Juli 8, 2019

Episode 9: Gojek's Growth Dilemma

Episode 9: Gojek's Growth Dilemma

Intro: Welcome to GO FIGURE. My name is Nadiem Makarim, CEO and founder of GOJEK Southeast Asia's first Super App. GOJEK does ride hailing, food delivery, payments, even on demand massage. You name it. We do it. GO FIGURE is a podcast dedicated to expose the inner workings of ambitious tech companies in the emerging world. We like to talk about things we like and talk about things we don't like. There's a lot of myths out there that we want to dispel. So keeping it real is kind of our mantra. Hope you enjoy.

Nadiem: Hey guys, welcome to GO FIGURE. Thanks for being here.

Crystal: Thank you for having us.

Jasper: Thank you.

Nadiem: Awesome. I'm here sitting with Crystal, our head of growth platforms, and Jasper over here who is one of our heads of marketing. And today we are going to be discussing growth. What does that mean to grow a business in the tech space? What are the pitfalls of growth? What are the misconceptions people may have about it? And I'm sure this will be quite a lively debate because everyone here is very opinionated about what growth means. So let's, let's, let's start with what growth means. Let's, let's start with the scope itself. What is growth? What is this? Is it user growth? Is that what we're talking about here? Is that transactions growth? What is it?

Jasper: So my definition of growth is market share growth and market share growth. Marketer shares always based on two things. One is penetration and the other one is purchase frequency. Penetration X purchase frequency is market share. So that's my definition of growth. Growing your market share in your particular category.

Crystal: I think for me, the market share analogy works with the already established businesses. But given GOJEK is in so many, I guess newer verticals I think of growth as the expansion and education as well as penetration of the true awareness of your product in the market. And that would cover things like whether or not they understand what your product does, whether or not they find it easy to use and whether or not it truly solve the problem that they consider a problem.

Nadiem: Okay. Well, I think that one of the biggest debates about how to grow a business is how much effort do we invest in spending money to grow the business and how much effort do we spend on non financial improvements in the product or experiments, right? This is an ongoing debate, both take engineering resource, which is the most valuable resource. One takes a lot more money obviously than the other, but both are valuable. Where do we strike the balance between financial incentives and non financial incentives in growing a business?

Crystal: I think incentives can give you the jumpstart to get a large enough user base to understand how users are interacting with the product. So promotions or discounts can be used to leverage a position in the market and put yourself against the competitor and move their user base to yours for that given point of time. But I think that to continue spending, uh, without a clear strategy or a reason why promotions and discounts are really a tactic and tactics should be applied to a longer term strategy. And so with growth, what we do is we will use vouchers or discounts or promotions in order to attract users to complete an action or encourage or incentivize them to do a certain type of behavior. Whether that be a topping up for the very first time or doing your third GO RIDE transaction, but you still need that little top of mind nudge in order to complete that behavior. But the behavior itself in this example would be part of a longer term strategy so that you can evaluate whether or not that third transaction really is important and not just for the sake of trying to get people to three transactions.

Nadiem: Is there like a specific moment or threshold by which you decide, okay, now it's time to spend money on this product and could you spend money on a product that is actually extremely wasteful to spend money on it initially? What? What's the criteria to use whether? Alright, now it's time.

Crystal: I think when enough users understand what the product is, then you can start spending money. If money is being used to in a way to replace a go to market strategy. I don't think that that's a good way to use money. For example, I think, you know, movie pass. Are we allowed to name exactly?Oops.

Nadiem: That's fine.

Crystal: I think movie past as an example where you know, you are using a high subsidy model just to get people in the door, but that the strategy never changes or it never evolves to pivot users to different types of behaviors. So for us, for example, I can say where we decided not to spend money when we run on experiment, we have an understanding that, you know, many of our power users use us in the morning rush hours. Now that hypothesis as a group, we discuss and figure out that morning rush hour is actually the longer commute time than evening rush hour. Because most people have to get to work by a certain time. But most people can, you know, wait to leave work at a time that is less, uh, I guess has less traffic. And so we assume that given user's experience a high amount of core value during morning rush hour because we were saving them so much time because we get them through traffic. That is how we get the best power users. And so we created a feature to provide vouchers that were only usable during morning rush hours. Now to test whether or not that incentive worked. You're not really, you're not really testing the incentive itself. You're using the incentive as a tactic to understand the strategy of using early morning rush hour as a way to acquire new power users. And so when we ran that experiment, you have an AB test where the control group gets to use the voucher at anytime of the day or only during the rush hour period. And then the experiment group is only allowed to use that voucher, let's say, you know, between the early morning rush hours. But what happened was, you know, our users did not actually increase their power hood, their power user likelihoods. And so that was a signal to us that incentivizing that behavior isn't worth it. And so we killed that experiment.

Nadiem: Right. You just, just for it to kind of highlight the scale of magnitude of what your product group does, how many experiments are actually running right now today on the GOJEK Platform?

Crystal: Oh Man.

Nadiem: Give us a number.

Crystal: Dozens, probably ...

Nadiem: Dozens, right now?

Crystal: We launch at least one experiment per day on average.

Nadiem: On average, a new one?

Crystal: Yes.

Nadiem: But then experiments are running in parallel, parallel all the time with different segments, different price points, et cetera. And so basically, uh, I think what's critical, I think for the audience to know is that growth is such an important part of GOJEK's success and potentially failure if it doesn't work out well, that we've created an entire product group at the same level as like food, like food delivery or transport called, which is what you lead growth platforms that rapidly becomes the voice of truth. And instead of Nadiem telling crystal, "we need to launch this kind of a promo on GO RIDE using like GO PAY vouchers et Cetera. Crystal is the person who goes, wait a second, you are not the voice of truth. My experiment will be the voice of truth and let's see if your assumption is correct.

Crystal: Exactly. Yeah. And I think that a lot of that investment, um, is multiplied their impact multipliers across the org helping in this from the performance marketing teams who run, you know, app store optimization experiments to the data science teams who help with creating the frameworks by which we, uh, constitute whether or not something is successful or statistically significant. Um, so there's a lot of work cross functionally that goes into this.

Nadiem: Okay. You mentioned about performance marketing there. Um, this is, this is one thing that I think is very interesting. Jasper, we've had multiple debates about this and I think you've shared a lot of insight on this. Conventionally. Tech companies have, a lot of them have this mindset and we used to have that mindset as well in the beginning that digital marketing, online marketing is by far a superior form of user acquisition and awareness because it's attributable, you can directly link it to an install or a conversion on our platform. And many people believe that, you know, offline marketing, TV, magazines, billboards are a waste of money because you can't attribute it to shit. Right? So, um, what is your opinion of this and is that true?

Jasper: So like I think, I think there are two things in what you're saying you like one is attribution. Can you attribute a action to a certain channel that a user has seen and the other one is measurable. Like yes, you can measure digital clicks, digital actions, but the attribution, if someone downloads our APP and like books a first trip via a promo marketing campaign that we ran on digital, it doesn't necessarily say it. 100% of the attribution should go to that digital ad. I could have told you many times, hey, start using our product because we've got a fantastic product. And then you run into an ad, you click on it and you start using it. So that means I could have told you a million times and the attribution is more on me on the offline telling you instead of you clicking on that APP. Um, so that's more of the attribution side over the like the measurability of things. Um, I think if you look...

Nadiem: So just because I served an ad to you digitally doesn't mean that when you convert it, it's because of that digital ad

Jasper: of course not

Nadiem: it come from a variety of different offline things. Even a referral from a friend and that lies the complication with attribution, even in digital marketing...

Jasper: Correct. Attribution on a channel by channel base is not possible because it's too complex. People are very irrational in the choices they make. People buy Nike shoes, not because they've seen an ad, but because of like 40 years of brand building. Uh, they, they make a certain decisions like how do you attribute that purchase to a certain campaign that you've run in the last few days? It's impossible.

Crystal: Even see that in the data as well. When when marketing ends up spending a lot more for a specific season, we'll see organic go down and the multitouch, uh, touch points will actually show in the, I guess performance channels that, you know, it looks like a lot more users are converting. And our, our, I guess the CAC is much lower because they have so many offline touchpoints.

Nadiem: CAC is Customer acquisition cost.

Crystal: Yes. Customer acquisition costs for the uninformed.

Nadiem: So you're saying that actually the effectiveness of performance marketing, performance marketing is just another way of saying online marketing,

Jasper: Yes because all marketing is performance marketing.

Nadiem: Yes. Sorry. So let's just say online marketing. So most companies and probably marketers are overly attributing the effectiveness of online marketing. Right?

Crystal: Yeah. Absolutely. Yep.

Nadiem: So they're taking credit basically for something that could be not caused by the campaign itself. Okay. That's really interesting. So what do you...

Jasper: And the only reason why is because it's measurable. So people feel confident in that excel slot that I guess okay. We've increased our user base because of this concern.

Nadiem: So the marketers and the planners of the marketers that have to take accountability to the CFO and the budget can clearly see, I spent this much and this is how much transactions I got. Right, and I can attribute it that when I serve this to this ad, this is what happened.

Jasper: Correct. If you look at like this isn't, this isn't really...

Nadiem: When that person could have watched a TV ad or seeing say a GOJEK driver on the street or a friend recommended or they saw a friend order GO FOOD and he just happened to be served that ad and then decided to download the APP and click when it had nothing to do with that.

Jasper: Look at what we did in Vietnam was probably the best example. Or in Singapore, we enter a market, there's a lot of noise. By doing PR on news or where ever it is and we see our user base grow when we don't have to spend a single dollar...

Nadiem: Without digital market, without significant ....

Jasper: Without any market. Well like you could technically say PR is marketing as well. It's just noise noise around your brand. Whether it's paid for, it's free or it's, it's in digital. It doesn't matter. Like consumers will react to what they see and they will, they will act on it. Whether it's on digital, offline or PR, whatever it is. But what's interesting though, now that we talk about like digital marketing, I'm a big fan of digital marketing by the way. So the things that I'm saying is not to diss digital marketing at all. Uh, but if you look at the last 10 years of marketing effectiveness, the moment these big digital companies enter the marketing scene, you saw a shift in the average effectiveness of marketing campaigns. It's going down. So the effectiveness of marketing as a function is decreasing. And one of the biggest reasons for that is, is what they call short-termism. Short-termism is only looking at short term results instead of long term brand building in longterm results like web

Nadiem: What you're saying is that the window of assessing effectiveness matters greatly.

Jasper: Correct.

Nadiem: So if you are just looking at things from a one or two month time period and assessing impact, then you're, you're not actually occurring the true impacts because there's a delayed,

Jasper: There's a massive delay.

Nadiem: How big are we talking about?

Jasper: So I don't think there's a direct standard, but when you do, when you do a brand building campaign, uh, say we do a big TV campaign, the actual results of that, um, go on for like six months on average. So you do a four week long, big TV campaign, outdoor, digital, whatever it is, entire campaign. The results of that might have a long tail of around six months on average.

Nadiem: And you think most marketers just observe month to month numbers?

Jasper: Yeah, because like...

Nadiem: And therefore the numbers kind of lied to them in a sense.

Jasper: Again, if, if you, if you look at, uh, a board, right? Do you have a CEO, CMO, CFO, and a lot of other Cs. The CMO has to the shortest tenure of all chief level people simply because I have to, um, deliver results now. And if they don't, they get fired, so it's all down to short-termism. But if you look at, when someone needs to... Look at Coke,

Nadiem: Sorry, what's the term again? Short-termism.

Jasper: short-termism.

Nadiem: Shortism or short-termism?

Jasper: Short-termism. If you look at like Coke, right? One of the, probably the best examples of, uh, uh, marketing brands, brands that are 100% on marketing like Coke as a, as a product doesn't fulfill any functional need whatsoever. It's very irrational to drink Coke. A lot of people do it cause it probably tastes good. Um, Coke was doing like poorly up until, I don't know, 2011. They saw their like, sales, like decrease. The new CMO stepped completely changed the strategy, but the impact of that was only shown like two, two and a half years later because you're changing like completely oiled machine. Um, in terms of how people see your brand, feel your brand, how easy it is to buy, how easy, how easy it is to recognize stuff like that all matters.

Nadiem: So this, this super urgent need to prove, I guess, "attributability" right. If that's even a word to prove that I can attribute every dollar spent, which is the great golden hope of digital marketing. Right. That's what the mantra has always been about digital marketing has actually skewed the perception of marketing away from investment into longterm. Into short term gains, right? Because the framework is still small. But then why, why even invest? So a lot of, you know, a lot of tech companies have completely dismissed TV, say, oh, it's just a waste of money, et cetera. Do you agree?

Jasper: No, of course not.

Nadiem: Why?

Jasper: But like, okay. So if you, if you look at a user base of any company, right, not just our company, any company, it consists of, uh, people that buy our brand every now and then. People that buy our brand every week. And people that buy a brand every day, the majority of your buyers are none or very light buyers,

Nadiem: People who use it infrequently.

Jasper: Yes, yes.

Nadiem: There's a natural distribution whereby the majority of your users always are the less frequent, correct.

Jasper: I don't know the exact number, but it's a very big group.

Nadiem: It's the majority.

Jasper: It's a majority, the far majority, right? So, uh, that the majority, uh, needs to be rich and that's a very big, but we're talking about like tens of millions of people, uh, needs to be reached on a very frequent base that whenever they feel like, "Hey, I am, uh, I have a desire to like buy a cup of coffee or hey, I want to get from hub using a ride share option." That is the moment that, um, that whole group needs to think about a brand, our brand first instead of our competitors. You can only do that by using mass marketing, uh, over a longer period of time and building those back brand attributes, which you were talking about like in the beginning, like educating large groups of people. On what the like consumer benefits of certain products are, uh, why you should like take our brand instead of our competitor brand and you can only do that, uh, like the, the biggest group to influence is that lighter non user group and the best way to serve them, it's you used mass marketing and that is in this market the biggest challenge TV. It's as simple as that.

Nadiem: So above the line, marketing has by far the highest reach, it is like TV. And therefore it is affecting disproportionately the biggest number of users that you need to move up to the higher frequency, uh, items.

Crystal: Interesting. But until without a way to very clearly understand what that budget translates into, right. Because there's so many things going on in a company, especially with us, you know, we have multiple campaigns running, we have multiple promotions running. So how, how does one know how much budget, let's say marketing should get, or this campaign should get, like what are the best practices there?

Jasper: Good, good question. I think if, if like if you look at companies like Coca Cola, right? The majority of the money goes into marketing because there are 100%, we're not hunting, but they are 90% marketing companies, the majority of the budget we'll go to that. Um, we are slightly different company because we are still like tech company, we're product driven company. Um, so in order to say like, how much money do we need to spend on marketing versus something else, I don't know, looking at our user base and that like the frequency that we want to reach them, say on a weekly basis I can say that we need x amount of money to do so. Um, or like this is the entire category. How much would it cost to reach them on a weekly basis? Say twice across all the channels that we have that would cost you x amount of Rp. So for a full year, it's this amount.

Crystal: I see.

Nadiem: One of the things that was quite interesting when we had this debate before was this concept of previously we thought that if you hit a user again and again and again with a given image? With every incremental intervention, you are increasing the probability of that user to convert and finally become a user like say a GOJEK user, but in reality given a limited amount of resource and if you're paying for these ads, what you are telling me or correcting me on was that actually it's extremely wasteful to try to attack that same user again and again again to get to them to convert, but it's much more effective to spread wide in order to, because there is a kind of almost like a fixed probability, random probability of someone converting. So widening the number of people that see that is the most important thing when you're targeting growth. Is that right?

Jasper: Yeah.

Nadiem: That intuitively makes sense, but then why is there such a different conception than why are people paying digital ads to keep hitting the same user again and again and again? Based on audiences and segmentations. Isn't that wasteful based on your...

Jasper: 100% do you want to, do you want to get the real reason for that or the fake reason? So the real reason for that is those people that sell us those ads, right? Those big digital companies, the easiest way for them to make money is to reach a user over and over and over again

Nadiem: because they already have established those audiences.

Jasper: Correct.

Nadiem: And they want to keep monetizing because if they have to capture new users all the time, they might not be able to know where to find them or they might not have enough.

Jasper: So if you look at like TV channels for instance, um, it's similar to our um, buyer base. So you have daytime TV people, like people that watch daytime TV, they watch a ton of shows, they will watch like 10 hours a day. And then you have your prime time TV, say a live football game, which attracts millions of users, but they only watch TV say every once a month. Very big group of light TV consumption happens in primetime for that TV channel. It's much easier to sell ads to those people that spend 10 hours a day because they can just serve them ads and ads and ads and ads but that doesn't mean that that group is the right group to target at because in that prime time football game, you have a lot of like light TV viewers who are more expensive to reach via TV.

Nadiem: Right.

Jasper: So if you talk about frequency versus reach, uh, and why, um, these big digital companies would say, uh, you need to invest in frequency is because it's easier for them to reach those heavy digital platform users than it is to reach those slight users. Does that make sense?

Nadiem: Yes.

Jasper: But from a strategic point of view, always invest in reach instead of frequency. Uh, there's been a ton of research into that. If you spend that reach has a up to 2.5 x higher return investment than investing the same dollar in that same person.

Nadiem: Wow. So if you look at a curve of user frequency from nonuser here to power user here, you want to spend all your money down here, you want to spend all your effort and money down here. Right? And that's the biggest bang for the buck because that's really interesting. So, so what does that mean for, for growth platforms, right? What does that mean, uh, for GOJEK or any other tech company? Like how do you re engineer a product and the marketing of that product to gear towards new users are very low frequency users. And do you think that makes sense?

Crystal: Yeah, I think that the same principle, it does apply that you want to focus your efforts on increasing the frequency and conversion and top of the funnel of those low frequency users. They are the ones who will return the most value for every dollar that you spend given they are doing less transactions. They have done, I guess they've maxed out, they haven't maxed out on their ceiling of usage yet with the platform. And so for us, we do see a better, you know, cost per incremental transaction. That's one of our KPIs on the growth side. When we invest a dollar into a user who is doing, you know, one or two transactions versus someone who is doing, you know, 20 to 30 transactions because the amount of gain that you are able to get out of a low user, uh, frequency is much higher, right? There's, there's much more behavior to shape and mold.

Nadiem: So getting like a GO FOOD user who just uses, uh, GO FOOD maybe once a week if you just get them to, and they'd say they use it only on weekend, but say you get them to use it for a lunch on a week day. That's already doubling the number of transactions. Whereas like, uh, my cousin who orders GO FOOD literally every night, you know, even if you're a, he ordered it like probably like five times a week, right. Getting them to order one extra order is not going to necessarily move the needle and it's much harder to make him buy that extra order because he's already empowered using.

Crystal: Exactly. And plus

Nadiem: So should we just ignore our power users so we know what you're saying?

Crystal: We don't ignore the power users. I think that the power users will symbolize a set of Aha moments that they achieved in order to get to where they are today. And so we can use power users to better understand what is the maximum utility that a user can gain from our platform. And how does that manifest in their different behaviors? Was it someone who first used GO RIDE during rainy during the rainy weather? If so, you know, perhaps that's a good time for us to maximize our education or a new user acquisition marketing because this happens to be a good entry point for the types of users who are going to convert to our platform and find it useful for a long enough period of time during the rainy season. So I think it's better for us to look at power users as a way of providing the base of understanding for us to better convert the low engaged users.

Nadiem: Can you give me an example of how the story or the journey of a power user could inform how to transform a low frequency user into apartments? What's an example?

Crystal: So I think that it will show in cases like, uh, your cousin where they are using this every single night because it provides an option for them to buy food while they're on the way home already. It's become a habit for them to, once they get into the car, they open up, GO FOOD. They look for a merchant that they know is near their home and they book while they're on the way there. Now we look at that type of power user behavior and we realize that this is a potential use case that uh, our users who are light touch users perhaps have not created the habit for yet. And so we create features in the transportation page to show a user who is traveling somewhere. Hey, here are some different food options for you that are near your destination. We surface that behavior or we guide them into the behavior with better product features because our power users tell us what they want our app to be. They tell us how they use it and we should be creating features that actually eliminate the friction that they're manually doing as a power user.

Nadiem: So that's super interesting. I'm a GO FOOD user. Uh, and I always book it, let's say I'm a power GO FOOD user and I use it every time on my way back while I'm riding my GOJEK home, okay. And you are also a power GO RIDE user. Every day you go back home but you never ordered GO FOOD. So the algorithm or our models basically say there is x percent probability that if I start giving you, for example, like a voucher for food delivery during this time that you may build the same kind of habits that I'm building and therefore potentially become like me.

Crystal: I think that what it does is give us us an insight into what are the use cases that we should reduce the friction for our low engaged users. Right. And it doesn't even have to be a voucher. The voucher can be used as a way to test whether or not users will even adopt that behavior because the voucher removes the existing friction of, I don't even know what merchant I want to eat at. I don't know what merchants are near my destination. And but the voucher will serve as a way to force the product, um, away from that friction. Right? It removes the friction because it incentivizes the user to just get past that friction because they got money.

Nadiem: I've never ordered food delivery before and that's a friction. That's a mental barrier to try to get me past that.

Crystal: Exactly. Correct. And so once, once you see the uptick of users adopting or let's say redeeming the voucher that has a one hour window expiry from the time you book your ride, then you start to be able to prioritize that this feature should be built, that we should create the feature that will make the product so great that you don't need the voucher to remove the friction for the user.

Nadiem: So if you draw again that chart where you've power users, you've got low frequency users and non users here on the top side, these guys over here, this is your basically your role models. These are the guys you dissect, analyze and see which behaviors and Aha moments have they experience and then you apply it to the low frequency users but they're already, then the low frequency users are the ones you want to move up this chain, right? So these guys have a much higher uh, or lower cost per incremental transaction than someone may be in the middle between a power user and a lower user. But then why even work on this isn't the biggest return on investment really the non users. So completely new users. Like why are we just spending all of our effort on non users and getting them on our platform?

Crystal: I can give you an example from our experimentation framework. So there's a limited amount of new users you get every week. We actually run out of customers to run experiments on from the new user base.

Nadiem: So you'll hit a wall.

Crystal: We literally hit a wall. Okay. And so that part of it is just there is an operational wall from uh, how much money we could spend on these users. But beyond that, I think it's more about creating user life cycles. There's not just black and white new user and power user, right? There are a lot of stages in between that and allowing the team to focus on and identify what are those different users stages. We'll give them the insights around how to create those step-by-step journeys for a user to become a power user perhaps. Right.

Nadiem: Can you talk a little bit about user journey? This seems to be quite a buzz word, but in the growth perspective, like how many user journeys do you have mapped out right now across the GOJEK platform?

Crystal: I think it depends who you ask. I mean again everyone is, has pretty open access to the data and the research teams and so they're free to create their models for what is the best GO FOOD user, what is the best transport user. And then we have platform wide user states, uh, which we work on with the data science teams, with the research teams, the marketing teams that...

Nadiem: Platform wide means across, all of the Super App.

Crystal: Yeah, exactly.

Nadiem: Right. So GOJEK's quite unique in the sense that it's not just a journey within your GO FOOD journey, but how does the user behave across multiple services and how we can integrate that. And that's this concept of the Golden Triangle users. You know, we've been refining this concept for some time. Maybe you want to explain why Golden Triangle users are so special.

Crystal: Yeah, I think that, you know, every company will have a very easily communicatable Aha moment for the company to rally around. And this simplicity beats out the complex, more even more accurate, uh, Aha moments. And so Facebook has, you know, seven friends in 10 days. We have, you know, multiple, at least three use cases within 30 days. Right?

Nadiem: Right. Three different services within 30 days and using GO PAY.

Crystal: Exactly.

Nadiem: That's kind of the holy grail.

Crystal: Exactly.

Nadiem: Our turn rates drop significantly. You become a user for life and that's an Aha moment. So just for the audience that may and listeners that may or may not understand the concept of Aha moment, an Aha moment is a discovery basically based on analysis of user behavior that most of the times you would never guess, right? It's kinda hard to guess, but there's certain things that happen in the user journey that lead them to either end up using this product consistently or it could be negative as well. It could be a series of negative events that end up churning the user out. Right? So you can have a positive aha moment and you could have a churn Aha moments

Crystal: Absolutely, we have those negative aha moments and the framework that we use is actually what is it that retained users do that churned users do not do. So it's a disproportionate, I guess, uh, what's the word when you like, could compare and contrast between, you know, what do you retain? What does the majority of the retained user base do that the majority of the churned user base does not do? Right? It's even stronger than just pure correlation of what do most retained users do? Because you could say that, you know, a thousand trips, it is highly correlated with retention, but most of our retained users don't do a thousand trips. Right. Um, so what we're really looking for is the biggest overlap between the most of our retained users that the most of the churned users do not do. You're looking for those types of behaviors. And when we look into each service type, I think that the reason why we have a, an organization that is both embedded and core is because there are two pieces right to every uh, I guess marketing life cycle. There's always two pieces to every product development life cycle, right? Where GO FOOD wants the best possible GO FOOD users ever and they what they consider to be the top of the funnel of awareness, understanding, consideration, uh, et cetera falls under things like purchasing GO FOOD or using the search feature or using the categories feature or using reorder. There is a depth to the product of GO FOOD that perhaps on a platform level we don't necessarily consider right. On the platform level we consider awareness and consideration as things more shallow. Right. Does the user use GO FOOD and GO PAY? Do they use pay later? Do they use a multiple send service types and shop service types? Because the way that we have to do the, I guess the structure of the growth team would be more around is the user highly engaged in the platform versus in that particular product. And I think that there are a lot of fair trade offs to debate between the two.

Nadiem: That's really interesting. So the signals for retention not only is defined by how many different services on our Super App that people use, but within each of those services, how deeply penetrated are they on the feature settings is and why I, that's to me, I think the logic is that the more invested I am, the more I penetrate the feature sets of an APP and the more I use it on a regular basis, the more psychological investment I put into this platform. And therefore the more attached I will become to it. Is that generally true?

Crystal: And like honestly, hopefully the more core value you extract by using more of those features, right. Otherwise those are bad features. Um, but the more times that you use, let's say the on the way home food service to see, you know, what food is near destination or the more times that you use, let's say the categories or the recommended food items, hopefully the more value you are extracting from the platform.

Nadiem: Fascinating. Let's talk about a little bit about market share and its impact on product growth, right? Is market share something? Why is it important and if so like how closely do you need to monitor it and follow it? Right? Um, in certain high subsidy environments, market share is extremely volatile and it can go up and down depending on who's running up from what any given time. Right? But how do we think about, or how do other startups also think about market share, uh, as a signal either leading indicator or lagging indicator of growth. Like how important is that?

Jasper: Key. So like if you.. Well like in the end, you as a brand, whatever, what your brand may be. The only thing you want is to grow, right? Growth is the fundamental driver of your business. If you don't grow and your competition grows faster, they will outgrow you and in the end crush you it's as simple as that because they will earn more money than we do, um, and they will have more resources to like again crush us.

Nadiem: Well that's not always the case,

Jasper: Over the long term?

Nadiem: Over the long term, yes.

Jasper: That's always the case.

Nadiem: In technology space probably more acutely

Jasper: In the short term it's, it's different and that's just what we discussed earlier. Like it, it depends on what your strategy is. For any, for any other company probably apart from ours is you need to uh, drive penetration, meaning getting new users in on a weekly, daily, monthly basis in order to grow your market share because when you grow your penetration, the average purchase frequency of your consumers, meaning loyalty will increase with that line.

Nadiem: Say that again.

Jasper: When you increase your penetration of household penetration, so like the number of users you have on your platform, when that increases, right? There is a direct correlation with an increase in average frequency of your consumer base. So the loyalty will increase. And the reason for that is the bigger your base becomes, the more light users you have in your base, which usually are hundred percent loyal user because they only make one purchase in a year. So the bigger the base is, the more loyal that base becomes. That's the law called double jeopardy marketing. So the higher your penetration is, the higher your, uh, average purchase frequency of your consumers is. The delta between the two is that the difference in penetration between your competitor always be bigger than uh, in the after purchase frequency that's relatively small. So in order to grow market share, you can only do it through penetration and the average purchase frequency will follow. It's actually pretty simple if you think about it, right? If you're a nonuser, uh, so you've never tried the product before, you become a first time user, which is the biggest group. And then you either decide, hey, I like this product or not so you move down the line of one purchase, two purchase, three purchase and to become like that thing that we call power users. So the more people you get from zero to one, the more likely it becomes that someone turns into a power user.

Nadiem: Because assuming everyone over time on average has the same probability of becoming a power user.

Jasper: Well like if you look at like, um,

Nadiem: It's not like all the people in the beginning who adopt early are the only power users. That's not how it works. Right. You could have a really late comer come from the back and still have an equal probably probability of becoming a power user. They just discovered it late in their stage.

Crystal: Every cohort averages out. We've seen that too.

Nadiem: So you seen that in the data. Can you just explain that for a second.

Crystal: Every cohort there is a base line of given their spend in the first month, we can project what their spend will be three months down the line and there is just...

Nadiem: And those ratios don't change.

Crystal: Those ratios don't change. But what does, what hasn't worked is on a per user basis. So on a cohort level, yeah. Uh, I think that, you know, Jasper is absolutely right because we see the users as an aggregated base. They come in at the same, uh, I guess density or distribution of types of users or their capacity of power usage. Assuming you don't change significantly the different channels of acquisition, you have a pretty clear baseline from month one what they will be like in one three as a cohort.

Nadiem: As a cohort. That's really interesting. Because intuitively you would assume that all the power users are getting early - early adopters, right? But that's simply not the case.

Jasper: No.

Crystal: And I think what we do see though is that early users aren't the best users to research their motivations for using the product or how they want the product to evolve because they are of a very different user type. I think the early adopters, whenever we do research or user testing on let's say homescreens or on new features, we are very careful to have a mix of people who have never used our product before as well as loyal users as well as people who, you know, are in the middling because users who are later adopters tend to have less input on what our product should look like or what features they want because they don't understand the product and they aren't part of that early revolution of people who had a desperate need for the solutions that we provided. So they think that it, this is more of a problem for research or evaluating the decisions that we make on a new feature basis.

Jasper: And look in this is where it becomes interesting when we talk about what we call loyal users or power users. Because like, we share our buyer base with our direct competitor. Like it's the exact same people, exact same demographics, exact same psychographics. It's exactly the fact that it's um, it's basically the same category group. Um, so when we talk about power user, and I'm gonna take Vijay as an example for this. Vijay is a sneaker head is a massive power user when it comes to sneakers. He has a preferred brand.

Nadiem: Vijay, our creative head?

Jasper: Vijay our ECD of our creative head, he loves sneakers. He's an absolute sneaker head. I think he purchased like a new pair of probably every, every week or every two weeks.

Nadiem: I can tell man, very trendy.

Jasper: It's ridiculous. He probably has a room the size of this with all his sneakers in it. And he has a preferred brand, and his preferred brand is Nike. But that doesn't mean that he'd occasionally buys Adidas as well. Like he, he probably has on every 10 pair, maybe seven pair of Nike and three pairs Adidas. So when we talk about like the misconception about loyal users is that a loyal user of a brand only buys your brand. That's not true, simply not true. Like of all Coke users, I think it's like the number is 67% also drink Pepsi. Like who would look just,

Crystal: Yeah, that's not intuitive to me at all.

Jasper: That's not intuitive, it's like that super counter intuitive, but that's, that's reality. Like loyal users as we as the definition of saying, okay, you only buy one brand in the category, do not exist. Actually the more the more power user you become, the more brands you use. Like I said, I can't if you're a very light user, so you only use the only buy pair of sneakers once a year, you will probably buy the same brand over and over again because that's the only brand you really, really know in that category.

Nadiem: So let's think that through. So what's the mechanism that's happening there? If I am using, if I'm a power user of a ride hailing app, let's see, I might find myself in situations whereby I can't get a ride higher frequency and therefore I might try the alternative uh, service. I would also be a lot more conscious of the promotions that are going on at the time

Crystal: You are a more educated user.

Nadiem: I'm more educated users so I am aware of when the other side is promoting and not and I may want to optimize even though I have a general preference for one brand. Um, so it becomes a bigger part of my life and therefore my propensity to dabble in competitor products is higher. Whereas if it's a very, or a smaller, a relatively smaller aspect of my day to day behavior, I don't need the cognitive load of having to think, you don't your optimize all of them. So what you're telling me is that winning those light users is more important because they are probably just only going to use you.

Jasper: It's fundamental for, like I said, for two reasons. One...

Nadiem: Seems so counter intuitive to what everyone knew about power users

Jasper: and that's that's, that's so cool about that

Nadiem: and you're blowing our mind right now. It's confusing. So power users are to study, to learn but don't spend money on them. Then why are there loyalty programs?

Jasper: Great question. They are, can I say bullshit? They are complete bullshit.

Nadiem: We have a loyalty program too. So in general loyalty programs are bullshit, why?

Jasper: It's a waste of money. Because they're very expensive. Has anyone ever done the analysis, take a disability and pump into it. Not just from a marketing perspective or from a overhead costs, investment in tools like the whole shebang. And what the outcome of that.

Crystal: Couldn't you argue the same thing on offline marketing that Oh, it's not attributable, but you need it. Like it's just, it's a hygiene for a product.

Jasper: So how we, how we attribute it, uh, it like that. And that's what I'm saying. It's not channel by channel. You look at the total investment over a period of month, uh, in TV, outdoor digital combined. And look what the outcome is. And that's similar to loyalty programs like you look at, we should look at the total investment going and what the ROI is that comes out.

Crystal: But what if it's more long term facing, right. What if we have the same problem of the attribution being much more future? Uh, I guess recurring.

Nadiem: It also depends on the level of engagement to that program that some people have. Like, for example, I'm super happy in my airline smiles program and I will make sure that I don't lose my status, by even changing some of my dates in order to make sure that I can hit that airline. So yeah, it's not, you know, I'm not gonna go out of my way to increase my travel time, but push comes to shove. I will try to always take that airline in order to maintain my status. So I don't know if I want 100% agree with you, but maybe,

Jasper: sorry, I'm getting really excited about this.

Nadiem: OK. Go ahead. Go ahead.

Jasper: So if you look at airline programs, right, you're probably talking about Singapore Airlines, the majority of the people on Singapore Airlines and which statuses buy their flights based on company budgets, it's not out of their own pocket.

Crystal: Oh, absolutely.

Jasper: So if you look at, if you, because I was a huge fan as well, like in my previous companies I could fly business class all over the world and hey it was awesome. I get like massive amount of miles. And I went like from, what is it like gold, platinum to what or whatever else, and I was like, this is so cool, but the moment you have to book your own flights, pay for it from your own wallet. I didn't fly SQ anymore. I just picked another like pick another airline. So how does that loyalty work if it comes out of your own pocket?

Nadiem: I see. So you think there's a generally an overvaluation or overvalued nature to loyalty programs in general.

Jasper: A hundred percent, a hundred percent it's like it doesn't... The cool thing about the SQ, and that's actually the only one I really, really like because it's actually a marketing tool, not necessarily to change behavior, but to show, I mean like you would tell your friends like I do. I'm like I'm a Gold member, which elevates ...

Nadiem: I don't but I love my SQ miles, I also love my Garuda miles, I'm all types of miles.

Jasper: Like whatever your miles status is. That gives you a bit of bragging rights. Again, no large user would give a crap about that, but like power users will give a crap about that. But let's think about like, uh, like a simple coffee shop, right? Use me a power user, like I go to Liberica every single day here and I get my Cappuccino. These loyalty programs at some coffee shop would say, okay, you get your 10th cup of coffee for free. It's basically 10% discount. Why would you do that? I would go to, I would buy the 10th cappucino anyway. I don't need to be incentivized for that. You know what I mean? That's a loyalty program in itself as well.

Nadiem: That's a tall order because that's considered one of the most successful loyalty programs of all time. By you're saying they would have made that purchase anyway.

Jasper: So you're giving me a 10% discount or nothing. The reason why I like,

Nadiem: But here's the thing, I think what you're missing out, despite the fact that I would have ordered that coffee or book that ride or book that plane anyway, there is a massive component of net promoter score in the feeling I get from that additional perk. I would say that, you know, my willingness to promote that brand to another person, which by the way, for all of those out there people that don't know, ultimately your product is going to grow not from marketing but from word of mouth. No matter what it is, it's going, the conversion point will happen through word of mouth. Ultimately that's the stage. Feel free to challenge me on that.

Jasper: Yeah, I know. You're absolutely right. Word of mouth is one of the most,

Nadiem: Wait, wait, let me finish my point.

Jasper: Okay. Sorry. Go ahead.

Nadiem: So because of that and because of the importance ultimately of word of mouth, whereas all the ads that you see everywhere, it's just reinforcing that moment for conversion. The fact that if loyalty programs can make you feel special, a special connection to a brand, I think it can very powerfully increase that word of mouth score or the NPS (net promoter score) or whatever it is. That would be my perception. But let's, let's talk a little bit about the word of mouth because you know, people are still skeptical. People think. I think a lot of not only people who advertise, marketers, even growth people, um, founders, all the same, and I was also guilty of believing this previously is just like, this is how a user that they look at an ad, maybe on a social media site, et Cetera. Um, they look at it at a few times and that's all they're looking at. And then at one point they click on that ad, go to a APP store, downloaded, and then I am that user, right? This is apparently the furthest from the truth that could happen, right? That's not how it actually happens when we actually did a lot of this research on what happens when you ask a person, how did they start using it? By the service. I mean, I don't know exactly the number, but more than seven times out of 10 that person says, oh, I heard, saw what my friend talk about it or use it in front of me or go run. And it all happens in the offline space and not in the billboard offline space or TV ads or and literally coffee shops, family gatherings, class, work environment. It is the unattributable conversations between humans that ultimately lead to that conversion. And you cannot avoid that. Sorry. You cannot increase that without having product excellence.

Crystal: Absolutely, Yep.

Nadiem: So at the end of the day, no matter how much money you've got, no matter how much cash you've got and how much promotions you're running, you will not be able to run, hack your way or spend your way around great products for those.

Crystal: Those aren't sustainable. You need the viral loops. I think viral loops help sustain those conversations. So for example, what Dropbox did was a natural extension of their product's core value, which was sharing and collaborating. Their referral program was a natural extension of that because, hey, you're already sharing this document. Why don't you join me on this platform as well? And that becomes a viral loop because you're already talking about working on a document and this discussion about, uh, joining this platform is a very natural evolution. Therefore, so I think designing those viral loops to match the organic occurrences or are super important.

Nadiem: And let's see, like what are some of the things that are content worthy for viral conversations? Getting a promo or discount generally, no one really brags that much about getting a discount, right? You don't go to work and go like, Yo, I got five rides free. It's amazing. You know, like, and that doesn't happen as much as you would think. Right?

Crystal: Unless. The Promo is buy three, get four, and then you start telling your friends, Hey, I wanna buy this food, but I need three more people. Who else wants in? Those are the viral loop.

Nadiem: Bingo. It requires participation of someone else to do it. Sending a referral code and then being able to get it back, sending a voucher, something that requires an engagement or a conversation. But product feature development has that same effect because it's like a mini release of a product or service. So there is a new thing even though you use it once every 10 times within that product experience, once you found value in it, height, that's a, that's a, that's a buzz worthy topic. Yeah. So product development has that very fundamental relationship with word of mouth marketing. It's not really marketing, it's just word of mouth.

Crystal: We just did a, we just finished, you know, some, some homescreen, research testing, user testing, and we got a respondent who is saying, uh, Hey, uh, you're, you're asking me about how I feel about GO NEWS. You know, I never, I've never actually seen GO NEWS on my platform, but someone was telling me about a new story that they read on GO NEWS. I tried to look for it, can't find it, how can I get GO NEWS? And so this was super inspiring to us. It shows that, you know, you, you need those different product development tidbits that are fitting into the organic life cycle of a customer.

Nadiem: That's right. And there is no, there's no business model to GO NEWS. There's no cost to, you know, it's content on our APP. And so that's an example of a non-financial kind of incentive. Yeah. Right. It's engagement on the platform.

Jasper: I just want to tap into that word of mouth thing, like you said, like within the marketing tools that we all have, like I said word of mouth is probably the most important one in terms of like influencing a consumer's decision to make a purchase.

Nadiem: And so you agree with me?

Jasper: Yeah, I agree. But here comes the but. If you again like split between very light users, which is again the majority of any company's space and power users who are say uh, subscribers to a loyalty program. Very light users will not resonate with that loyalty program so much, so if you have a hundred consumers and five are your power users of that, uh, and say 80 users off that hundred are every light users, they will listen to a different kind of word of mouth than you were saying with your loyalty program. So you have five people talk about the loyalty program and the other say 95 out of a hundred would talk about other words of mouth. Hey, I've tried this product like for the very first time it's bloody awesome. Like that's a different kind of word of mouth. Then, hey I've got my 10th cup of coffee for free or whatever it is and that's that five only like a small base of your group who actually resonate with that.

Nadiem: I would stretch that even further. I think the propensity to share like a negative NPS is much higher from a power user.

Crystal: Oh for sure.

Nadiem: Right? Because if you're using it all the time, what are you going to complain about? Like what are you going to talk about your service? I'm not going to say about how awesome it is anymore. It's already become like, like water for me.

Crystal: Take it for granted.

Nadiem: I am exactly that user. I am this, I use, you know, GO RIDE every day, twice a day. And most of the times when I'll talk to someone about it, even if I'm the CEO of the company is to complain, I'm like, ah, why is this doesn't work, this is could've, should've been like this or is there, but you're right. And on the other end of the spectrum, if that, um, user, uh, had that light user had a good experience or the first time experience the potential for virality and word of mouth is higher.

Jasper: Yeah. And that group is just much bigger. Right? Like I said, if we talk about a hundred people, uh, usually like 80 or 90 people of that are infrequent users and only 10 or 20 are heavy users like you are. So that that group is just simply so much bigger than that small group so that the voice of that big group is much bigger than that small group so that 80 people are able to influence many more people than that small group.

Crystal: I would challenge that a little bit because even though the base of the new user group size or the light touch user group size, it may be by absolute volume much bigger. Are they more likely to speak about the product than the power user at that volume and at that level of credibility? I guess because as a brand new user, I guess I would have much less top of mine for a product or I would have just, you know, less invested. I care less. Right. Why would I talk about a product that I've just used once versus someone who is addicted to a product? Like, I'm a huge Guava Pass fan. I use it every day and I've literally gotten, you know, like three different people on my team to use it as well. But then you have people who maybe went to one class and they're like, oh yeah, I use it once. Like the, it doesn't come up in conversation. Like it's not something that they're not motivated to talk about it. Right.

Jasper: Let's test it like, because like

Crystal: I've always wondered how to test this ...

Jasper: That's pretty awesome, right? This, this could be a pretty awesome test because...

Crystal: I would like to know actually, why do we not focus on social? If we think viral loops and offline and word of mouth and people and communities are so important, why are we not investing more effort in the product development life cycle around social.

Jasper: Social as being a channel?

Crystal: Let's say I have a social feed on my app that shows, or I'm allowed to post. I just bought food from here. Just got 10% discount from there, just bought offline here, just dropped, dropped off here. Would you be able to use a product test like that as an experiment to see whether or not brand new users are they willing to post or not versus a power user? I'm willing to bet power users are more likely of course to submit to that.

Nadiem: In the spirit of being a scientist. I would like to make this a wager between you two so that we should experiment.

Crystal: We should design an experiment.

Jasper: Yeah. That's pretty cool.

Nadiem: I think that one thing that we're not factoring here about power users that I'd like to challenge you on is that GOJEK is very unique in the sense that the service that most people use, which is transport, food and pay happen in the real world, in the physical world, they are not things that are done in quietly on your own in your smartphone. Okay. In order to get on the back of a motorcycle or into a car, you literally, you know, have a motorcycle come in there. And when you book, even when you're in any social setting, everyone can see like I am literally, if you just walked down to our building, I can immediately see what everyone is booking, right. And which app they're using. Food is an even crazier part. You don't have to, a power user doesn't have to do anything. The power user just has to order food and it arrives and everyone else who he's with at that time, we'll see and pick up that him or her, pick up that food package in front of them. So when you have a business that fundamentally lives in the offline world, the word of mouth is not just a function of literal word of mouth, it's also visibility of the action and transaction.

Crystal: Driver jackets were like the biggest reason most new users knew our platform. Right. That's why it's super important.

Nadiem: Exactly.

Jasper: But that's marketing.

Crystal: that's literally marketing yeah.

Jasper: For me that's all part of marketing. Like the Coke rack like you, you will walk in the supermarket and you go down the aisle and the first thing you see is Coke rack but that's, that's marketing. That's like brand codes that been trying to imprint in people's brain. So again, the moment when someone wants to purchase a food, if I go food or like he's hungry or he wants to go from A to B. Marketing's function is to be that first brand that pops into their head. That's our function. Like our fundamental role that we need to play.

Nadiem: First to mind.

Jasper: Yes, first of mind in a purchase occasion. So not in general because I like when I say uh, what's your name... Name like the top watch brand. A lot of people will say Rolex but that's not like top of mind when it is not that relevant because when I'm going to buy Rolex is not in my set, but the top of mind around this is really hot. Right. So it's more like in like during that occasion moment, what is the brand that what comes up first? That mental share needs to be highest for us in order to win.

Nadiem: And mass marketing reinforces that.

Jasper: Correct. Mass marketing builds those. I mean, when I say coke immediately there's a red thing that goes into your mind. When I say Nike, you see the switch immediately in your head. So those coats, whether it's our logo, whether it's our color, whether it's shapes like that, like the Coke wave. Like those are things that we need to imprint in people's brain over a long period of time.

Nadiem: Well, good thing we were the first green jackets out there on the streets.

Jasper: But that's it. And like, but like, and it doesn't matter that your competitor might have the same color. That doesn't matter at all because like these coats together form your brand. I mean like a Twitter and Facebook have the exact same colour, sorta blue. But that doesn't matter because if I said Facebook, you would think about, oh yeah. Blue F, uh, Mark Zuckerburg probably there will be like your three brand cuts and with Twitter you would think, oh yeah. Blue, little birdie and uh, controversy during the elections in the US. That's okay. You know what I mean?

Nadiem: Well guys, we've run out of time but incredibly awesome discussion. Love the lively debate. I think we'll be continuing to debate this for the many years to come. But I think that the key insight here is that all of the notions that people assume about channels that are fundamentally better than each other or approaches that are fundamentally better than you, that there is no absolutes in this. There are just a series of it depends on, right? And I think that nuance in both digital, offline marketing, growth, experimentation, financial incentives spend, non financial incentive spend, all of these things have to be assessed and experimented accordingly so that you get to the closest source of truth. Don't believe in Dogma and Paradigms and just run the experiment and see what happens.

Crystal: Yep.

Jasper: And over and over again because the landscape changes your consumer base changes.

Nadiem: Thanks guys for coming. Hope to have you again on the show.

Crystal: Let's run the experiment.

Jasper: Let's do it.

Outro: Hey guys, hope you enjoyed the podcast. If you liked it, please hit like, subscribe and follow us on social media. Thanks so much for tuning in.


See Full Video