set up like cross team agreements or what confidence being, it is often a function of like having experience. But a good way of saying it is like, you can take a percentage and be like, well, I'm gonna say 20% confidence if this has never been tried and 90% confidence if this has been done already on a different side of the product company and we've proven results. For example, this worked on Android, it's gonna work on iOS very likely. Not always, but confidence is higher. And you have cost, cost is just like your constraint function, right? Like what is the number of weeks available from the engineers or like your marketing budget and so on, right? And these two other, these three elements gives you a priority score. Here's a practical example. Let's talk about like, I think that there's a lot of users that are misclicking and exiting my creation funnel. So I'm gonna create like a confirmation that says like, are you sure you wanna get out? Apologize for this bad design of mine. And here you have like a real example from Instagram. So the idea here is increasing friction in exiting the steps, the creation funnel by mistakes. Now, how do I build my ROI? My ROI is going to be the number of users that enters a flow and then exit it. And I am assuming that some of them will be executing by mistake. You can try to approximate how much or how many of those you can use like different type of experience that you have where you've noticed that there was a mistake percentage. But let's assume that I come up and says like, I think 7,000s of the users that enter every day actually are exiting by mistake. And then they're discouraged and they don't wanna get back in and redo it. So they're like, they give up. I am 50% confidence. This has never has been done in some shape or form in a different type of the company, but not in this exact context. The implementation was different. So 50-50. And it's gonna be very quick. It's gonna take half of a week for me. And so here it is. My return of investment here is 7,000s per week. Now, why do you wanna do this, right? Because you wanna have a long list of this type of projects and you wanna look at them and being able to compare them. So imagine that you have like a long list of this. You wanna have a very, very, very thick list of projects in growth. And then you start looking at the end score priority and you immediately notice how some things don't end up being as high as in priority as you might expect. For example, let's say that you have like, oh, I'm gonna make an AI generated post machine, right? It's gonna have massive ROI, like 2,000 users every day is gonna use it. But I'm not really confident. I don't know if this is gonna work, if I'm able to build the solutions and it's gonna have a high cost six weeks, right? And then you compare it with like, I'm gonna add this tiny dialogue or I'm gonna increase the visibility of the CTA. And then you look at the expected impact and the confidence and the cost. You make those ratio and you immediately realize this is much more worth it. Like it's less ambitious in many ways. It's less innovative in many ways, but it's gonna just work. And that's what you wanna go when you're talking about growth. You wanna make things that are higher, return, lower risk. You're more risk adverse in world stage. So let's talk about the next step, experimentations in growth when you have projects, then you wanna start thinking about like testing them. Typically be testing generally last like one to four weeks. And once the decision is reached, it's generally ideal to keep a small hold out and we'll see why. And the process I think is very similar, A-B testing review, launch, iterate. Not all the times that can happen zero to one. At times since zero to one, you don't have enough volume to do a fair A-B testing. At times you also don't have like the possibility of not launching. Like it's not a choice. Like you invested too much just gonna get launched and you don't have enough data to say like, I don't, I shouldn't. And so when in doubt, you just have to push through. When you're in growth stage, because the product are so in between quotes cheap and you have so many, not launching is actually an outcome and an acceptable outcome if you think that's not working. And now the reason why you might not wanna launch is because there's different type of growth. And so some products you'll notice give you growth that you don't actually want. Now a disclaimer, these are absolutely made up names. There's no such thing as a universally understood as cumulative growth or state growth or regress. These are just like things that I made up right now to share the concept. The idea behind here is that growth can be cumulative. So it can continue to increase over time, right? The feature that you do keep, like increases the value that is delivered to user with times, right? And we'll see an example. It could be a state change. So it's just building the product from one level to another. It's a step function, but once it brings all of the user to that level, it just stabilize. Like it's not gonna continue to growth indefinitely. And then the last one you have regressing growth, which is your change has initially drove a lot of value, but then it's actually all that value has disappeared and it's mostly a function of novelty or some unexpected results that you don't want that you haven't understood from in the network. And we'll see some examples. So let's go into very practical examples, right? First, very famous one, people you might know from Facebook, idea being you come on Facebook and you wanna retain the user, you wanna have them come back. And the Facebook team realized that like friendship was a very good indicator. If you have a number of friends, like I think it was 16 or 14 in the first 48 hours or something, you'll come back. And so the idea was like let's increase the awareness, the availability, the ability for users to actually make friends. Let's put that at the very top of the main page. And the interesting thing is that this is a long-term impact, right? Like it's commutative in the sense of once you remove this feature from the top of the mainline, your friends continue to be there. So you continue to actually deliver, like you continue to extract value from having make that friendship, from having used it, used this type of feature. So it's the best type of growth you can hope for. It's something that you do and it's a change on the operating mode of the user. In this case, the metric was friendship, probably had some guardrails, but that's a very successful example of growth. Now, another example here is autocomplete, right? So here you're trying to reduce friction. You're trying to make things simpler. And the idea is I wanna make it so that users don't have to spend too much time into typing and I wanna help them. And so you reduce the friction of the search experience or you just the friction of like replying to messages or replying to request and so on. Now this is not a self-sustained product growth feature. Once you remove it, user will not retain the value of having used. The next time I make a search, I don't have the past searches there. Next time I have to reply to a message, if you remove that, I'm not gonna retain anything, right? So your numbers, when it comes to like number of searches or like number of replies, will go down as soon as you remove those features. While in this case, if you remove this, your DAU will not go down because those users are retained at this point. They have the friendship it's there. So in this case, it's a little less self-sustaining, right? Another example here is the discoverability aspect and the friction aspect. So there is a Facebook feature where if you swipe right and left, you see like this music thing popping up. And the idea is like, let's make it simpler for users to discover music and to enter the music creation flow. And a similar use case is on the right is a in-between stories card that says like create a new story, right? So you're trying to increase awareness, like making people aware that these things happen. And you're trying to make it simpler for them to enter it. And the idea is like, let's click away from you to go through this flow. Now this is again, not self-sustaining in terms of tomorrow you remove this and nothing like your number of creators will go down. But it has some network effect, more stories, more content, more engagement and so on. Another example here is the mimicry aspect, right? So the idea is like, users see something they like and they wanna be able to do the same. On Instagram, you have like the use this audio thingy when you're consuming a story with like a specific song, you can like, you know, use the audio and build immediately something with it. Or you have like on WhatsApp, something that I found interesting was when you're talking with a business on WhatsApp, you have like an ability to check that they're a business and once you're on their profile, you can get your own business account. So the idea being like, oh, I'm a business, I'm engaging with this business. And I'm like, oh, I also wanna have this. How do I get it? Here it is, this is like, you notice it, you see it, you wanna replicate. And so again, this is non-self-sustained for the audio aspect, but it is self-sustained for the WhatsApp aspect because once a business is converted, it's converted. So that's a shift in the operating point of the user. Like it's changed something at the very core of that entity. So we're talking about some of these example, let's look about the risk behind. The two main risks behind growth and the first one is short-term prioritization because what we discussed right here of the way that you construct priorities, you will lean towards short and safe projects. And so you will eventually end up in a situation where you have short-term prioritized everything because anything just falls below the list when it has a little bit more daring, it's, I don't know, tech depth related, it is quality related, it just doesn't make the cut. So it's very easy to reach a local maximum where you kind of painted yourself in the corner of the room and now you don't know how to get out of it. And it can be quite painful. Another example of short-term prioritization and it's the reason why we spoke about holdouts is that sometimes growth, and this is the type of growth that you don't want, can be regressing, right? So in the short-term, you see this impact and then on the long-term, this actually sets you back. An example is reshares, right? Like reshares can like send up the number of creators, you made it much simpler to create, but then what you end up having is a situation where your network is much easier to spread fake news and the quality of your content just decreases because everybody's resharing the same things, right? So you don't have new content coming in and that decreases the value and then long-term impact can be negative. So you short-term prioritize for immediate gains, you didn't look at what was going to come ahead. The second big risk is being leapfrog. You don't take big swings on the growth team. You don't take daring features and so you might be left behind. To get out of tech, a good example, it's caught up focusing on improving their chemical imaging and then someone else was working on digital cameras and then they got leapfrogged. Another one is Facebook, we like stories. They were working on improving like posting onto timeline and someone made stories and you got leapfrogged and they included stories and then someone made reels and like the TikTok thing. So they had to copy those as well. And so it's very easy to see like you're established, you don't take big chances and so then you need to learn how to incorporate what's being done in the market and kind of groid yourself again. Now, this is kind of what we have is very high level. I think it's a good starting point. Let's recap very quickly. You know, growth roadmap is about maximizing your return of investment by focusing on project that have a clear, quantifiable cost, outcomes and risk. The main ingredients for this are generally a north term metric, a user funnel and a growth model. And once you have those, you can leverage the five common growth levers that I mentioned to identify projects out of any specific user funnel to move the metrics through your growth model and eventually impact the north star metric. The way you prioritize all of this project is through continuous functions that take into consideration cost, return and risks in a way that you cannot do on zero to one. And then while you do this exercise, the biggest risk is always to reach a local maximum and be leapfrog by other products. And so that's what I meant when I said not all products are not all in zero to one or all in growth. You need to have a balance. Some teams work on growth, some teams work on zero to one for every product. And that is the way that you prioritize, you balance short-term prioritization through growth and long-term sustainability through zero to one teams. And so when you are unbalanced into that split, then you start getting into problems. And you know, there is no formula, there is no silver bullet, it's just a try and error and you need to find what is the right time to push on one or to push on another. Out of this world is useful, useful. If you have any questions, please feel free to reach out and comment and it was great to talk to you.