Here is interview #2 that I did as part of research for my book on Product Market Fit. This one is with Anshuman Bapna, now founder of ClimateTech startup Terra, and erstwhile Mygola (acquired by MakeMyTrip). I reached out to Anshuman on seeing a tweet by him describing his failed pitch to Joe Gebbia, Airbnb’s founder. Having seen this and his previous tweets, I thought what he had to say would be interesting. And boy, was I right! The conversation with Anshuman was consistently interesting and illuminating, covering such topics as the biggest mistake founders do, the best metric to use at a consumertech startup, as well as his thoughts on product market fit. Here it is for your reading pleasure.
The transcript has been edited to make it more readable. There will of course be errors; please excuse me for them. The interview took place via a Zoom call, on 18th August 2022.
The Mygola journey, and the biggest mistake founders make.
SP: Hi Anshuman. Yeah, thanks for this. So, as discussed, I’m going to kind of just ask you a few questions, feel free to answer them as is or use them to think out loud. Let’s start with your Mygola days. And if you could just take me through one, was there a conscious desire to navigate to PMF? Was PMF, the word, concept something in your mind in those days? Or was it okay, I have got to drive growth, and okay, I’ve got to get customers? Was PMF a conscious decision at all when you were building? Or was it more like something that was not conscious, but something you’re aspiring to?
AB: Yeah, No. That’s a great question. I think, just like how the art and science of Product Management has kind of evolved and become something more of a set thing over the past two and a half, like maybe 20 years, I think, PMF, I am not entirely sure if I had that terminology, when I started out, but I did have something else, which was probably closer to that insight.
We started in 2009. At that time, the problem we wanted to go after and which we did for the next five years was travel planning. And travel planning was that elusive problem that I knew even back then that multiple startups across many years with sometimes a lot of funding had actually failed to crack. So, in that sense, there was this desire to find that bright shining thing, this metric, there’s proof that would prove out to me and to the team and to investors that this time, it’s different. And of course, as entrepreneurs, we are eternal optimists. So, I actually ended up interviewing about a half a dozen failed travel planning entrepreneurs when I started out, even before I started out, to find out why they died. And lo and behold, entrepreneurs also have kind of this, a strong sense of like, trying to leave this reality distortion field, which works on them first of all, before it works on others, right. And I convinced myself that at this time, it’s different.
For what it’s worth, Hacker News had this famous thread on the worst possible startup ideas in the world. And I believe travel planning was number two. If you want to find out about number one, I can tell you that too.
SP: Ha ha….what is number 1? I am very curious.
AB: Personalized news aggregator.
SP: Okay, that’s interesting. Good one.
AB: Yeah, so even though I was kind of deluding myself a little bit, looking at all these interviews, there was also this more intellectually honest desire to find that shining metric, or that shining proof.
SP: Got it. So, the sense I get is that PMF was not apparently conscious, it wasn’t consciously Oh, I’ve got to work towards PMF. Because PMF a topic or concept is a lot more in the air now. But intuitively you are working towards it, and I love the fact that you researched six or so failed founders, etc. So, when you started this, what was the aspiration? What is the goal to move towards? Were there X, Y, or Z metrics that you had in mind for Mygola?
[3:26] AB: Yeah, yeah, I think for us, the theory of change as in the nonprofit world, was pretty straightforward. It was that, look, travel is one of the largest economic categories in the world. But all the money’s actually made at the bottom of the funnel. There are literally two companies which are at the top of that funnel instead, one, and they are the exceptions that prove the rule. These are Google for acquiring these customers, and TripAdvisor for helping them research. And there was literally nothing else after that.
So the thinking was that if it is truly the largest possible, one of the largest ecommerce categories, if you could somehow sit at the top of the funnel in the broadest possible way, like Google, TripAdvisor, and yourself in some way, you would have a way to put your finger on all of this money flow that was happening. So that was thesis number one, second one was fairly sound and continues to be sound even now. I think the thesis number two was that oh, and the way you get there is by helping people in something that takes an enormous amount of time for them, which is travel planning. So, I think this jump, that travel planning was a chore that everyone hated. And not only hated it but also hated it so much that we would actually pay something or someone else for that. I think it was this subtle jump that we made without even realizing and went too far down that path in my opinion.
SP: Ok, this is interesting. Can you double click on that “too far down the path”. This is great.
AB: I think that maybe I’m introducing another framework here. But one of the things that I’ve realized across the four different startups, companies and so on I have been at, is that you can innovate on at most one axis at a time. If, as a startup, if you try to do more than one axis, then you’re, then you wouldn’t have enough runway to be able to crack both of them. So, we were already innovating massively on the product side. We ended up getting a US patent on text mining. We ended up building this algorithm that could automatically read any article or blog and convert it into an itinerary, which was not done before and so on. We did this beautiful interface; we won a lot of awards for that. We did a ton of product innovation, but you’re also trying to do a business model innovation at the same time, which is that somehow all this money that goes into travel will start flowing in a slightly different way.
And that business model search, we didn’t take into account that that was actually as big a problem as the product discovery and search was. And, frankly, after the acquisition (of Mygola by MakeMyTrip, when I came in to Make My Trip, it was plain as daylight to come in and say, Look, there is literally one business model in travel, which is selling tickets. And there are many ways to sell tickets. But if you add anything, any kind of structure that is before selling tickets, and yet makes money only on selling tickets, then by definition, your conversion percentage will be some percentage of what happens in ecommerce (selling tickets). If Make My Trip has a 5% conversion, you will have a 5% multiplied by some other leaky funnel. And no matter how much you optimize that funnel, you’re only going to get to essentially Make My Trip’s economics. So that realization that we were innovating on the business model at the same time was the one that I felt that we jumped into that too far too long.
SP: Okay, a lot of it is intuitive to you but I just wanted to understand this a little bit. So the basic, the product set is very clear. You read a Conde Nast travel article on touring through the Maldives or Bali, and you just parse it through the Mygola algorithm, and it creates an activity for you two days at this place, go to this temple, this beach, etc, etc. Now, I just want to understand the business model, please help me. So you were selling tickets, you were making it all happen. You were like a travel agent, you would book everything and give it to them; was that the revenue model?
AB: Correct. So, the itinerary would have hotels and activities and in some cases, flights and transport also, they were all bookable. But what was really happening was, and if we look at the business model, the business model is now saying, Look, remember, Make My Trip’s three tabs or Booking.com’s three tabs on flights, hotels and activities. Now imagine a beautiful interface that takes you days, weeks, hours to get to that interface. And any ecommerce Product Manager worth their salt will say, “Dude, you’re gonna lose like, I don’t know, an enormous number of people, even before they click the activities or the hotels tab, even if they consider to start purchasing.”
SP: Got it. Fair enough. Interesting. Would you say that you overdid the product side? Did you feel that in hindsight, if you were to go back now with all the wisdom that you have, how would you kind of redo it? What would be the change?
AB: I think the one thing that I have realised, and I’ve seen this now as an angel investor also is that entrepreneurs sometimes make the mistake of solving not the biggest risk, but the risk that is most solvable by them first, right? So, in my case, the biggest risk was actually the business model, which is how will you make enough money to actually make this all work? But what I really knew was how to build a great product and make it highly, highly scalable. In fact, I think the Airbnb thread that you saw on Twitter that we first connected on, it was literally like four years into this journey when I finally realized how far I have strung myself along when in that Airbnb conversation, Joe Gebbia basically said “Dude, we don’t even care about scalability. You’re too scaled for us.”
You are this tiny startup of 20-25 people, and we are a $3 billion company at that time, we already scaled, because that’s not the biggest risk in this problem. Somebody will solve scale, you will solve scale. And yet we are building all these great things that could automatically scale to everything on the planet. That’s what I would do differently. In a way I kind of almost admonish myself, and to angel investors, entrepreneurs that I meet, I say, “Look, I know you’re a great product guy, you’re a great business guy, you’re a great this guy, whatever your background is, can you run your business for six months on just an Excel sheet and a phone call? I don’t care what the product is. I don’t care what you’re doing. But can you just do that, and force yourself to not build anything because building is easy, and gives you a false sense of progress?”
SP: Got it. In hindsight, you would just check in with the market earlier, like instead of building something to 75- 80, let’s say 100 is max, and then checking with the market, you’ll probably build it at 10-15, and then take it to market. There’s no code, or make it low code and take it to the market. That’s what you will do?
AB: Yes, No code or low code and take it to the market.
SP: I’m curious, I know from Mygola you went to MMT (Make My Trip). And then I think Terra.do, but before that there was something in between?
AB: Make My Trip merged with Goibibo. So I was the CPO, and became an CPO of Goibibo, to kind of integrate the companies.
Anshuman’s definition of PMF as stacked S-Curves.
SP: Got it. So given you do angel investing etc. Do you now have a definition of PMF? If I was a young founder coming to you and say, “Anshuman, Hey, what is this PMF thing?” And you had the patience to explain it to me, how would you kind of give me that definition? I mean, don’t have to put it in too much English. But even Hindi is fine, whatever you’re comfortable with.
AB: The qualitative part of it almost sounds like the way you describe love, right? Which is, I can’t tell you what it means to be in love. But when you’re in love, you’ll know it. So that’s kind of the completely over a beer kind of a version of PMF. But honestly, I think there’s again, I’m obviously talking from a very consumer internet kind of standpoint, maybe there are parts of it, which are relevant to SaaS, but maybe less so to enterprise, and so on and on, but to me, it feels no matter which way you arrive at it, and whatever metrics you use, ultimately, the true measure of product market fit is growth. Right? When you see stuff growing for example, or it might be that customers are arriving at your door without you doing any marketing spend. Or it might be that customers are actually coming in and doing NPS and the NPS is like crazy, good NPS, and then crazy good NPS is leading to referrals. So, stuff seems to happen without you having to do much. And that starts to translate into growth is to me is one aspect of PMF.
Now the difference is, and this is something which is, now for example, I’m talking about Terra’s experience right now, which is that for the first time in my life, we’ve actually built a product that has an NPS of 88. Right, that’s awesome. Like, when we run all our programs, our referral rates are crazy. About 50 to 60% of all people coming to our programs come from referrals, and the kind of keyword or words that they use to describe the experience at Terra is outstanding. Now, the challenge with that, is that why that is still not product market fit is because all that we have proven is that for a very small micro segment of the market, we are golden. But PMF, the M part is obviously as important. And what we have to prove now is that this market is large enough. So in that, and another way to look at is these the classic S-Curves, right? PMF is not one S-Curve, it’s S-Curves stacked on top of each other. And to me, therefore it feels like maybe what you’ve done is that you’ve done one of those S-Curves. But you need to find our next S-Curve, or we don’t have a company, or we have a company which is not interesting enough, large enough, impactful enough and so on.
SP: Very nice. I might use this! “PMFs are S-Curves stacked on top of each other”. So, I think one of the frameworks that I’m coming to and it’s not entirely kind of created by me, but I have seen some versions of this out there; so my framework is that, PMF is actually two parts – one is PPF or product to problem fit. And the second is MMF or what I call motion to market fit, which is there’s a go to market motion for the market, and you need to crack both stages. So different people have different terminologies. The founder of Regrowth, Brian Balfour calls it product channel fit. And there are different fits; different people have come up with many kinds of terminologies and all that. But it seems to me that you’ve sort of, what you’ve solved is really the problem solution fit or product to problem fit. So there is intense customer love but it is a niche. And now you’re saying is that niche a market? Can you expand it to a market? And so, that’s great. So you’re saying S-curves stacked on top of each other?
AB: Product channel fit is exactly how I would talk about as well. I have increasingly gotten to the point where, to me, it feels like, look, there’s always a very large market out there. It might be that a large market might exist in a parallel universe. But as entrepreneurs, you could actually make that parallel universe happen. If the right kind of things come together, and the thing that really is getting critical, critical in that, is the product channel fit.
Given the power of the internet, with 4 billion people on the internet, there is a large enough segment of almost anything that you want to do. NetFlix had proven that for us, Reddit has proven that for us, and so on. I would think that almost as an axiom, there are a few millions of everything out there. But the channel discovery is the difference between life and death for a company. So at Terra, I think we have a product market fit where the market is kind of a tiny market. But there is a much larger market out there waiting. But we don’t, we haven’t yet discovered a channel to access that. And it will not be one channel. Yeah, but will largely be one channel that we will get right first, and then that channel will start exhausting. And then we’ll add another channel on top of that, hopefully the channel that you are picking is actually an internet scale channels like Google AdWords, for example, or SEO or things like that. But it’s fine, it doesn’t matter initially (that it be internet scale initally).
There was a time when for example, for Mygola, Twitter was a fantastic channel. It just turned out that in those early days in Twitter back in 2012-13, we acquired half our customers through Twitter. What would happen was that…Twitter had this dynamic where two things are happening. One is that people were still early in the days of trying to figure out how to use Twitter, and one of them was, well, hey, we’ll ask a question and see what happens. And second, was that unlike other channels, where people had already been completely jaded out from businesses reaching out or responding back to them, Twitter hadn’t gotten there yet. So people would post and say, Hey, I’m headed to Berlin with two kids, what do you think I should do? And turns out, not enough people have followers who have a big intersection of Berlin and kids and the second, the Twitter stream itself is ethereal, it comes and goes, and so on. And so the vast majority of people will actually find, essentially, like the abyss staring back at them or their activity, right, on Twitter. And if a business actually reached out with the right intent, then they would actually find it useful. And we realized that we could automate the hell out of that process. You could literally write a simple NLP engine on top of all Twitter streams and identify things like Berlin and kids and intend to travel next week. And automatically tweet back and say, Hey, what kind of stuff do you like when you go to a new city, for example, and then people will respond back, and so on. So that became a great channel for us. But again, the channel worked well, because it was the early days of Twitter. And then Twitter essentially converted that into something that looked like that same product.. the Twitter ads product had a feature that did exactly this going forward. And lo and behold, everyone jumped in, and therefore CACs (customer acquisition costs) went up and so on and on. Just like every channel does. Actually, the Brian Balfour essay was one I was going to refer to you also. It’s a lovely essay.
Anshuman’s views on GTM.
SP: It’s a good one. Yeah, this is great. It brings me to a question that I had wanted to ask further down, but I’ll bring up here on the motion or channel. You mentioned that it may not be one channel, but it may be multiple channels. So, typically, one of the things I’m coming across is that there is typically one hero channel for successful products. And one I’m also trying to kind of see if there’s a framework for identifying a channel, and I wanted to get your thoughts on it. So, if it’s a high intent product, then I think SEO or even performance marketing could theoretically work. Because it’s very high intent I’m very keen to ask. For example, I’m going to Berlin with two young kids. It’s a very high intent question, right? But sometimes it may be, you know it is not a high intent context, it maybe a niggling pain (not intense) or something like that, and you may subconsciously have wished for a product, which kind of eased the pain. But you’re not going to go and directly search for it. Maybe it’s a different channel that works for them. Perhaps content could work very well there. Maybe, I don’t know. So, is there a framework that you have, like for Terra, how did you think of the channel? What are the channels you’re thinking of? Maybe you could just take me through this thinking? Similarly, if Mygola, you explained about Twitter, but after Twitter, what happened? So, two questions around both early stage zero to one channel hunt.
AB: So I think there is a pseudo framework I have implicit in my head, and let me see if I can make that explicit. For better or for worse in my case, I think, my experience is connected to consumer internet high consideration purchases, right? So, travel is like that. And so is working in climate, right? Learning in climate? You’re essentially already kind of primed to do a pretty deep commitment and so on. In that the sense where I have gotten to is that there is always a channel one and channel two stage. So, channel one is to find an existing user behavior that you can latch on to. Right, and it might not be very large. And, and so it might be, for example, people, I mean, this case, in MyGolas’s case, it was we realized that about 100,000 tweets are being sent out every day back in the day in 2012, which are about a travel intent question saying that I’m doing travel, help me. So, there’s already this undiscovered, but already large behavior happening there. The similar behavior that we found, for example, in climate is that people tend to join these micro communities. “I want to do something climate, we have no idea what to do with climate, let’s create a WhatsApp group. Let’s create a Slack group.” And these micro communities are actually out there in the literally hundreds, each of them with like a few 100 people, sometimes a few 1000, at best, but the sigma of them is pretty large. So that’s an example. Both of those are examples of channels.
In Terra’s case, the channel was how can we embed ourselves inside communities in a way that is a win win for the members and owners and for us. So that became part one of the channel hunt. Now it turns out that there is a specific kind of such community or micro communities, which are potentially significantly larger, if really acted upon and deepened as channels. These are employee groups inside large companies who are all interested in climate. So, these so-called interest groups, typically, like Google will have for example, a Google Anthropocene, which has 1000 people, who are highly interested in climate. Amazon has something of similar scale, Salesforce has probably something 10 times larger than that, and so on. And it turns out that when you are embedding ourselves in these public communities, we were seeing a small reflection of each of those employer groups. So, 10 people from Salesforce would be inside some other publicly visible community. And turns out that this is potentially a second channel, which is significantly larger, and has the characteristics where you can actually go to a Unilever, they may or may not have such a group already. But you could insinuate yourself pretty easily, get yourself started in that channel.
I’m giving a specific example to essentially say that it’s almost like when you look at your M1 M2 retention, cohorts, etc., you can potentially look at all kinds of data for all kinds of cohorts of users. But one set of users that is kind of standard product management shibboleth is to look at the happy cohort, right? Especially in the early days when there are many reasons for people to be unhappy with you. But the people who continue to be happy despite the crappiness of the early experience, the power users or whatever you call them are the ones you actually want to double down on. To me it feels like channels have a similar characteristic, which is, inside your small channels are actually embedded big channels. Now, the media that you use to access them, for example, it might turn out that to access those 5000 Salesforce employees, maybe I should still use Google AdWords. Likely not. Or you actually become a sponsor at Dreamforce; this is kind of a channel or Google Zeitgeist is your channel or Amazon, whatever the conferences are, it is the actual channel or the media, not the channel.
I don’t know if this is helping in any way. But to me, it feels like the framework is step one – try to see if there is an existing latent consumer behavior for even a small set of consumers to begin with, insinuate yourself into that. And in there, seek out the happy users and see if – and make a distinction between channel and media – see if happy users are actually a large channel for you, just that the media might be different. Media is much more easier to experiment on how to reach this particular channel. Because you can do like five different experiments, say, Okay, let’s use our contact list, let’s use our investor base to reach out to Salesforce’s head of sustainability. But let’s also run a Google ad. And let’s also make a free product just like Slack had a free product for enterprise, right? Any Salesforce employee can use their salesforce.com email address and sign up and get 10 more of their colleagues to start using Terra’s product and so on, that is relatively easy to experiment with. But knowing that there is this universe of users sitting inside this channel called enterprises interested in sustainability is the example there.
SP: Got it. So that is channel one. Channel one is wherever there is a natural existing slant towards a certain activity or a predilection towards a behavior that’s already happening. And you basically go to that watering hole. So, it’s sort of a watering hole strategy. Second one, where you sponsor a Dreamforce etc., Is that part of the same strategy?
AB: Yeah, think of it this way, which is that the second channel is enterprises already with net zero commitments, right? And their employees, right, are the user base and companies with the net zero commitments are the path to reach them.
SP: Okay. Got it. So net…. I’m not that clued in on climate tech, to be very honest. So net zero means something.
AB: Net zero means a company that has actually raised their hands and publicly made commitments to all their shareholders and to their universe of users, consumers, that they are going to reduce their net carbon footprint to zero.
SP: So, these are very evolved companies like a Patagonia types, or close to that.
AB: No! That’s the amazing thing in climate, which is now for a variety of different reasons, 30% of all, S&P 500 companies have net zero commitments.
SP: Okay, wow! so this is this, like the next stage. And where you can go directly to the companies?
AB: Correct. Maybe the distinction that I’m making, maybe it’s an artificial distinction, I don’t know, is that, imagine that this company X, which has made net zero goals, that company X has 20 people at these watering holes showing up consistently all the time, and that company X has 1000s of employees. These 20 give you an insight that there are actually potentially 1000s inside there. Now, you still don’t know the channel right now to reach those 1000s in the company. Right? The channel might be to, to talk to the chief sustainability officer. Or the channel might be to actually set up a free version of the product just like Slack sneaks into enterprises, right? Maybe you sneak into the enterprise. And you use these 20 to somehow take the free product and actually get another 100 people using it by the time you actually have a conversation with somebody inside, and on and on.
SP: Okay, so now it becomes clear. So when say channel one there are 20-30 people and channel two is much larger, but two ways to approach one is like the highly committed watering hole for channel one, and then the larger channel two via a separate channel, got it. So for any high intent product, this could work, not necessarily for a low intent product, where you say I don’t know what I want. There, this may necessarily not be the best approach. So, there you may have to look at a different kind of approach.
AB: Yeah, I didn’t mean high intent as much as high consideration, which is that your intent might be low, high medium, I don’t know right now. But if you ever go down this path, you will not make a decision that helps my bottom line immediately. You will take step one, step two, step three, and eventually somehow you will blink. Travel is like that, right? Which is that even if everyone knew about Mygola and you were thinking whether I should go to Turkey or whether I should go to Dehradun for my summer vacation. Completely different intents and therefore by definition Turkey is pretty low intent, right? You will though check it out. It is a high consideration thing. You will actually spend X number of hours before you spend dollars on this.
SP: So, this is interesting, this distinction. So just from a Terra.do perspective, what is the next step? Like how do you kind of figure out the market now? I mean, just think a little loud. So you discovered a niche. How do you now make it a large enough market?
AB: Yeah, I think firstly, macro picture, which is this whole belief that the climate economy is going to be bigger than the internet economy. That’s how big a deal this is. Right? And for a variety of reasons, but especially sitting in America, which just passed yesterday, this $400 billion act, which is unheard of in American history. And $400 billion of American government money will catalyse another 10 times that much money coming in through all kinds of different sources and so on. So we’re off to the races as far as I’m concerned in America. And it feels very much like I started my first company back in 1999. It feels very much like that in Climate right now. So, one way or the other, to me, it’s a given that we’ll have, instead of the 2 million people who work in climate right now, we’ll have two hundred million people working in climate in 10 – 15 years. It is both because the crisis is so big, but also because the opportunity is so big. That’s part one. What follows from that, for me, is that therefore, anyone who builds a piece that allows people to make this transition from outside climate into climate easier, will have a large business and large impact. Now, what is that piece? I’m completely agnostic about it. I was very agnostic about when we started, I continue to be agnostic about right now. What we did was try to capture it in the name and it said, Look, we’re not Terra dot learn; learn is not an end goal. It’s a means to an end. And the end goal is do, which is people have to be deployed in climate. They could be starting companies, they could be working on projects, they could be volunteering, they could be doing this or that, etc. but they have to be working on climate, and hence Terra.do.
So what we realised was that just like in the early days of the Internet or the early days of crypto there are the believers. And so, you have to first build for the believers. So we were very clear we are not in the game and will probably never be in the game of trying to convince people to work in climate. We are going to be working, building something for the believers who say look I am going to, I don’t care, come hell or high water, I am going to work in climate. And I just have this 1-2-3 kind of problem. So, when we listen to them very carefully, we realise that the number one problem for them was I don’t know where to begin. You don’t need to convince me that climate is important. You don’t need to convince me that I should be working on it. I just don’t know where to start. And that sounded like a learning thing. So first you see we will have to solve learning. So we built learning, but because this is also, like imagine the internet in 2001 or maybe even earlier; turns out when people graduate from our programmes, it’s not like we’re running a data science programme where there are 1000 companies at the other end just waiting to catch these data science graduates. We have to figure out the discovery and employment problems and so on. Also, because nobody had solved it and so that’s how we got into the jobs space. And the thing that is different between climate and say fintech is that in climate you’re doing it for something which is slightly more than oh my god there is much money to be made here. So, there is a very strong sense of community, which is a very fuzzy and kind of crazy word to use. But to me it feels like on the consumer internet, on the internet, the hardest thing to build is community. You can build self-driving cars and whatnot; you can’t build community by just throwing money at the problem. And there are literally very few examples of truly scaled community ever on the internet. Back in the days the Open Source movement, maybe Crypto right now, maybe Reddit is one example which proves how hard it is, as opposed to proving how easy it is, and so on. So, learning, community, and jobs these three things that is what you really need to put together. In what order I don’t know. It seems like this is the order
If you look at what, who else has combined learning, community, and jobs, the most obvious example scaled example is LinkedIn, which worries me, right? Linkedin is like electricity, it is everything, depending on which lens you look at it. That’s how platformised it is. So, back to our story. So, our story is that we think 100 million people are going to move into climate so therefore by definition our product has to be really broad at the top. Day one, PMF1 was fine. It was all about, let us take the believers and really solve their problem beautifully, which got us to 88 NPS. And we have people like Rajan Anandan who are also students in our programme and rave about it. But so is that waiter in New York, who just committed his life to climate, and investment bankers and ex-firemen and so on and all that who are part of our cohorts. We built for the believers and we’ve gotten a great NPS from that. But we are now breaking up everything that we do, all our learning content behind these 1000s of dollars, and making it 15-minute chunks and making it freely available to anyone who wants to work on the climate of the planet. Because we know that if we truly believe that it is going to be 100 million people then we have to be broadest at the top, not a narrow straw which is give me $1,000, give me 10 hours across 10 weeks of your life and only then I will let you experience Terra. I think in that sense it’s a pivot for us. It’s truly like a PMF2 search, which is a bit of a bet the company kind of a search.
SP: Got it. So, you’re fundamentally just looking at what succeeded, learning, which is expensive. You’re saying broaden the top of the funnel to attract more people into it. Learning may not be…you may not monetize learning as much as you may monetize something down the path , something like jobs. Or something like that, is that sort of the idea?
AB: Sajith, the thinking is pretty simple. As I shared recently, innovate only one axis at a time, right? I think the fact that we are in climate so early, is the only innovation that I want to take on. Not any other innovation, and not definitely on the business model side. Now on the business model side, the way we make money are the oldest ways of making money on the internet, which is one is we charge people for these courses. And it turns out that there’s a lot of appetite on the employer side to pay for their employees going through this learning, skilling etc. Very straight forward. And on the other side, in jobs, which is companies coming and hiring from talent that want to work in climate. All climate companies hiring will pay us a commission.
Just to give you a sense of the value, PwC announced that they’re going to hire 100,000 ESG consultants. This is a big company but not that big. And if you think of PwC is almost like a canary in the coal mine. They are actually seeing how many dollars are gonna flow, billions and billions of dollars, every leader in every sector will say, Yaar (Bro) I also want to be sustainable, and go net zero. How do I do it? PwC says don’t worry, I’ll tell you how. You see for PWC, McKinsey, Deloitte etc., sustainability is the new digital. In the next 20 years, the next 50% of their partners are going to come from sustainability transformation, not digital transformation. That’s why they are retooling. In that sense the need for upskilling massive sectors and the need for jobs in massive sectors is what we are betting on.
SP: Got it, you said how many of the Fortune 500 or S&P 500 are net zero, 1/3?
AB: 30% of them are committed to Net Zero.
The most important metric, and why he never overweights the V1 version of every product.
SP: Well, this is interesting. I kind of joke that I am a Web2 VC when I talk to crypto guys, and I’m currently a smokestack VC for you, but this is something for me to read upon. This has been extremely useful in enlightening me on the sustainability part. Okay, and we could move to the last section. You are a PM as well. So metrics. Okay, so when young founders come to you and say, hey, what are the metrics that we should focus on, what do you say? Are there 2-3-4 metrics that you encourage them to focus on? Is there generic thinking about frameworks about these metrics that would help? You could talk about Terra as well? What are the metrics you focus on? Clearly customer love is one of them.
AB: Yeah, sure. I think the metric that I value as a PM over all other metrics are metrics that are good leading metrics as opposed to good lagging metrics. Because the velocity is such that you can’t really wait for everything to kind of pan out; like LTV to CAC. It is great for your business team to eventually get there in six months time and so on. But it doesn’t change your… it doesn’t help you make a decision today or tomorrow, right? So therefore, I focus a lot more on leading indicators and there, for example, for a consumer internet startup, something like a DAU to WAU. Daily active users, weekly active users is something which I care a lot about. Then I look at the experience difference data in the cohort between the first session and second session. There is onboarding and then what happens after onboarding, and then what is the delta between the first session close and the second session close, which is kind of the frequency of the product in people’s mind. To me those…
SP: Just double click into this.
AB: To me, any product that might be high consideration or not, has to exhibit very fast follow-on behaviour, for it to become a habit in the user’s mind and habit even on a weekly scale; I am not talking about lifelong habit. Which means that your second session has to happen quickly enough. And you can tweak that up and down depending on what product you’re selling. But the second session has to happen quickly enough. And that is another example of a not super precise, but a leading indicator of whether you’re doing well or not. And what I do then is I then, just to kind of gut check my intuition, I do simple correlation of these kinds of metrics, these fuzzier leading metrics with actual lagging, high quality metrics that I care about three months later. Like the M3 retention example, or CAC to LTV. So, I do a correlation analysis and say, my intuition was these two matter or these three metrics matter, so which of these actually is a higher predictor of the ones that truly are the final metrics. Or it turns out my intuition on this one was not as right or maybe I was looking at this factor slightly differently. So, I’ll go back and tweak it.
SP: Got it. This is useful. Okay, nearing the last few minutes. Any kind of caveats, thumb rules like what you said about, don’t innovate on two axes. Sort of like a warning sign; are there favourite warning signs you tell founders like yeh kabhi mat karna ya yeh dekha karo (Don’t do this ever, or look at this) or something like that, like grandma’s wisdom, like what’s the grandma’s wisdom equivalent?
AB: Yeah, I think yeh toh woh wala tha ki (that one where) you want to gut check that you are working on the riskiest part of your assumptions. Or are you just working on the easiest part? The third is which is more of a product kind of a thing, intuition, which is that I actually don’t get anyone of the V1 feature. I only care about the V2.
SP: Okay. Interesting.
AB: Because V1 is rife with all kinds of false, both false assumptions, but also false beliefs and false confidence and so on. And there is this whole massive deflation that happens after V1 is shipped. Everything was just hard. It seems like you’ve already done 90% of what needs to be done. When in reality, theoretically, we all know it actually has just begun. And typical startup and organisation mojo are such that they prioritise V1 better than they do V2. So V2 is where, and V2 is kind of a placeholder for V2 and V3 and V4 or whatever is actually in my opinion what we should focus on. The second one, if you look at most of these product market fit stories also, I mean, they are like lightning in a bottle kind of PMF. It almost feels like Boss isme kaun sa kismat hai kya hai, bahut badhiya (What amazing luck!). I have no idea what to do with that (PMF from V1). How can we use that as advice for anything? The vast majority of us are actually getting PMF from V2s. So all the intuition, all the organisational structure, even incentives in the company, aap kya reward karte ho aur aap kya punish karte ho woh sab sach mein V2 mein dikhta hai (what you reward and punish is actually seen in the V2). To me, the focus of the founders should be to use their good offices to get really high quality V2s out, V1 jo karna hai hey kar ley. (Ignore V1s).
How he learns, and his advice for how and when to quit.
SP: So, just a few things that are resources that you recommend. Like you clearly seem to be a reader, like, PMs also read a lot more. So are there resources you recommend, could be podcasts, could be books, could be articles, could be people, could be YouTube videos, for the zero to one phase, not necessarily on PMF, though fundamentally everything’s about PMF in that sense, but any zero to one resources you recommend?
AB: So, I have gotten a little rustier on this now.
SP: You can also send it.
AB: Yeah, it kind of used to be my bread and butter a few years back. But I have been off for a while and it is 2022, and you might say, what is he recommending? I had more of a process of learning, so what I got really good at is to figure out this is how I learn. I realised that I have two modes of learning that really work well for me, which I actually look for a lot in product centric founders a lot. One is this exam time, last minute mugga wala jo mode hota hai (crammer mode). In six hours, you have to decide whether to invest in this prop tech company or not. And you have no idea what the proptech world looks like, what to even look at, what the dimensions are and so on. So I realised that is one mode of learning which really works well for me. And the way that works really well is to actually just be power users of Google. And one of my hacks is to start at images.google.com not google.com. Because it turns out that when people are trying to be really really pithy about their deepest insights, they will try to capture that in an image or a graph or a framework or something like that. So I start there. There will be five different images and you can over a period of time, get a sense of the sector. So that’s one way of learning. The second is this daily habit of essentially having a curated Twitter feed. To me it feels like Twitter has gotten to the point jahan pe (where) people care about building an audience, have essentially honed it to a craft and therefore my feed and some simple discipline around consuming that feed on a daily basis will get me enough to keep my reps happening on a daily basis. On the product side, I definitely look at the Lenny Rachitsky’s of the world and so on. But I also look at for example, people who are philosophising on how behavioural psychology works, for example, all the people who are in that whole world of okay, what kind of false assumptions do people make all the time… what is the guy?
SP: You mean Kahneman
AB: Yes, that universe of people a little bit.
SP: Yeah, interesting. Any other?
AB: I am going maybe a little esoteric here. For example, I love how architects think and the urban planners, because in a way, they have similar problems like product managers right? There is this ideal product and then there is the product as these fools called mortals use. And there is the whole world of architecture, full of architects and designers, who design perfect cities, which are unfortunately sullied by real humans living in it. So I follow urban planners and architects to get some intuition around this.
SP: Oh, that’s interesting. Is there a question that you expected me to ask but I didn’t?
AB: I think there’s just a question around pivots, which is, when do you know that you need to pivot? And maybe I’ve done so many of these and there is a little bit of history around that. I think first of all, I think the hardest part about that is whether product market fit, not whether you have product market fit or not, but whether product market fit is just around the corner or not. And our standard psychological biases are designed to kind of send us that in one direction or the other. If it is the optimistic one, it could be what could be the next feature, that next market segment, this one that will actually get us there. Problem is only so many cycles can you do that.
I think maybe the only way to solve it, in my opinion, is to decide how much mental runway you have, not just resource runway. And obviously do that before you even ship out any version of the product. Which is to say look I’m going to take a problem statement, I’m gonna get Indians who start staying in somebody else’s apartment. America mein hota hai, India mein nahi hai. and for whatever reason Jain food and ladka ladki etc. (This behaviour can happen in US but not India for cultural reasons). You are a PM at a travel company or you’re a founder thinking that this problem can be solved, and it’s a valuable problem to solve. Timebox it; say I am going to get three shots at the goal and I’m going to get a year to potentially get to this. Having on day one, the humility that you will have three shots allows the pivoting to be significantly easier. Because to me it feels like the balance is weighed so heavily in favour of continuing with status quo. Yaar kar diya hai, (Bro I have done it), I have convinced all my employees and my investors that it was the right thing to do. And thoda thoda toh (A little bit) if you squint really hard, you can hardly see the data, kuch toh hone wala hai (something is gonna happen). But if you know that you have promised I’ll take three shots of the goal. It just makes it take the pressure off you and you can say okay, is it the time to take the second shot? Because I have only six more months left. Yeah, it is okay, fine. There’s no drama or there’s less drama and significantly less drama.
SP: This is interesting. Yeah. Did you actually go into Terra with this thought in mind?
AB: Yeah, Terra is very much like this and honestly I ended up doing that in Makemytrip also. I launched a couple of businesses there, wahan bhi yahi tha ki (There too I took) three shots in the goal. I only have so much time.
SP: Thank you so much for this. Really appreciate it, Anshuman!