Ranjeet is one of the more interesting figures in Indian startup twitter. I have, for long, enjoyed his tweets and thoughts on the challenges and learnings from building a content platform in India, Pratilipi; one that paved the way for a wave of vernacular language apps to inspire and follow and expand the market. Hence, I was excited to get him aboard for a PMF Convo. And he didn’t disappoint. In our PMF Convo, Ranjeet talks about why he doesn’t like the concept of PMF (too vague), and why a better framework is to see the early startup journey as one of verifying each of the hypotheses that need to hold true for the vision to be realised. He then goes on to explain the Pratilipi journey in his framework of hypotheses verification, such as how for Pratilipi the first hypothesis was to prove that a UGC content platform could be built, then the second was to validate if they saw network effects, then the third hypothesis was if they could translate the success of this format to other languages and geographies. 

In the convo, Ranjeet explains how network effects businesses see engagement metrics such as time spent increase with scale, unlike say a pure media business. Along, he shares his thoughts on the importance of liquidity (or supply of inventory) as a way to measure the health of a marketplace with network effects, and how to use an atomic network (or basic building block) to see how you can achieve sufficient liquidity in this atomic network, and then replicate what you did across other markets. Another framework he shares is how to address the biggest risk for a business first. Finally he touches upon to understand which metrics matter, and why frequency of usage is almost always the most important metric for consumer app companies. I enioyed chatting with Ranjeet and learning from him, and so will you.

(Please find below, the transcript of my conversation with Ranjeet Singh. The transcript has been lightly edited for readability. There will of course be errors; please excuse me for them. The interview was conducted via a zoom call on 11th June 2023.)


Sajith: Ranjeet, I wanted to know a bit of the journey that led to Pratilipi, and then we’ll talk about PMF (product-market fit). But along the journey, do talk a little bit about whether you consciously worked towards PMF, that is, were there any overt conversations about PMF? Later you can give your definition of PMF. But broadly, the Pratilipi journey so far. Before Pratilipi, I know you worked two years in Vodafone, then did a startup programme and then the last eight years is the Pratilipi journey.

Origin story

Ranjeet: So, I was born in a very small village, of less than a thousand people, in Rae Bareli, UP. Almost nobody older than me knew English in my village. Typically if you made ₹7,000 – 8,000 – 10,000 a month, you were decidedly rich by any definition. That kind of a village. I was the first engineer in maybe 50 – 100 villages. I was also a voracious reader, I used to read 140+ books a year, roughly, give or take. I started by reading Hindi because that’s what everybody else read. Went to pursue my engineering, realised that Hindi content was simply not available as much, so shifted to English, but I also started telling my friends that it should be my choice if I want to read Hindi or English or something else, it shouldn’t be because of lack of access. However, I didn’t really do anything about it at the time.

After that I went and did my MBA and worked at Vodafone for a couple of years. That seemed a little too easy and I don’t like easy things so I quit Vodafone. The plan was ideally to find a startup that I wanted to join. I’d done a little bit of computer science, a little bit of retail, a little bit of banking. I interned in Citi for 3 months, a little bit of strategy, a little bit of sales, a bunch of other things. So, I was like okay, I can learn and do whatever. I just wanted to work with the best boss that I could find. So, I roamed around the country for about four months, met a lot of people, liked some, but did not like anybody enough to say, okay, this is who I want to work for. So, around this time some of my friends called me and said you have been cribbing about this for such a long time, so if nobody else is solving this, why don’t you give it a shot? And that’s how Pratilipi started.

On PMF; I don’t like the word PMF because it’s too vague….basically you like a company who say they have achieved PMF, you don’t like a company who say they haven’t achieved PMF and then to justify that you’ll come up with 5,000 other explanations like PMF is a journey and that you can find PMF and you can lose PMF and all kinds of bullshit which is obviously true but all of that also basically makes it not actionable, which basically makes it useless.

Working backwards

Ranjeet: Rather, the way I advise people and the way I think about Pratilipi is to come up with a set of hypotheses. Start with the vision. If you are right, what does the world look like in 10 years? What changes? If you are right and you are building X or Y or Z, then what does the end result look like? So, start with that and then work backwards from there to find out what needs to be true for your vision or your thesis to be true? So, come up with that set of hypotheses that these are the five things that have to be true for my vision to actually be true. And then basically you solve one of those, two of those, three of those and the more of those that you keep on solving, the more the business thesis starts making sense.

So, as an example for Pratilipi itself, when we started building Pratilipi, the overall vision that we thought of was to build an ecosystem where stories could start in one format, one language, and one geography and then travel across all other formats, languages and geographies. In other words, anybody should be able to share a story with the rest of the world in any language, any geography, any format they want. And then we should be able to pick the best stories by using data models and we should be able to leverage these stories into all of the other formats, all of the other languages, all of the other geographies. 

Obviously, you can’t start by doing all of this. So, we broke it down into, in fact, our earlier deck would probably have this slide as well. That step one would basically be to prove that we can do the first part of the story, which is basically that anybody can come and share a story and anybody can come and read that and that itself works. In fact, give me a minute, let me try and find if I can find that old deck.

Sajith: Sure, that’ll be great.

Ranjeet: So, it might be very useful. (searches for it)

Sajith: Yeah.

Ranjeet: Unfortunately, I can’t seem to find it.

Sajith: No worries. Tough to find these things in the moment.

Ranjeet: Actually, I still found something which would kind of give a good sense of it.

Sajith: That would be great.

Ranjeet: (shares the screen and shows the deck) So, this is a draft vision document which was prepared maybe two-three years back. So, for the first part of the story, if we are right that we can build a democratised storytelling platform, then how does that work? Why would that make sense? So, this is where that can only work if it’s a true UGC platform. Why UGC is important is because otherwise basically the winner is decided by who has the most money.

In marketplaces where you don’t have network effects, it’s about who can out-execute, either that or who has more money, who can outspend. Versus in businesses with network effects; it basically means once you get there (where you have strong network effects), it’s very, very difficult…you’ll have to actively fuck up and it’s still hard for you to die. Look at LinkedIn for example, looking at Craigslist for example. They have kept on making mistakes after mistakes after mistakes. And both of these still continue to work because the network effect is so strong.

Pratilipi and Network Effects

Ranjeet: So, our hypothesis becomes that Pratilipi can become a business which has very strong network effects. And why that would work or how that would work is that we’ll basically get, let’s say a few writers to come in and write on Pratilipi. Once we have a few writers who are writing on Pratilipi, they are going to share their stories with the rest of the world because, well, it’s their story. Some of these readers who come to read, they will become engaged readers, they’ll start reading other people as well. Once that happens, then some of these readers will share their links with other people and they’ll come and start writing. So, this flywheel is why Pratilipi would start making sense and result in network effects over time. 

The definition of network effects basically is that once you have a new user who joins the network, then the existing users get more value from that network. If the existing users are actually getting more value from the network, that should be reflected in your engagement. That should be reflected in your retention, that should be reflected in your frequency or if you’re monetizing, that should be reflected in your revenue per user.

So, this is the first part of the thesis for us. This is all that we focused on at the time, till the time that we did not see that all parts of this hypothesis were true and that every time we acquired new users, our time spent per user would go up, our retention will improve, our frequency will improve. So, we knew for sure that we had strong network effects. Nothing else really mattered at all. Only after we got to this point is when we started expanding into other formats and other languages and now other geographies. So, this in my opinion, is a much better way to look at business versus trying to look at a touchy feely thing, which is somehow magically true for all businesses.

Sajith: But this point, it was only Hindi, you said you’ll only solve for Hindi first?

Ranjeet: No, no, we started with Hindi and Gujarati, then we launched very quickly eight languages. So, launching a language for an internet product, it doesn’t really mean much, right? It’s basically just a language and 15 years back, of course it had its own technical complexities. When we started, language was just a language, it was not really technically very challenging. Rendering was still a bit of a challenge at the time, but not a real difficult challenge.

Sajith: Got it. So you said what’s step two? Step one was anyone can share a story and anyone can read it.

Ranjeet: So, step one was basically seeding the platform, which meant getting writers to come and write and then people could read that. Then of course after that, the product keeps on getting better and better and better. But step two was just waiting till the time that you have network effects. So, that you know whether the thesis works or doesn’t work. So, once the thesis works, then you go to the second part of the thesis, which is basically the future thesis that says that the stories travel to other formats, other languages & other geographies.

Going multilanguage and multiformat; and measuring Network Effects

Ranjeet: Once the first set of hypotheses got proven; the second hypothesis was if a story works in one particular language, one particular format or one particular geography, do the odds of that story working in other languages, other formats, or other geographies also go higher?

So, that was the second thing that we focused on. We launched multiple different products starting with two, Pratilipi comics and Pratilipi FM, and we saw what is common between the comics which do best on the Pratilipi comics or what is common between audio, which does best on the Pratilipi FM. And if a story is working really well in literature, does the odds of that story working in FM and comics is that much higher than a story which is not working very well in literature.

Sajith: When you say FM, what does it mean?

Ranjeet: Pratilipi FM is like Pocket and Kuku basically audio books or audio stories. 

So, that was the second part of the thesis. Does that story work across formats? Does that story work across languages? Once we have proof points of that, then that part of the thesis is proven, geography is still unproven. So, that is how I at least think about whether the business is working or not.

Sajith: Got it. So broadly three steps. Step one was seeding the platform. Step two was waiting for network effects to kick in. Step three was now if it works in one language, can I take it to other languages, formats, etc.

Ranjeet: Right. Languages, formats and geography.

Sajith: Got it. If I may ask, 2015, you launched it. So, what was the journey like? How long did it take for network effects to be visible according to you? Do you remember now

Ranjeet: Internally the team started seeing the beginning of network effects at the beginning of 2017. It was obvious in metrics by maybe by the end of ‘17, beginning of ‘18.

Sajith: End of ‘17, clearly visible. Okay, and how did you figure this out? Did you actually look at the cohort of who came in and them becoming writers and their audience coming in? How did you track this?

Ranjeet: We did track this as well, but a much better way to track whether you have network effects or not is to ask, does my time spent per user for example, increase with scale or decrease with scale? So, for all other kinds of companies except for consumer internet companies with network effects, as you go from a million MAU to 2 million MAU, your time spent goes down, your retention goes down, your frequency goes down. But businesses which have network effects, your engagement goes higher as you scale, your frequency goes higher. So, we looked at whether engagement is increasing over time or not as one proxy. Our time spent per DAU, when we raised our Series A used to be about 27 – 28 minutes and even that people said was really, really good. When we raised our Series B, it was touching 40 minutes, so much so that Qiming and Shunwei who led that round, they actually asked for technical DD to make sure that the metrics are actually measured correctly. When we raised our Series C, it was 60 minutes per user per day. When we raised our Series D from Krafton, it was about 84 – 85 minutes per user per day. Today it is about 95 minutes per user per day.

Sajith: Okay, and when you do this, are you using this for a particular cohort of users or is this across the thing?

Ranjeet: Across everyone.

Sajith: Wow. People are spending 95 minutes per day. Wow.

Ranjeet: This is for all users, even if you open the app by mistake, even then. For people who actually start reading something for them, this would be about 115 minutes, for paid users it’s about 210 minutes.

Sajith: Got it. We can delete what I’m going to ask you now, that’s not a problem, but if you can just give me what are the broad top-line number of users and paid users.

Ranjeet: Roughly, Literature would be right now at about 2 million DAUs. Paid users would be like if we only count active premium users, which is the primary monetization for literature, it would be about 300,000 now.

Sajith: 300K. Got it. And what are the other verticals apart from Literature?

Ranjeet: We have now basically five verticals. We have literature, we have Pratilipi Comics, Pratilipi FM and we acquired IVM podcast and Westland books. Outside of these five, we also licence our stories to other people for movies, web series, TV shows, stuff like that.

Ranjeet’s views on PMF

Sajith: Got it. That’s interesting. Yeah, so, do you want to just talk a little bit about PMF? I think I’ve got a glimpse of it. You don’t like it because it’s vague. Different people have different views about it, not actionable.

Ranjeet: Yeah, so I was saying the way I explain to people or the way I advise people is that from their business fundamentals, how do they know if the business is working? A good example is to think of let’s say something like Ola or Uber. Let’s say Ola for example. Now if you get a cab, let’s say you book a cab and you get a cab in three days, are you going to wait for a cab to arrive in three days? Probably not. Are you going to wait for let’s say five seconds versus 10 seconds? Is that going to make a meaningful difference? Probably not. So, for different people the answer might be slightly different, but typically maybe somewhere around three minutes to 10 minutes is where getting a cab would probably make sense for you. So, as a business, the way for you to understand whether the business works or not in a much better way than with what people call PMF is basically to get to the point where somebody can get a cab in three to five minutes or three to 10 minutes.

Once you get to that point, then you can look at whether economics makes sense or not and accordingly whether the business is viable or not. For different businesses, the answer will be different. There’s no single magical proxy metric that you can look at every single company and say, okay, this is PMF or that is PMF. In fact, as I say, I don’t even like the word PMF. Generally the word that I use for consumer companies is typically liquidity. Do you have enough liquidity in the system? But even that would not work for everything and that definitely doesn’t work for enterprise. More valuable is how do you know if your product is working and there is no right answer, you have to define that. How do you know whether this product is working or not?

Sajith: Now this word liquidity, can you just double click a little bit into this? It’s an interesting word and can you give me an example of it?

Ranjeet: Liquidity matters for companies which are marketplaces or clearing houses of some kind. So, for example, for something like Swiggy let’s say, or again Ola or Uber, let’s pick Swiggy for now. We did this exercise five, six years back because a friend was doing a food tech startup. We found that if you can get your rider to do 11 rides a day at that time, 2016 – 17, the economics made sense. If you cannot get it to that point, then it does not make sense. So, you need to have enough liquidity in a hyper-local area. Let’s say Koramangala, if you have a thousand orders in Koramangala, the economics makes sense, but if you have a thousand orders from entire Karnataka, the economics don’t make sense. So, what is that point where you would say that this is where the economics make sense? What is the clearing after which numbers actually start to make sense?

So, that is what I refer to as liquidity, which would matter or which becomes important for any kind of marketplace types networks. So, think of something like Pratilipi, for example, if you don’t have enough stories that you can read for the next three months, even if I acquire you, you come and you read for whatever two months and 29 days after that there are no new stories you leave. What would you do if there are no new stories? You need to have enough stories and enough users where that makes sense. So, for Ola like in the earlier example, from a driver’s perspective, as long as I make at least ₹1 more than what I was making before Ola, the system makes sense. If I’m making ₹1 less than what I was making at all before Ola, obviously the market doesn’t make sense, so my previous salary plus one is the clearing price for for a driver in this case.

Ranjeet’s advice for founders

Sajith: Got that. So, let’s say friends reach out to you or someone reaches out to you saying, look, Ranjeet help me, is my business working? So, one clearly is you look at the product and whether it’s meeting that customer need, second, if it’s a marketplace, you look at liquidity. What are the other things that you’ll advise founders? Are there any specifics for people who come to you who you may have angel invested in, etc. Are there specifics like guidance you give them?

Ranjeet: It depends on context like very, very little advice works in general, which is why I try to avoid giving generic gyan…which is why I think the disconnect is because you are looking for an answer that works in every scenario versus what I’m trying to say…

Sajith: No, no, that’s fine, go on.

Ranjeet: When people come to me, what I start off with is especially forget the jargon. Think of it in the most basic sense, what is the core value proposition that you are trying to solve? What is it that you’re trying to do, from a user’s perspective, not from your perspective, and then if you are trying to do this, does that make theoretical sense? Sometimes you can make sense on paper itself, sometimes it cannot. In which case, what is the fastest way for you to experiment and find out whether this makes sense or it doesn’t make sense? That is what we focus on. So for example, as I am saying, for something like Ola or Uber or Swiggy, it’s difficult to know for sure on paper whether something, whether the rider economics and the driver economics will work or not.

So, my advice would be to pick a very small area, let’s say Koramangala fifth block. It doesn’t matter how much money you spend here, it doesn’t matter how long it takes. Basically figure out if I get enough liquidity within Koramangala fifth block, does the economics then work for rider and driver? If it works for both, then I know that the thesis works. It’s basically just about scaling and execution and all the problems that come with it. But till that time, don’t worry about unit economics, don’t worry about anything else. For certain things on the other hand, you can make a theoretical sense even on paper. If I’m right, then this works in which case somebody else has already done something. Which is why there is no fixed one line answer, except for the fact that you should find what is the simplest way of describing your business from your customer’s point of view and then what are the biggest bottlenecks or biggest hypotheses that have to be true for you to be right?

Beyond the product

Sajith: Got it. So, let’s say the product is working okay. Somebody’s coming to you, it may not be a marketplace, it may be a consumer app product that is working. What next? Of course, you don’t use the word PMF, but as the product is working, what next, because there’s a journey of obviously economics, there’s a journey of GTM, etc. How do you think about that? What are your frameworks or internally you have divisions like FM, Literature which seems to be more advanced, clearly that has hit PMF in a way, FM or Comics may be yet to. So, how do you think about what happens next after the product is working?

Ranjeet: Again there isn’t a single piece of advice that works for everyone, but what I try to look at both in the Pratilipi and outside is that you try to remove the biggest risk as soon as possible. For certain kinds of products. Let’s say for example, let’s pick comics Now, for each comic that you produce, there is a cost involved. You have to spend some money to produce that comic. So, you’ll think very differently versus think of literature or think of Kutumb app for which unit cost is basically zero. It’s like messaging a text or writing a story in text. The only variable cost is effectively AWS cost, which is basically very close to zero. In that case, my advice generally is forget about monetization at all. Monetization is literally a feature. It doesn’t even matter how you monetize. There are 5,000 different ways to monetize as long as you have a large enough and engaged enough user base versus I give example of Pratilipi comics where I worry about monetization.

Now there could be other cases for example, e-commerce, but e-commerce not marketplace where think of consumer brands for example. You produce something for ₹5, then you’re selling it, but monetization is not a feature anymore. Monetization is what the business is. You cannot produce something for ₹5 & sell it for zero. You have to sell it for something higher than 5, there is no one fixed answer. If I can give a fixed answer, basically try to remove the biggest risks first and then only get onto the next phase of the risk for different businesses, what that risk is would be different.

Sajith: Got it. So, comics cost a certain finite amount to create. Maybe it’s ₹5,000 – 50,000 – 60,000, whatever. Okay, so how do you remove the biggest risk for comics then?

Ranjeet: Now there are two different problems. One problem is basically what you asked. Second problem is, if you remember I was saying earlier that for different kind of marketplaces or certain kind of businesses, at least economics don’t make sense till you have critical mass, unless you have critical mass economics do not work that’s it. Businesses with network effects for example, you cannot build the ads business for a social media company until you have at least a few tens of millions of users. Before that you simply don’t have enough data to build an ad engine in the first place. So, for something like comics for example, you would not know for sure, sure till you get to that scale, but you can do two, three things to get a reasonable idea. One thing is that you know, okay, this much is what a comic costs.

You can experiment with different pricing models to see what can probably stick and then you can work backwards from there and say, okay, these many users need to pay for this comic for it to be able to pay back the content cost that I’m creating. So, that is one proxy way of knowing whether it can work or not. Second proxy way of knowing of course is benchmarking businesses which are conceptually similar. That doesn’t mean other comic companies, that doesn’t even necessarily have to be other digital businesses. Businesses which are relatively similar in terms of concept, find out those businesses and then benchmark from them. Do they make enough money to cover their content costs so that triangulating other companies or other benchmarks. Let’s take comics for example. You can look at, let’s say Naver Webtoon, which is the largest comic company. You can look at Kuaikan or other comic companies in China.

You can look at Kakao Webtoon, which is another comic company and you can try and find out their monetization, how that works, how much content costs them to produce, and then triangulate from that and think that my content cost is this, my ARPU (Average Revenue Per User) is going to be roughly this and does that roughly fall in the range of where this would make financial sense? So, that is the second way to move. Third is of course sometimes you feel so strongly that you basically say, okay, I’ll take a leap of faith. I’ll get enough content and enough readers where I have enough critical mass and then I’ll see whether the monetization works or not. All of these three approaches are fine. Ideally the advice is that you triangulate all three or at least two of the three.

Ranjeet’s views on go-to-market or GTM

Sajith: Got it. How do you think about the go-to market? I think for you, the go-to-market is inbuilt into acquiring customers because you’ve got to seed it initially. How did you think about go-to-market at Pratilipi? Did you have to advertise to get people on board in the early days? How did you seed the market?

Ranjeet: For us, the beginning was writers sharing content. For example, if you write an essay, you are going to share it on your Twitter and your Facebook and your WhatsApp. It’s just what people do. So, that was the big growth lever early on. Then of course Pratilipi is a long-form text content, that’s like a holy grail of SEO. So, SEO became the second big chunk and third big chunk, maybe not for the first one and a half years, but after that was basically paid advertising. So, the combination, I think till year back it was basically roughly one third, one third, one third across all these three.

Sajith: So, paying people to come on board as writers, getting writers to….

Ranjeet: We have never paid writers not even acquiring writers by direct performance marketing. So, first was basically people sharing links, second was SEO and third was paid ads. Paid ads for demand, not supply side. Having said that, normally for most people, my advice is that you have to try out 10 channels. Sometimes you might be able to guess that this is what should work or sometimes you might be able to figure out, okay, these are the possible 2, 3, 4 channels that can work. Generally a good advice is that you spend a little bit of money on eight different channels. Sometimes it’s money, sometimes it’s time, sometimes it’s a little bit of both to try and experiment with eight, find out which one or two work. Basically spend 90% of your money or energy into one of those that is working and then 5-10% experiment with newer channels.

Picking the space

Sajith: Got it. Fair. Just before you decided on Pratilipi, were there other ideas that went through your mind and you picked one? How did you arrive at the space?

Ranjeet: Looked at quite a few. Towards the end I had basically shortlisted two. One was basically what Dunzo has done but with a different GTM and second was Pratilipi, so these were the two that I had shortlisted.

Sajith: And what are the process of coming up? I know it’s a bit a long time now, eight years, but what is the process of coming up? Did you follow any structured processes or something like that?

Ranjeet: Generally or in my case?

Sajith: Both.

Ranjeet:  In my case because I read a lot, so I kept on reading about both other startups and ideas and kept on talking to people randomly and some ideas pop up.

And after that as I was saying, think about whether this makes theoretical sense, if you’re right and it makes theoretical sense, how large can this be? And third, is it something very obvious in which case there’s going to be a lot of competition or is it something where there’s some kind of potential difficulty that exists in a starting and once you get to a critical mass then you have some kind of defensibility. So for example, I don’t like building or I don’t like generally, typically businesses where there are no network effects or economies of scale. That doesn’t mean they’re bad businesses obviously. It just means that it’s more about execution and more about how much money can you raise, how much money can you spend? And for some people that’s a good decision because they believe that they’re good at execution or they can raise a lot of capital, they can deploy that capital. I just don’t think that’s my strength area, either of them. So I would much rather prefer businesses where the winning team is the team that has the best strategy and then you have some kind of a moat which basically makes you immune to execution risks or monetization risks.

Ranjeet’s reading recommendations

Sajith: You said you read a lot and I just thought we’ll just head there a little bit. What have been your favourite sources of knowledge and insight?

Ranjeet: Almost entirely basically two – books and essays/research papers. Almost entirely reading. Just reading.

Sajith: Okay. And anything which comes to mind as something that you recommend internally or to the founders or the operators you work with?

Ranjeet: Usually what I recommend is to go on Twitter, and find out 10 – 15 people that you respect. Then go to their followers list and follow the people that they follow. That is the best way of building serendipity unless you already read a lot. Because for books and for a lot of the best books as well, over a period of time, context evolves and then some of that advice may not be relevant anymore or sometimes advice is relevant, it’s just that it’s relevant in a specific context and context changes or evolves.

Maybe a month back, for example, Sam Altman said that he feels guilty about the advice that he gave at YC because he basically had to break all of that advice while he was himself building OpenAI. It’s not that older advice was bad or this advice is bad, it’s more like it depends on context, what you are building, how much competition you have, the timing and so on, so forth. And I think Twitter does a great job of this, of basically exposing you to different viewpoints versus saying that okay, this is the one true way of building a business. There is no one true way of building a business. Typically, what I advise is to go on Twitter, spend half an hour, find 5 – 10 people that you respect, go through the list of people that they follow and keep on kind of churning from there.

Sajith: Got it.

Ranjeet: Obviously, if you have to start at some place then Paul Graham’s essays are probably the best place to start.

Sajith: Yeah. Any YouTube channels, any folks you particularly have enjoyed following apart from Paul Graham?

Ranjeet: Non-fiction as in business related? I think someone that I have both learned quite a bit from but also disagree quite a bit from, two people would be Marc Andreessen and Naval Ravikant, like 10% of their stuff I find hard to agree with, but 90% of their stuff is very useful.

On metrics

Sajith: Interesting. Yeah. Got it. Let’s segue to metrics. In general, what are the metrics that you specifically over-index on? Again, it might be very specific to your business. One of course, it is with liquidity, but in general for other businesses, what are the metrics you look at?

Ranjeet: Depends on company to company, there’s no easy answer, but for consumer companies, maybe 80% of the time frequency is the most important metric. So, you can define frequency in multiple ways. You can look at DAU/MAU ratio, you can look at the power user curve, how many days a user comes in, you can look at different other ways as well. The important part is that you don’t lie to yourself and look at one metric and keep on tracking that, not like this month DAU/MAU is down, So, you’re looking at sessions per user. Next month sessions per user is down, so, you start looking at time spent and all of that. As long as you’re consistent, you are not lying to yourself. Then I think for 70-80% cases in consumer internet frequency is the most important metric, but again, it depends on business to business.

For Flipkart, probably AOV (Average Order Value) would be a lot more important, and gross margin would be a lot more important than frequency as an e-commerce company. For something like Pratilipi, which is long-form high shelf life content, engagement would be the most important metric, not frequency. For a messaging platform, frequency should be so high that it doesn’t even matter, so you have to look at retention as a better metric. In high CAC businesses, retention makes more sense. Different businesses would have different metrics that would be important, but at least the startups that have I have met, I think 70 – 80% cases, frequency is the most important metric.

Sajith: When you say frequency, it’s typically DAU/MAU.

Ranjeet: It’s basically how often a user uses your product? Depending on the use case, sometimes DAU/MAU but sometimes even a hundred percent DAU/MAU would not be good enough. Like messaging for example, it should not be DAU/MAU, it should be multiple times a day. You want to look at sessions per day, all days active per week for some kind of businesses. 

Ranjeet’s thumb rules

Sajith: Got it, okay. So, typically when we folks come to you for advice, etc., are there things which you definitely tell don’t do or other caveats, warning signs or thumb rules?

Ranjeet: I think the biggest mistake that I have seen smart people make is that… till now at least, I have let’s say, invested in 15 companies more or less, and it’s not really an investment, it’s not a meaningful part of anything, but just driving home a point, I’ve only seen till now three cases where the company’s performance hasn’t been significantly better now versus when I invested. In each of the three cases, they have made the exact same mistake and almost every smart founder or 80% of the smart founders that I’ve seen fail. They made the exact same mistake that they listen to VCs and let’s say a Sequoia would say, do X, a Nexus would say do Y, and both of these could be good advice, don’t get me wrong. It’s just that when you combine X and Y, it’s absolute shit advice.

You could have done X, it could have been a great company. You could have done Y, it could have been a great company. If you went with your initial thought process Z, it could have been a great company, but once you combine X and Y and Z, then it suddenly becomes bad advice. For example, if you keep on changing your tactics too quickly & you don’t give it enough time to compound, again, the same thing as network effects for example, by definition only work for people who are on the network. Which company has the strongest network effects in the world? WeChat, WhatsApp, YouTube. But let’s pick WeChat. If I don’t use WeChat, what does WeChat’s network effects mean for me? It means shit to me because I’m not on that network. So, if I was building WeChat, the only thing that would matter to me is how do I get my network size to be as large as possible before anything else I would worry about.

Versus markets are down & somebody tells me that let’s focus on revenue, so I start focusing on revenue. Somebody else comes up and says that, okay, AI is big, so let’s focus on AI. 

None of that matters if it’s a network effects company because you want to increase the size of the network, same thing for other kinds of businesses, but basically keeping focus and working on a strategy and avoiding advice that doesn’t align with that strategy. Similarly, avoiding what is hot, which doesn’t align with that strategy. I think that is the single biggest mistake that I’ve seen smart founders make.

Sajith: Interesting. Yeah. Anything I should have asked and haven’t asked?

Ranjeet: I dunno. I came with an open mind, like anything that you wanted to ask, I would be happy to share types.

Sajith: Got it. Ranjeet, it’s always interesting talking to you. So, the way I look at PMF is that I’d like to think of it as two parts and you touched upon some aspects of it. One is you said, is it a working product? That is what I call Product to Problem Fit or PPF. The second is what I call MMF or motion-to-market fit where you’ve identified these kinds of people, who are paying for your product. Now we’ve got to figure out and find more people like this across the world or across India, and you need an efficient, effective playbook, which doesn’t take a lot of money or which is the GTM motion and figuring that out, combining the two PPF and MMF gives you PMF. So, I have a more structured definition, but that’s fine.

Ranjeet: I’ll tell you why I don’t think this works.

Sajith: Yeah, tell me.

Ranjeet: Think of Facebook. Who do you think their target user was?

Sajith: So, Facebook initially started out with Harvard College students.

Ranjeet: That is a very reductionist MBA’s answer. I’ll tell you what the answer was. Chamath, who was their VP Growth asked Mark, what is our target customer and how do you want me to reach them? Which is basically what you defined as your second step. Mark’s answer was, everybody who’s fucking alive, I don’t give a shit how you get them, just get them. Billboards, TV ads, use SEO, use whatever else you have.

I’m saying 99.9% of cases where the large companies that work out, they do not work out in an MBA class case study type manner. They work out by saying, I’ll try out 10, two of them will work, I’ll scale them.

Unfortunately or fortunately, what we are trying to find is a generic answer in a world which is driven by power laws. Only the most important companies matter and the most important companies will break almost all the preconceived rules any which way.

Sajith: That’s true. But there are very few companies like Facebook where everyone in the world is the target audience.

Ranjeet: But I’m saying the only companies that matter, they will always break conventions. OpenAI was a good example. Sam (Altman) said, all the advice that I gave was incorrect. The problem is in a power laws world, it’s the companies like OpenAI and Facebook that matter.

Sajith: You are absolutely right, but at the early stage it had to find love in a certain audience. Well, for a long time, Facebook needed .edu to sign up. Then they opened it up. You’re right, eventually, great products are universal. But how do I know it’s universal till it reaches a certain stage. So, my definition of PMF is a scalable GTM (go-to-market), enabling predictable, repeatable unit and positive customer acquisition with high retention. It’s a mouthful and it covers all bases, and of course it’s a very MBA definition, all that type. Scalable go-to-market because you need a predictable, repeatable unit positive customer acquisition and there has to be repeatability of your acquisition and high retention. So, yeah, you’ll say that now I’ve created a definition which covers all bases and I broadly feel yes. But for a founder, when I give him or her this guidance, I feel like they know what to work towards because the biggest thing is great founders, people who have succeeded had a very intuitive way like you, Kunal Shah etc. Now you’ll say, why are you putting me in the same category and all that. I find you are creative business people, you have a very intuitive way of grasping things and you struggle sometimes to put it into words and you feel like it’s too generic. So, I feel like you have a very intuitive way and I can’t rely on that (route) because I’m not a founder, I’m a VC. To young people who come to me, they’ll say, Sajith you can’t talk like Ranjeet, you can’t talk like Nischay, that I think is a challenge I have, and I have to be true to that.

Ranjeet: Makes sense.

Sajith: I know. Anyone else you think I should talk to? Who should I talk to next according to you? Who are the people you go to for sparring, discussing, etc?

Ranjeet: It generally depends on context, but typically the people that I reach out to who have been helpful for me would be Sahil from Delhivery. Then maybe Nishith from Locus, he is in your portfolio. So, he should be more easy to reach out. Then Vidit from Meesho. Both Gaurav & Hemesh, but even more Hemesh at Unacademy. I have only met him once, but I really liked talking to Hari from Big Basket back in the day, four or five years back, I don’t even know if he remembers me. But still, I think he has different thoughts compared to most of the founders. Maybe not the strategy but some kind of question like culture or People Ops-related questions. I think Mekin (Maheshwary) is really, really good with the top names that come to mind.

Sajith: Yeah, great. Cool. I think I’ll pause it here. Yes, I’ve got the names. Thank you for your time, Ranjeet, it was a pleasure!