Prukalpa was one of the first founders I thought of speaking to, when I conceived this series of talks. She was one of the few founders who had tweeted about PMF (Product Market Fit) and her take on it. I had also heard her speak on the topic, on an audio talk where we were fellow panelists. Nonetheless I hesitated, for I wanted to be better prepared for our conversation. I finally wrote to her last month, and she agreed to join me to share her thoughts on PMF.
A bit about Atlan. Think of it very simplistically as Github for data teams; a collaboration hub creating context and trust for all members of the data science team. In March’22, it announced a raise of $50m led by Salesforce Ventures, valuing it at a post money valuation of $450m.
In our conversation, Prukalpa talks about the transition from SocialCops, a profitable fast-growing services play with finite limits on scale, to Atlan, a product co with theoretically no limits on scale. Prukalpa and Varun, her cofounder, were obsessed with scale and in that pursuit of that obsession, they had to take very tough calls, including deprecating two profitable but ultimately smaller scale ventures than Atlan. In honing in on Atlan, the founders spoke to over 150 data scientists to understand their workflows, and suss out their pain points. This deep understanding of user pain informed the product deeply, and clearly was a factor in its eventual success.
Over the course of the talk, Prukalpa covers her definition of PMF (pull, repeatability of motion), why founders selling is a dangerous crutch for the business especially when founders rely on selling largely to those they know, and her recommended resources for founders seeking PMF. I really enjoyed this candid conversation with Prukalpa, and came away deeply impressed with how in their pursuit of scale, they were willing to make hard sacrifices including shutting down profitable units, to focus on the one thing that could scale. I hope you do too.
(Please find below, the transcript of my conversation with Prukalpa Sankar. 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 13th May 2023.)
***
Sajith: You seem to have a fairly intentional way of defining PMF. I’ve read your Twitter thread which told me about your 3-tier approach to PMF, problem-value fit, problem-solution fit and then finally there is product-market fit. I’ve also gone through 2 of your podcast transcripts, Moonshot, which was really interesting and the other one was Return on India. We also appeared on a talk where I took notes on your views. So, this is a conversation I have wanted to do forever, but I felt I should do a few more conversations before I did this.
So, what I want to do is start by focusing this conversation around the first part of your framework, that is problem-value fit for the first 15-20 minutes, and I’ll tell you why. My definition of product market fit is that there is a product to problem fit (PPF) and then there’s motion to market fit (MMF) and it’s almost sequential, except a little bit of a loop that you need to go back to iterate on PPF depending on what you find in the market, but.
The first part of product to problem fit is the pick, where founders take their time to hone in on the problem. What I found interesting was you did 150 interviews with Humans of Data as you call it, I was utterly fascinated by that. So, I had been collecting these cases where founders do these large number of interviews before they decide what to build. So, I found it quite interesting.
It might help if you could just talk a little bit about the SocialCops to Atlan navigation and how did you pick the space & the interviews that you did and process maybe could just double-click into it before you finally decided on the space. All yours.
A search for scale
Prukalpa: So, in our case, honestly, the intentionality of product market fit comes from a lot of failures more than it does from success. We failed many times in doing and so it came down to we just have to get this right. So, it was a journey to get to that point and happy to talk about that.
For us, the place where it all started was we happened to be a data team at Social Cops, we were running a consulting business & we were doing quite well. I mean for a consulting business where you’re growing 2 to 3x year on year. We were working on dream projects with the UN and the National Data Platform for the government. So, our journey to figure out what’s next was not coming from a place where a lot of startups start from, which is a founder says, I want to go do a startup and it didn’t start from there.
It actually started from like, okay, we’ve scaled to this point but at this point, we have this journey where we can scale a consulting business. You can scale 2-3x year on year, you can do really well and it’s very profitable, but that’s the only way you can really scale. You can build a high-quality business like a McKinsey-ish business, you of course can build a Infosys kind of business and I have a lot of respect for entrepreneurs who do that. But it was just not me; Varun and I couldn’t build that company. And so, that’s where it started from, which was that scale mattered to us. We had in five years achieved a lot of scale. We never meant to build a consulting business. At Social Cops, we meant to solve problems, it just happened. The only way to solve those problems was to do consulting for large organisations because they had the reach.
And so we had hit a lot of the places it all started from us, we had this Dream Wall in office when we started where people would put up their dreams and it said things like, we want to power the Prime Minister’s Office and we want to drive national level policy and we want to solve problems like this. And every year, when we started the year, people would move the chips on the Dream Wall from dreams to achieved. Around 2017 – 18 a lot of these dreams had come true, so we asked what’s the next decade? That’s basically where the starting point for us came from. So, that context might help in terms of why we had to be super intentional in our journey to figure out PMF.
So, at that point, we happened to be one of the very early modern data teams just because of who we were. Today, it’s more obvious what that looks like. In 2018 it was not; the world was very, very different. Snowflake had 12-13 customers or something. Very different world and we were one of those early teams and around 2018 we built a bunch of internal tools for our own team to make our team efficient and agile. A lot of the scale that we saw, and profit margins, a lot of it was because of the internal tools that we had built and the place where it all started for us was a question which was like we build these internal tools for ourselves. So can we go figure out if these tools will help other teams around the world? That’s basically where the starting point was.
There actually were a bunch of internal tools & Atlan was one of them. We had two others actually as well that were internal tools that we had built and that had actually the potential of getting somewhere. So, we basically bet on three different possible product markets essentially around 2018. Funny enough, these two products actually hit about a million dollars in ARR (annual recurring revenue) actually. And we had to make a very difficult decision to shut those down and say we will only do Atlan. But we were very clear that if we do the next phase, it had to be a very big problem and a very global problem.
Because the reason we were doing this in some ways was a search, it was a personal founder search, for scale. That was where it came from. We had a lot of fulfilment with what we were doing at SocialCops. So it was not coming from that, it was coming from the thought that in the next decade, we want to build something iconic that has scale, fundamentally big, where you hit hundreds and thousands of teams that are using your product that was where the search (for scale) came from. And because it came from there it was very important that whatever we picked had the ability to go that big. That was where the starting point came from. And so the first thing we had to figure out was, is this a big enough problem in the first place and is it a top of mind problem?
So, that’s basically where we started interviews. We actually used to do two things, the way it started was we had our own data team internally, and so we used to actually do something called Friday <incomprehensible>. My co-founder Varun used to run this session where we would call all our analysts and we would sit down and it used to be this open-ended problem session where people would rant and we would capture that. So, that was actually the first set of interviews which were internal and then on we talked about what really contributed to their work being better.
The second was open-ended interviews with data people. We at that time made a decision to go after not interviewing leaders but interviewing practitioners. Our hypothesis at that point which was that the leaders were too far removed to actually understand the real day-to-day problems.
And back then, I still remember 2018-19 when I would tell people we were thinking about building something in data, people would be like, all data problems are solved. Because if you remember that world, there was Hortonworks and Cloudera IPOing. So, they said everything was solved, we don’t understand what was left to be solved. And so the only way to figure out if we were real was to talk to more practitioners like us and check if it is just that we are running very unique projects and so we face these pains or do all practitioners face them. So, that’s basically how we got started with the Humans of Data.
The three product lines, and why Atlan won
Sajith: I found the part about the 2 other projects really interesting and I just wanted the sequencing. So, did the product creation come after the 150 interviews were done as in did you have a hypothesis that we are facing these problems, so build these three tools, let’s talk to these 150 people and get a sense of which of these tools really matter to them?
Prukalpa: So, actually, there were three product lines that were not for the same audience but different markets. So, there were three product lines or tools that we had built internally for ourselves for the types of problems we were facing. One was actually around primary data collection, the other was Atlan in its current form, which is helping manage your internal data / data management and the other was around external data, alternative data. So, they were completely three different products, three different product lines and three different markets in themselves. One of them was actually a very well-built product, it was essentially covering a tonne of the national data collection drives across the country. So, there wasn’t as much of a market discovery exercise as much, but it was never SaaS. So, we had to figure out SaaS stuff on it & see how to monetise it so, that was one angle.
The other was an external data product or an alternative data product. Think of it as a data marketplace. So, we built internal tools where we could predict affluence down to a building level. So, let’s say you give me a latlong, I could tell the person’s affluence, probably this is how close they are to banks, ATMs, and things like that. So, we were able to predict where diseases would break out and so there was a tonne of things that we were able to do. That was the second line and the third line was Atlan in terms of the product. So, in retrospect, this was a very good decision. We actually didn’t make a bet on which of the three they were. So, Varun and me, we split and said, let’s go figure out which one makes the most sense.
And we did that. For some time we didn’t make a decision. But for Atlan, we did 150 interviews, we also did a lot of other stuff for the other two products & that actually went on for a while. It was maybe about a year and a half in that process of figuring out what makes the most sense to go into long-term. And through that actually is why we became so maniacal about product market fit. Because we hear these things like the Marc Andreessen post, that the market is everything, it’s not the product, you read it and you think you understand it and you actually don’t. You only really understand it when you go through it and you realise it. Our other two product lines, I mean the external data product I was telling you about, it still probably is the coolest things I’ve ever built in my life.
Till date, I don’t think I have ever built anything that is cooler. If you actually think about the impact, I remember one of our customers, they made an additional hundred crore of revenue that quarter because of us. It was impactful stuff, but it was not that the problem was not there and the solution was not there and the product was not there, all of that was there. It’s just that there wasn’t enough of a market. It couldn’t scale. We could have easily taken any of these other product lines to ₹20-30cr a year in business, that would not be a problem, it was just that there was nothing beyond that. The market wasn’t ready. Maybe at some point in 10-20 years, the market would be ready for what we built, but the market wasn’t there at that point and that realisation of the market was probably the turning point. So, eventually, the thing we did was we just bet on the market with Atlan. We just bet on the market. We didn’t bet on the product as much as we bet on the market and in retrospect that was probably the best decision we made.
Sajith: Got it. So, just to articulate broadly, around ‘17-18, you figured that, there are these three opportunities, first was primary data collection and you had set up tools built for that, you figured out the government or some UN-type multilateral organisation should value it, but it would best become ₹20-30cr and it was not a scalable motion, not repeatable. You had to go in each time to convince someone.
Prukalpa: Not necessarily one was an online SAAS product by the way, up until now it’s been very hard to replicate it, we got 200-300 customers on the product, completely online, credit card based etc. So, it wasn’t that it wasn’t scalable, it’s just that the market was, not big enough.
Sajith: Got it. This is the primary data collection product.
Prukalpa: Yes. It was basically a mobile app to collect data. So, we had people using it for monitoring their retail stores or warehouses. We actually had customers in some 50 countries.
Sajith: Wow. You couldn’t spin-off? You could have taken three kids from the team & got them to run it or something like that.
Prukalpa: We could have, it was just too much mind space and effort. It was very hard, the product is still not deprecated. We’re finally deprecating it this year because we had customers that stayed on and their businesses ran on it and so, we stopped taking new customers and said whoever’s running it continue to run. We had customers asking us, can we just buy the product from you and things like that. We could have chosen to spin it off, sell it, whatever. But at that point we had already found Atlan and whatever transaction you do, it’ll still take time from the founders. And we were just like, it’s not worth it, we are not doing it for the money. So it was just not worth the effort in some ways.
The search for pain
Sajith: Got it. In these 150 interviews, what were the kind of questions you were asking? Were you talking to mid-level practitioners, people actually building modules or behavioural folks etc?
Prukalpa: Data analysts, data scientists and data engineers. Typically that was the audience and it would be like tell us about your day. What do you do on a daily basis? What are the biggest pain points in your day? What are the biggest problems that you face? Out of all these problems, what is the one thing that you would pay out of your personal pocket to solve? So, pretty open-ended type interviews to understand the daily workflows and the pain points in those daily workflows and what are the things that keep these people up at night.
Sajith: Workflows & pain points in the workflows, got it. And through them, you got a priority and felt there were 5-6 key points and that gave you a sense of what to build for, some of it you should have already built in Atlan.
Prukalpa: Yeah. So, the thing it really validated for us was how big the problem is because, through these open interviews, almost everyone we spoke to would talk about the pain points that we had faced as a team ourselves as one of their top. And so it validated for us that actually irrespective of whether it is a data team in India, whether it’s a large enterprise, a small company, whether it’s the US, whether it’s the most ahead of companies, whatever it is, all of these people actually still face this pain point and it’s a real problem for the practitioner. I think that helped us validate not just the problem, which is the mistake we made with the other two products because in both cases the problem existed, it’s not that the problem didn’t exist, it’s not that the solution didn’t exist.
The problem was that the problem was not big enough, meaning there were not enough people in the world that faced the same problem, which meant the market was not big enough to build something that was iconic. And that was what we took away from Atlan or the early layers of Atlan; the internal product used to be called Athena at that point & for Athena, it was like, everyone faces this problem, it’s real. We talked to the top data teams in the (Silicon) Valley and they were telling us it was a problem, but we also talked to large enterprises in India and they said, this is our problem. And we are like, okay, this is a problem everywhere. So, that was the biggest thing we took away from those interviews.
Sajith: Cool. So Atlan was originally Athena, but I suppose it became Atlan for the keyword (SEO) and all of that.
Prukalpa: Yes. Athena was the internal product name.
Sajith: Internal one, okay, got it. If I may ask, what was the real pain point or two? One was collaboration, I presume it was very painful to collaborate. Is that one?
Prukalpa: Yeah. So, our biggest insight and it’s beautiful how this kind of lives through to today. My co-founder Varun had written an internal vision doc back then in 2018 in terms of the problems. Actually the doc basically talks about the problems and that over the next year, we will explore these problems to figure out what the solution is. But it came down to our core thesis which was diversity in the data team. That the data team, unlike other teams in the world, need a lot of diverse talent to come together and collaborate. So, to make a data project successful, you need a data engineering skillset who can scale pipelines, an analyst who understands the data, someone from a business context who understands the business & you might need machine learning engineer etc, I don’t want to go into all of that, but they’re just very different from the way the data engineer works, very different from scientists, very different from analyst, very different from business.
And because of this fundamental diversity, these people don’t understand each other. So, there’s a very big lack of context and knowledge. There’s a very big lack of trust because if you think about it, the problem we would have is that a number on a dashboard is broken. If so, can I trust this data, it could be because the pipeline broke and it could be because the data engineer’s not doing their job. It could be because the analyst wrote the wrong model. It could be because the business user is understanding the wrong metric; we don’t know how ARR is defined. All of those things could lead to the number not being trusted. Now the problem is as an analyst I can’t see a pipeline, I don’t know what a pipeline looks like. As a business user, I don’t understand any of these things. I’m just like, man, can you give me the number? Why the hell is it so hard for you to give me a number? Because I only see Excel and I’m like, I can do this on Excel. Why can these guys, with all these investments I’ve made, why can they not figure this out? But it’s not that easy because live data pipelines have n number of things that can break in it and there’s lots of complexity in the data pipeline and there’s complexity in the people and the end tools. There are 5-6 different tools, there are 5-6 different people, that was the core thesis. And our thesis was that this diversity never goes away. This diversity is the biggest strengths, it’s also the biggest weakness. And so how do you help these people work better together? That was basically where it started. So, how do you build trust? How do you build shared context? How do you, you make this team as successful as it can be is basically what our core hypothesis or thesis at that time was.
Prukalpa’s definition of Product-Market Fit
Sajith: Got. Very cool. So, do you have a broad definition of product market fit? I know you have a 3-tiered framework, problem-value fit, problem-solution fit & product market fit. But as such, do you have a definition if a young founder come to you and asked, how would you say, I mean doesn’t have to be a formal definition, but whatever comes to your mind.
Prukalpa: Yeah, it’s actually really funny because I still keep going back to that Marc Andreessen blog post. I read it when I was 21 but I only understood it when I was 28. But it says this thing where when you get product-market fit you know it feels it’s product market fit. It feels like someone is pulling the product out of you faster than you can build it. And up until that point it doesn’t feel that way. And I still remember this moment, it happened with us for Atlan, where in Atlan also we had to rebuild the product twice and there was this one quarter where I remember we relaunched the product. And up until then for about a year or year and a half, we had one customer. And suddenly in that quarter we went from our 2nd customer to our 10th customer in a few months. I remember I was sitting in this call with an enterprise customer in Latin America it was ridiculous and they were just coming to the pipeline and we were doing a demo and they were like, this is what we want to do in the trial and we want to buy in two weeks. I still remember that moment it was midnight in India and I was sending a message to one of our founding team members who was on the call with me and I was like, this is what product-market fit feels like. Up until then as a founder, it feels like a push, it feels like you’re pushing, pushing and at some point, it feels like you stop pushing and it’s working, like someone is pulling it out of you. Honestly, that’s what the feeling feels like and it’s the best way to describe product market fit.
I think from there you can find better definitions. Another definition for me was, can you do the same thing again and again and get repeatability. So, another soft thing for me was that I sold our first 10 deals myself and somewhere between our third and our eighth deal, I realised I was doing the same thing. I was not doing anything different. My call was the same, my discovery was the same, they were telling me the same problems, I was showing them the same demo, and they were like, yes, this works. And they were like, I’ll buy, it wasn’t any different. I wasn’t trying to solve a different problem, it’s a different solution to different problems or things like that. It was then I realised someone else can do this. The founder doesn’t need to be doing this. Someone else can do this and just do it repeatably. And that was another way that I used in the early days to validate product-market fit.
Fitting the product to the market
Sajith: Just what changed from the first to the second product, you said you had to rebuild it. Did you have to put in second functionality? What changed actually?
Prukalpa: With our first design partner-customer, we built the ideal solution. It was the ideal solution with zero constraints around change management we just built everything. And nothing really changed in terms of the front end of Atlan. We got to a tonne of user value with our first customer. We went from 100 users to 500 users organically in about 3 months. So, the product was working; the problem was that it was too much change management for most other companies to adopt it. So, we would do these demos with people and people loved it, but how do I get this set up? And we’d be like, you have to change your entire data infrastructure. And people would be like, I can’t do that. Can you find an easier way for me to, so what we really shifted was we shifted our deployment model and the backend to make the product very lightweight so that we could just fit in into an existing setup. It was not a lot of change management for someone to bring us in. That was basically the rebuild that we had to do.
Sajith: I also read that at some point you decided we won’t do any of these multiple things, we’ll just focus on AWS and Snowflake, which meant you had a narrower market, but now adoption could be faster because you geared towards it. Was that around the time you did this?
Prukalpa: Yeah, so basically what happened is we did the Humans of Data interviews and post that we were like, okay, let’s double down and actually build something so we can figure out if this works. So, we picked one design partner customer and we went very deep with them and we had a pretty core thesis of what we wanted to build. We knew what the solution should look like. So, we actually built for about, I would say about, six months we were just building and getting to user value and things like that.
Around that time we didn’t get enough data, actually, we made a bet. So, for the other two products, we decided the market was not big enough. So, as soon as we knew that the market was big enough for Atlan and we had early user signs; I remember with our first design partner-customer, there was this one time when the product was down, they were at about a hundred users then and the product was down for half an hour and the analysts said we have to go take a chai break because we can’t do anything without Atlan right now. So, they went for a tea break, which was the very early signs where we were like, this can be something that helps people work. It’s kind of a negative, but if not using a product makes people’s life hard, it says you’re building something that matters.
And so we saw early signs and so we made a decision to basically say we will shut down the other two product lines. It was actually very hard because with Atlan we were not making money at that time. It was like we were $10,000 a month, or sorry, under $10,000 a month with our first customer. So, we had a $100k ARR and everything else was about a million dollars. So, our other two product lines were actually much further ahead from sales and all of that perspective. But we were like, we just need focus, so let’s bet on the market. We shut down the other two lines and then we started buyer interviews. So, that was our next step, that was the market phase. So, we were like, okay, problem validated, solution validated to a certain extent, let’s figure out which market do we want to go sell this in.
So, that’s when we started buyer interviews and that was probably one of the hardest things we had to solve with Atlan because Atlan’s a horizontal product, so it means that you can sell basically to anyone. So, the same thing that made the market really big also meant that we could sell to banks versus we could sell to high-tech companies, we could sell in the US, we could sell in India, there was too much diversity in the customer base. So, we were trying to figure out who do we go sell to? So, that was the buyer interview process, which is who’s going to buy?
And through that phase, we realised that the best way for us to figure out our market was another traditional way people slice and dice, which was industry or company size but it was technology maturity. Because we were trying to find similarities going back to why would 10 people buy the product the same way. It’s like what is similar between these 10 people? And in our case, the similarity even today doesn’t come from the scale of the company it largely comes from the tech stack of the company. So, that was the first insight, which is that the wave we have to slice and dice our ICP comes from data stack or tech stack maturity. And then the second decision from there was what do we bet on? AWS or Snowflake was the decision we made. So, we decided not to do anything else and just do AWS and Snowflake as our first ICP.
Figuring out the GTM
Sajith: And the playbook for customer acquisition for the first 10, were they targeted off a list or had you worked with them in SocialCops? How did this evolve, the first 3-8, you said you got this Latin American customer which may have been second or third or fourth or fifth. How did you get these customers?
Prukalpa: Yeah, so one thing that was very clear to me that we didn’t want to hack our way to the first 10 customers. That was very clear. I knew we could sell it. I knew if we call people, we can hack it, we can easily do it. If you know how to sell, you can hack it. Honestly, I say this in a very nice way because people like founders and people want to work with you and people want to, if you are selling yourself, any founder in the world can get to their first 10 customers without product market fit. And so it can look like product market fit, but you don’t. So, our first 10 customers cannot be people that we know, it needs to be real customers.
Literally, by that I meant we were not using our investors, we were not using our board, we were not, we did not want anyone to do us a favour to get us to our first 10 customers. That was really important. So, then it was like, okay, then how do you go acquire customers? So, we did two things at the same time, we had Google, so we went after search because that was the closest to buying intent that people would have. If they’re searching for your product or the solutions, we went after a tonne of search keywords because we thought that would be definitely one way of getting the most relevant customers into the pipeline. And we also did outbound, so we had our ICP and we did outbound. Eventually, the unlock of the 2 to 10 customers was all inbound. So, it was all via search, it was all inbound. That’s how we closed the first 10 customers.
Sajith: Yeah, just one query. So, when you said you went after a search, was it a SEO or was it ads?
Prukalpa: SEO takes too long so, it was the wrong thing to invest in first. So it was ads.
Sajith: And that was the outbound motion and there was no inbound motion, right?
Prukalpa: That was inbound.
Sajith: Oh, so, what was outbound, sending emails?
Prukalpa: Sending people emails, cold emails.
Sajith: Cold emails. So, some of that happened and the inbound was ads?
Prukalpa: We were doing both. Eventually, it was the inbound that worked.
Sajith: Outbound mails + inbound from ads and eventually the unlock came from inbound. Yeah, got it.
Metrics to judge PMF
Sajith: Got it. So, in terms of metrics to measure PMF etc, by your definition, one is there is an intuitive push to pull transition and then there is repeatability of motion where you feel it can go to somebody else in your team & anyone can start deploying. What are the other metrics, anything that you would look at to say PMF? Are there any other metrics, anything that at all you would add to this
Prukalpa: In SAAS? I don’t think so. It just needs a lot of intellectual honesty from the founder to say that are these 10 customers really the same. And I don’t even know if it’s 10, maybe it’s five if you’re a large enterprise product or whatever. But are these the same? Did I do the same thing for, 10 is a good number, but you know could argue what that number is, but that it just needs a lot of intellectual honesty.
Her advice for younger founders
Sajith: Especially now that you are a little further down the path, younger founders will be reaching out to you for advice. Are there any caveats, some rules you kind of share with them saying, hey, do this but don’t do this. Anything at all?
Prukalpa: I think I’ve single-handedly convinced a lot of people to focus on the US market; a lot of Indian start-ups I have single-handedly asked to focus on US market. But there are two or three things that I would suggest as a founder, first getting a tonne of clarity on what you’re building for is very important what you want. It’s actually the last thing that people put. You always prioritise yourself last. But as founders it’s really important to know what you want.
Do you want an exit in two years? Do you want to build something global and large? How much time are you willing to give for that to happen? That personal clarity on what you want is important. And if that doesn’t happen, are you okay? A lot of people talk about the clarity of what they want. The problem is it’s hard to get there and so then it becomes hard to make decisions against it. So, for example, in Varun & my case, we were very clear, we are only playing to win and become a global iconic company, that’s it. And that means that we are okay to die. So, if we don’t get this outcome, we are okay to die. And it sounds like it’s a very easy thing to say, but it’s actually very hard as a founder because it’s your life’s work, it’s people, it’s humans, it’s all this stuff.
And just saying I’m going to shut it down if I don’t get there, is a very hard thing to come to terms with personally. But once we came to terms with that personally, the decisions we made, we had nine months of money in the bank when we decided to shut down a million dollars of revenue, that was a very hard decision. We had one customer, we did not have honestly have the data to say that let’s bet on this, we could easily have said that continue with this for some more time so we can survive. And if we had done that, we would’ve died because three months later Covid happened and we would not have gotten any customer discovery interviews if we had gone any slower than we did, because we put that constraint of we have nine months, it’s win or lose. We were able to get to product-market fit but if we had delayed by three months the company would be dead by now.
We would not have existed or we would’ve been in survival, we wouldn’t have died because we would’ve found a way but we would’ve survived. We wouldn’t have gotten to where we wanted. And so I see a lot of founders would say, I am in India, I know people here, let me sell to my first 10 customers here, but my long-term market is some other market. And I’m like, that’s 10 because you know what? Your first 10 customers take energy and time and resources. The hard thing about getting customers that are not your right long-term market is that you have to service those customers. So, you can’t find product market fit. You want your first 10 customers to be the ones that can help you scale your market.
They’re the ones you should go from 10 to 100 with. Now if you have the other market of 10 customers and then you are saying at 10, I’ll find another market, you just don’t have the time and the resources to do that. It’s not possible. And so if you know that this is not your long-term market, disqualify early and play for the long term. So, the third thing is, the reason people make the second decision is because it’s easier short-term, meaning you can show progress in three months, you can show progress in six months and we are all humans we need validation. So, it gives you validation. I still remember our second customer, how do you go tell a team that has been working for a year to get a product right with one customer and no validation beyond one customer. How do you go tell that team that actually this is a second customer and we have to build just two connectors for the second customer, but we think it’s the wrong direction.
So, we’re saying no. How do you do that? It’s hard. It’s really hard. And so most people don’t do it because you’re like, it’s okay. The team will feel you’re making some progress. Maybe one engineer will not quit. As a founder you need validation, but it’s the wrong long-term decision. Now these are all very easy things to say in actual practice it’s very hard to implement day to day, but which is why the first thing is very important coming to terms with what you want and what you’re ready to give up for that to happen. And if you have that personally, emotionally, then it helps a lot to make these very difficult, tough calls. They’re bets that you make. And they might be wrong bets. Maybe you are spending too much money on something, they’re all bets. So, the question at that point, if you can go back to why are you doing what you’re doing, what are you playing for? And if you have clarity there, then it becomes very easy to then make these decisions.
Resources and people who have helped her
Sajith: Yeah, thank you for this. I’m going to just go into the last five minutes. You have recommended Disciplined Entrepreneurship as a book, right? Are there any other resources that you recommend for founders either SAAS or otherwise?
Prukalpa: Disciplined Entrepreneurship and The Mom Test in the product market phase are the two books that changed my life.
Sajith: You also recommended that Segment founder Peter Reinhardt talk. Was there any other founders who you spoke with along the journey who you got clarity from or any investors you spoke with, any conversations that helped you?
Prukalpa: Yeah! So, going back to the first thing which is the clarity on what you want in life that is where I got one help. I still remember one conversation with Lavish Bhandari, he was the founder of Indicus, sold to Nielsen, a very renowned economist, who built some very cool things back in the day. And I remember this conversation with Lavish where I was telling him about what we are building and blah blah blah, Lavish looked at me and he said Prukalpa when you go to the US you’ll find entrepreneurs who are 1/10th your capability that are running companies that are 10 times your size.
He was much older, he had retired by then. He was like, I was not in a position personally to make that position, but you are, you’re at a point in your life, you’re young. So, if you can make that decision, you should. That was one conversation that for me changed just my ambition, the way I was thinking about the market, the way I was thinking about the biggest decision we made between Atlan and the other two products was the other two products were emerging markets-focused products. They were not global developed-market products and Atlan was the only product that could do global.
So, when I said we made a decision on the market, we just made a decision that there was nothing else that we made a decision on. And so that conversation with Lavish was very insightful. Wanting repeatability, that was another interesting one, I remember speaking to Vaibhav Agrawal from LightSpeed at one point, that the thing about the first 10 customers being very different is it’s a lot of fun. So, if you think about every customer they’re telling you interesting problems, it’s a different problem, you want to go solve that problem. As an entrepreneur you get high from problem-solving, that’s your thing. So, if it sounds like an interesting problem, you want to find a solution to that problem, that’s who you are.
I remember I said to Vaibhav I don’t know if I want to build a SAAS. It’s repeatable. It’s the same thing again & again, I don’t know if I want to do that. It was just personal discovery. I remember Vaibhav saying, what you’re describing to me sounds like a lab and I’ve never heard of a lab that scales and that was another very powerful thing for me where it helped me realise without repeatability you can’t scale. So, if you need scale, you need repeatability, which means you can’t get a high out of every conversation having a different problem. You just have to mentally change what you want out of a conversation. Maybe those two were my most powerful conversations. A lot of my conversations with entrepreneurs were around that first thing on getting clarity on what you want. I went through a journey going from SocialCops to Atlan to discovering what I wanted in life. It was hard because my identity was SocialCops for a very long time, and so it was not an easy decision to move from that and say something else.
Sajith: If I may just refer back to your Moonshot interview, there is one lovely passage which I’ve highlighted says quote “you have to in the early days solve the problems that are most likely to make you fail long term. That is, the hardest problems that are most likely to have long-term competing failure” and goes on. But I really love this part where you’re actually saying just solve what is going make you fail. So, that was interesting as well.
Prukalpa: That was not me, that was Patrick Collison. So, I just learned from him and applied it.
Sajith: For funding I think your first investor was WaterBridge & this happened before you got your first customer or after you got your first customer?
Prukalpa: It was before actually. I remember Manish (Kheterpal), I don’t know what Manish saw, man, I mean I have zero idea. I was like, we’ve built these internal things, this is what we do and we think that these internal tools can be helpful, we think there’s something here, but we have no idea which direction this is going to go. Manish just bet on the founders and I am very thankful at that time to give us that space. And he actually gave us that space. We were in product market fit for a very long time from 2018 to 2020.
Sajith: Oh, they came in 2018, is it? Oh wow. Okay.
Prukalpa: Manish gave us a lot of space to experiment and shut down. Things that a lot of people would not have given us, that space, and gave us a lot of trust to get that right.
Sajith: Thanks Prukalpa, this was a great chat!