Crunch Closure Note
Internal Closure note at Crunch via email from Aug 2024 repurposed here for record-keeping.
Hi Everyone,
Over the past few weeks, I’ve been reflecting deeply and we have decided to wind Crunch down and return unused funds to our investors.
At the end, if anyone has to be blamed, it has to be me. Startups are hard and it is that hardness which makes them fun but I made a couple of very big decisions that in hindsight did not play out as I hoped.
Choosing to base the engineering lead out of India made sense at the time for building a cost efficient team especially with the talent pool there but with AI disrupting everything so fast it turned into a real challenge keeping pace with the innovation happening elsewhere. I have spent time in the U.S. and have first-hand witnessed what a Bay Area-first engineering D&L looks like. I strongly think anyone truly innovating and building something substantial should have an engineering lead based out of the Bay Area.
Secondly I have always thrived in an urgency-first, move-fast culture. Speed has been my biggest advantage. Somewhere along the way balancing that speed with giving the team clear focused direction got misplaced. I was handling most of the customer meetings myself communicating every insight I gathered and pushing for quick experiments based on what I was seeing but it ended up creating too many directions at once, overwhelming at least some folks in the team. Sticking to one path longer might have yielded better results than jumping between experiments because we were all hoping for quicker wins and getting impatient when they did not come right away. We would sow seeds in different areas but move on before giving them time to grow and this loop meant a lot of ideas floated without getting the deep input they needed from the team leading to ambiguity blind spots and eventually affecting morale especially with some engineering gaps that built up over time. Part of the reason this happened was because of our sour memories of not being able to monetize what we started with which was DAOLens - an end to end infrastrucure tool for DAO management.
A venture scale business eats discipline for breakfast innovates at the edge of the world learns whatever is needed to innovate like that within weeks and most of all calls for a degree of hunger and aggression in closing without which a team lands exactly where we are, no matter what you build. We let numerous deals slip by or did not push hard enough on them and that breaks me more than a 100 customers saying no. This was largely because of my messed up chemistry with my co-founder. Something we should have fixed long long ago!
Startup journeys are hard. In the process, everyone int eh team responds with their own normal. While, I have always enjoyed the challenge of a venture scale outcome and get my inner drive from the same, there were mis-alignments between what an ideal scenario for the business should be within the team. It took me a while to realize the mis-alignment and by the time I did, quite a lot of damage was already done!
I tried everything I can to make Crunch and our product work. It did not. Today, when we’re returning back more funds than what most people even hope to raise, it’s with the realization that in the long run dragging this venture out will lead to nothing. The mistakes I talk about resulted into effects that will completely have to be undone for Crunch to be venture scale outcome - something that we always aspired for. The first idea of winding down came about a couple of quarters back and I immediately started flying to all my customers to give it all the lifeline it deserved before a decision is made. It would be unwise to post 100s of email thread finding customer discoveries here but those emails reflect how hard we tried.
Over just the past two quarters, I must have taken 250 plus meetings and 100 plus separately at events. I met whoever I could from all different types of industries, tried cringy things, tried new pitches distant from what we did as a business, pushed content on social (did Tiktok and Facebook reels), published articles into forums I never intended to post in, joined events that I would never typically attend (sound therapy session hosted by a couple of prospective buyers), pretended to be some other vendor and reached out to companies so that they do not have “who are you without any repiutation” as a question. I have harassed people in kitchen areas of co working spaces to push for leads. I rented an apartment with a founder helping people of colour get into YC (and tried my best to leverage his network). I have done media collaborations, given TV interviews, worked with coaches, joined communities like Pavilion where Execs hang out, invited people for podcasts so that I could speak to them, messaged people who went to events that I did not attend pretending to have met them and faking my way through it and occasionally getting caught, hosted webinars, did company wide presentations in town halls for some established brands, tried System Integrator route and initiated stuff there with folks like Accenture Acubate, hired US-First GTM team and got some US born friends to do some cold calls on my behalf as I sat next to them to navigate cultural nuances and participated in events that are domain specific away from the world of data and full of our buyers. I have tried creating fake aliases to get into events that are only for retail players. None of it seems to have worked to the extent that could save our business and scale us to venture level outcome.
The baseline is that in this market buyers do not want to take risks as it is their job on the line and no decision is the best decision. Moving with a decision others have gone for (incumbents) is the second least risky decision. Incumbents had a distribution advantage traditionally. However now tech wise as well they are equally capable of leveraging LLMs and there is no differentiation between what I have on top of Open AI and what they are offering in their Gen AI capability suite.
I have seen giant companies close small clients within a week and that is largely because they have a plug and play tool and can just move forward without having to build new things. In this era where Gen AI and getting a co-pilot to do something are the least differentiating things most of what I am trying to pitch comes across as a very average implementation. For folks in the US terms like a semantic layer we cure hallucinations we follow humans in the loop and we replicate human behaviour are no longer turning heads.
Overall there is a strong theme of building these capabilities in house and upskilling their internal data team to adopt AI and go bottoms up. More people today look at Co-pilot as something that should be developed in house and that is because over the past year that they have played they have only realized that it is not that tough to build co pilots on top of LLMs that get their specific jobs done. Picking a niche works well for some vendors as I have already anticipated.
I was trying to draw a parallel with other domains where new companies in recent times have gotten significant traction despite the presence of an incumbent. New products like Perplexity have a fault advantage meaning Google cannot implement the same thing Perplexity has overnight even if they could build it. However analytics has its challenges and is different from search. Here you cannot have inaccurate answers. The incumbents in the BI space have a line of trustworthy products and are adding Gen AI capabilities where the users are okay with errors. Trust is earned in drops and lost in buckets and if I play the analytics game the game I am playing it looks like I only have things to lose.
I have moved back and forth on our target market and all conversations point to one thing that there is no differentiation and there is no clear cut bottoms up distribution visible in sight.
I tried different pitch themes.
On the theme that talks about a personal assistant a delay in getting answers is not really a problem for most strategically important executives. They get the answers whenever they want. No one is in a rush to get to a place where they have to ask questions and get answers on their own.
On the theme that is around envisioning a world where things will be automated and executives would lean on AI for making decisions and reviewing what is being done need a cultural change. Cultural change is a decade old problem in data industry and there are things like governance etc that need to be solved beforehand. I might be too early but I have come to realize that being too early is the same as being wrong.
On the theme where I try to be another dashboarding tool I am still waiting to run into someone who needs better visualization and is struggling to find one. Some tools only do visualization and they get the job done. No leader I have met is asking if they can change chart types. In other words visualization is barely a differentiating factor and any energy I spend around that today will probably not help me get closer to PMF unless I pivot to be a tool that helps people visualize data in the best way. No one thinks of me when they think of visualization and I think that is understood why. Further I need to remember that I am talking about a venture scale business.
The next pitch theme is where I just go and pitch better decision making but in all honesty and through numerous conversations better decision making seems to be built best on top of an existing tool. The minute you start thinking of telling people what they should know to make decisions it may look like I am discussing a dashboard free future and tons of integrations. However human beings are limited by their experiences and dashboards make their way into the picture. The minute a dashboard comes in an AI powered chat comes in and then everything I have been building comes in. When you take a birds eye view of just this theme it becomes clear that an existing tool adding decision making capabilities will likely do much better.
What is worse is that people do not remember what I pitch meaning the differentiated value is not only absent but none of the other values resonate even a little. Whatever pitches I have tried do not excite them.
Now differentiation can also be in terms of pricing and not only features but there is not much that I have been able to put forward here because of my large dependency on models and how my cost depends on LLMs directly. In fact most financial services people feel that 15 dollars a seat is also high of a high price per month and this is because Power BI came up with this price in one of their experiments and did not get adoption. Establishing a proof is challenging with subjective evaluation criteria like making it easy to get answers. People feel if something is easy or not when you show them and the minute I show them what I have it goes back to what am I trying to replace.
As far as market verticals go here is my pulse and here is a summary of my learnings. I had already ruled out real estate (no use case) and airlines (strong players already have delicious offerings). Here is my thesis on the other market verticals. The automobile industry has a very low TAM (meaning fewer conversations) and paired with how I need to gain experience with the automobile sector makes it not an excellent vertical to start with. It takes a lot of investment for someone to be a well established player. Events are coming back but the entire event industry is again only a relationship and numerous stakeholders point out that events as an industry still not ready to focus on analytics. Then comes BioTech and Healthcare. I spoke to 3-4 people who have worked with biotech startups. The use cases that make sense are retail like for example placements of medicine. For anything that goes into the body of the patient there are regulations and consequences that come into the picture and it is best to steer clear of this vertical since I do not have a background in it either. Energy as a vertical sounds very promising and has tons of gaps but that is a hardware play which I am not very keen on. Financial Services are using tools like Snowflake and trusted vendors present on Snowflake. Also Financial Services have a strong motion to build things in house because of regulatory concerns. E commerce and retail looked like the best bet and even there for the past couple of quarters tons of conversations have led to nothing because of all the factors mentioned above.
At the Snowflake Summit in SF in June I ended up striking a few conversations with vertical owners on how I was thinking of positioning Crunch and if there was a Powered by play possible where Snowflake could support me and I build on top of Cortex (their AI co pilot infra play). Snowflake seemed open and as the chat advanced with respect to getting in front of specific customers to check viability they came back with question around how I was different than other established partners on their platform like Thoughtspot, Zenlytics, Lovelytics, Sigma and Lightdash. Though I responded in all honestly I never had a strong answer against what my unique differentiation is and what my unique market insight is.
There are some very very good learnings and this email will never do justice to all things I have figured out. Bay area is way too advanced. There are people moving mountains with teams smaller than ours and the kind of things people do in their MVP today is far ahead of what the world standard around MVPs are largely because of competition. In the bay area any vendor with no revenue also has benchmarks that they have tried establishing. For example our model is 95 percent more efficient than other models or things like 300 percent faster than any alternative etc etc. Established players were a headache anyway but I am also up against very established unknown competitors who are leagues ahead because of their engineering and then all the geographical advantage is of course on top. As a team despite all things I have built I think I have come across as less aggressive overall.
Overall it looked like there was not much scope and I might just have been terribly placed to make a sustainable business with Crunch. Ultimately the customers perception is the reality and that perception is that what I have does not work or excite them enough.
I reflected internally and came to the conclusion that this will not work and that wind down would be the best course of action. We will keep you posted on the proceedings and the process related to the return of funds. We would always be communicating clearly with you and we value the trust you have shown in us. I'm proud of how the employees who are now placed at the company are doing really well. Reflect on their time here and say that we had everything right and that we just needed to persist a little bit more on some of the ideas that we sunset a year back that we needed to be more. So. I speak to my investors and they reflect back on the journey. They say that we did everything that we had to do right. And it was a bunch of variables outside of our direct control which made our path despite so much of hard work so bumpy.
Ultimately, I was the captain of the ship, and I steered it to a place it should not have been steered to. When we started DaoLens (Now Crunch), it was with the idea of giving an honest shot to what seemed like a super risky but bold opportunity that should be evaluated thoroughly. We pivoted solely because we realised that even after a year of effort, we could not put a pulse on when and if DAOs would support a business where outcomes can be predicted.
Experiments followed, and Crunch was born. As we’ve worked over the past year on analytics, there were numerous hits and misses. Overall, we lost conviction for a respectable venture scale outcome, doing what and how we were doing. It has been a blessing exploring and working with some of the best investors and mentors, who are all now friends and supportive of everything we did. When we went with the idea of winding down, we discussed at length what our role in the world should be, how privileged we are and how it makes sense to keep failing spectacularly until you make it. I have witnessed first hand how you can be super kind with your gesture, and all I hope is to keep paying it forward.
Now, for my team - The learnings that came will not result in an ordinary outcome. So it’s a matter of time, and when that happens, I’ll forever be grateful specifically to every single team member who partnered with us, trusted us, and gave it all they had. Our journeys might be different, but personally, wherever I walk, I’ll walk with all your learnings, hopes, aspirations, and blessings — cheers to the Crunch Mafia 🥂.
If there is one thing this journey has taught me, it’s the usefulness of chasing bold ideas, whether they materialise or not. It shapes you for phenomenal future outcomes. Something in the journey shapes you, and I've fallen in love with what this journey has done to me. So, onto more and more dreams.
Thank you!