Once you figure out mindset and intangibles to startup, you can focus on the more tangible parts, which brings us to -
Part 2 - Problem / What to work on
Characteristics of a business worthy problem
All business is problem solving. However, not all problems make great businesses. Here are a few heuristics:
🎯 The problem should exist
This sounds obvious but a lot of business ideas are made-up ideas that nobody wants. Make sure what you’re working on is a real problem.
“We will make it work” is a great attitude but doesn’t work when there’s no demand.
💰 People should be willing to part with something valuable to solve that problem
Be it money in most cases or even time / reputation. If not, the problem is not important enough to be solved for that group of people.
In 2021, I was working on an idea for emerging Indian creators to monetize using communities. Realized soon enough that this was a good-to-have for them, not a must-have because they were not willing to pay or spend a lot of time.
🏄♂️ It should have some tailwinds
I do think TAM is overrated as a concept and I’ll explain why in a future post, but suffice to say that you should be working on problems that have growth and are not a part of declining trend.
💡 Working on the problem should be interesting to you
Great founders make great businesses. And a crucial component of being a great founder is interest in whatever you’re doing. I’ve written this last but it might just be the factor that ends up determining success or failure.
I’ve personally made this mistake earlier, where I picked something to work on because of the logical factors but didn’t give enough weight to my own interest. That’s not sustainable.
There are a lot of other factors like moats, brands vs commodity etc but they are a function of solutioning, not picking the problem and you should ideally separate the two initially. If you’ve an entrepreneurial instinct, it will be natural for you to see a problem and jump on it. But before you make it a business (in your own head), make sure you’re picking on the right problem to work on.
Non-existent problems
A key feature of a problem that startup solves is that it exists. This begs the question, why? Why is it not solved yet? 🤔
Firstly, business problems are not like mathematical equations, so they never get completely solved. That would mean all customers are as satisfied as they can be, and we know from Bezos that it’s never true. So the question actually means why is not broadly solved?
If you look closely, traditional business owners don’t need to answer this question. You have a kirana store opening closer to your house than your previous one, you will go and buy from there. Similarly for a lot of other traditional businesses.
There is definitely a problem that this kirana store is solving, you were going 1 km for buying groceries, now you can get it within 200m. But it is only incrementally better than what you were getting earlier. But you didn’t feel it until this new kirana store opened.
But with internet, going and buying groceries was no longer the most efficient solution. Hence, we went online. The business model of ecommerce however forced us to plan for groceries in advance. With quick commerce, this is the latent problem getting solved.
This example illustrates problems existing because of
➡ Not being solved by the most efficient technology (Ecommerce)
➡ Not being solved to the most efficient behavior (Quick Commerce)
There may be other factors also because of which the problem exists:
📈 Sudden increase in demand (E.g. GPU shortage)
📉 Sudden reduction in supply or supply quality
⛓ Forced constraints (e.g. regulations)
😏 Customers don’t care enough (e.g. most commodity businesses)
Any others that come to mind?
Anytime you notice a problem, it might be worthwhile to ask why it is not broadly solved. More often than not, you’ll realize it is already being solved to the optimal level that it could be given the constraints of the industry. But once in a while, you’ll realize that some other solution can do a much better job.
That’s when you’ve hit gold.
Why Now?
Every business needs tailwinds to succeed but they are often misunderstood.
Sequoia had a question which is now famous, “Why is this the right time to solve a problem? Why will 2 years later be too late and 2 years earlier be too early?”
The intent behind this question is to know if there’s a systemic shift in something which will lead to massive disruption. A lot of whether your business is VC fundable or not will come down to this question.
A rough heuristic for a VC business is if it can reach ~$100Mn revenues in 7-8 years.
E.g. India is becoming wealthier, incomes are rising, consumerism is increasing etc etc
This is true, and a good tailwind to build any business on. D2C brands are riding on this wave, so is ecommerce and a lot more sectors. However, this is not enough.
This is because this is an incremental tailwind. It has been true since a while and the growth rate of India’s consumption, while good, hasn’t massively accelerated vs the past.
Disruptive tailwinds have a much faster rate of change e.g. AI, regulatory changes, etc
One of my favorite examples is the US investing app, Wealthfront. Within 2 years, it reached approx $2Bn AUM (equivalent to $50Mn revenue). The tailwind then was the brokerages making APIs to make trading easier via integrations and the founder knowing about it. Back home, low cost internet, UPI and now UPI on Credit have been some of the biggest waves to ride on.
Some things to note though:
🦄 Disruptive tailwinds are very rare, and it’s not uncommon to confuse an incremental tailwind for a disruptive one
🤔 It’s very hard to predict these upfront, usually you notice them because you’re already in the arena / domain and get exposed to these
📈 Sometimes these disruptions occur later in the journey (e.g. Aadhar, UPI, Covid etc benefitting Zerodha) or can be a result of multiple incremental tailwinds
What does it mean for an entrepreneur though?
1) If you’re riding a tailwind, understand whether it’s incremental or disruptive and take funding accordingly
2) Even if it’s incremental, it doesn’t prevent you from building a large business, but might take more time or additional factors. So passion has to be the driving force here, not the tailwind.
TAM is overrated
Everyone fusses a lot about TAM, but I think it’s overrated. Here's why:
1️⃣ Firstly, TAM should be calculated at an order of magnitude level. It’s not important to know whether market is $800Mn or $1Bn (that is just guesstimate error) but more important to know if it’s $100Mn or $1Bn or $10Bn (i.e. log(TAM)). Seen this way, market size is a 5-min calculation, not something you spend days fussing over.
2️⃣ Secondly, focusing on just TAM ignores a very crucial component, market growth. If you want to build a high growth business, it is critical to enter a market which is growing very fast.
Generally, at any point, there are very rare large and fast-growing markets. You are more likely to find
➡ Large and slow growth market (e.g. Tesla)
➡ Small and fast growing market (e.g. Facebook)
If you have to choose between one, which one would you choose? It depends on what kind of business you’d want to build and your existing advantages, but I’m much more in the Peter Thiel camp here to pick the latter and dominate it first.
A large slow growing market will also have more competition from incumbents and puts more onus on differentiation whereas a small fast growing market provides a wave that everyone can benefit from.
Today’s Yahoos are not stupid enough to let the Googles fly under the radar - Differentiation vs incumbents is that much harder and comes over time. 🤷♂️
The limiting factor in choosing such markets is that the growth might tap out and then what? That is when founder imagination comes into picture. Usually there are adjacent markets that you can unlock through your existing capabilities. Some of these may not be obvious on Day 1, you only go there and figure out next paths.
One error you should definitely not do is pick a small market and say ‘this will grow in the future’. It is either growing right now or it’s not, and if it’s not, don’t be a fortune teller.
In summary, TAM is definitely important but a large TAM market becomes relevant only when you have strong differentiation vs incumbents or high market growth. If none of these are present, you’re better off indexing on market growth vs TAM.
Micro first, Macro aware
A lot of founders think like investors and lose the plot. It all comes down to micro vs macro.
Micro is all about being great at questions like -
👨💻 Who is the user?
⏩ What problem are we solving for them?
⏩ Why is this an important problem to solve?
⏩ How is the solution going to impact their life?
Macro is all about being great at questions like -
📈 How big is the industry and growth rate?
⏩ What slices of the industry are still untapped?
⏩ What are the barriers to entry?
Obviously, you’d ideally want to be great at both. But if it comes to micro vs macro, what do you think is more important?
It’s micro.
You can intellectualise about the industry all you want but you dont have a business if you dont crack the micro.
Why do people get trapped in macro though?
Because micro is a necessary condition for creating a business but not sufficient for great success. You may just solve an unimportant problem and not create a great business out of it. To avoid this pain, a lot of people start thinking macro first but never end up defining and solving the real problem. It is right for investors to index more on macro, but not for founders, because the founders have to live with the problem they are solving daily. So it better be real (and also important to them).
💡 The ideal combo is “Micro first, Macro aware.”
This is also linked to “Execution informs strategy”.
When in doubt, trust the micro. You will need to solve for the wings and engines of your plane before worrying about the winds.
Ideas vs Problems
Great businesses are made from problems, not ideas. What’s the difference?
When you start with an idea, you make the following series of assumptions at once:
➡ There is an underlying problem that is unsolved
➡ People care about getting it solved
➡ The solution the idea proposes is significantly better than existing ones
This is where a lot of delusion happens.
All 3 are not small assumptions and need to be proven / disproven. This is even before the fact that there’s an entire execution journey with the idea.
Starting from a problem means you first define your user and what gap exists.
How do you assess the importance of a problem?
💰 Ask users if they will part with something valuable (time, money, reputation) to solve the problem. Ask if they would be willing to do it now. If it’s not a clear yes, you are in iffy territory.
E.g. Enough people talk about creating hyperlocal news platforms, do a lot of people want it in the first place?
🎯 Ask how they solve the problem currently. If they are not making any efforts right now, it’s unlikely they will make an effort in the future.
E.g. For most SaaS companies, there’s some hacky worfklow already on Excel or other tools that solves for the current problem
😣 Go deeper to understand the pain with the existing solutions. Sometimes existing solutions solve problem optimally, technologically or behaviorally, but sometimes they are just placeholders until a better product comes along. This is hard to gauge but you’d rather be in 1st scenario than 2nd.
’The Mom test’ is a terrific resource to understand the art of problem validation better. This step is critical if you want to nail the micro aspects of the business.
Listen to your heart
Founders often skip the most crucial step during problem validation.
That is listening to your heart.
You find that the problem exists and you jump in straight on, without thinking whether you will really enjoy solving the problem.
This is generally a function of wrong motivation, but can also be impatience.
Why wrong motivation?
You were looking to startup / become a founder, finally you have found a problem that exists, this is your golden ticket to creating wealth and getting free, so you jump in. 🤑
This usually ends in a disaster.
Any startup takes years to build and the daily grind is so tough, that outcome can’t be your primary motivation.
My favorite example here is Zuck rejecting a $1Bn Yahoo acquisition offer, even when the board, including Thiel were in favor of taking it. His response was - “What would I do with the money? I’ll just create another social network and I already like the one I have."
It is crucial to dial up your human antennas and ask why you should work on it.
Here are some questions you can ask yourself at this stage:
1️⃣ Do I really care about this customer? Or am I just like “give me money and get my revenues up”?
2️⃣ Do you like the problem / problem space? If you were not working on it, would you still spend time thinking about it?
3️⃣ If nothing comes out of this problem for 10 years, would you still want to work on it? Would you regret not working on it?
4️⃣ Do you feel a natural pull towards the problem or does it seem like a chore?
Once you have validated that a problem exists, take some time off doing nothing about it, and see you frequently and excitedly you think about it.
Obviously, no one wants to fail when we start a startup, the questions are there to ascertain how you feel about the problem, because that feeling will generally stay throughout.
It is supremely hard to be honest about these questions, but you’ll save a lot of time and misery if you are. This axis is the one that separates missionary vs mercenary founders.
Ideas I worked on before starting up
Idea 1: Duolingo for STEM students to enable 'learning by doing'
Typical nerd founder starting out with a mission, and despite all the noise in edtech, education remains a powerful empowerment tool. So we were trying to productise learning for college students in coding, design, etc.
This was stemming from our own experience of working with college freshers and not seeing what we wanted to see in terms of skills.
We talked to a lot of students, identified needs for each persona and then ran an MVP.
Our MVP was basically a project that enabled students to learn by doing.
We ran a couple of batches with students to see the pitfalls in the idea. Realised quickly that there was huge delta in starting point of students even from similar backgrounds and level and that also determined high manual intervention required in coaching + doubt solving etc.
Essentially we got blocked at Bloom’s 2-sigma problem as we couldnt scale 1:1 tutoring then without ops and none of us were excited to take that part up 'at that time'. With GenAI now, that may not be as true anymore.
Anyway, a few of my personal learnings:
1️⃣ Mission-oriented problems are great for motivation :D
2️⃣ My strengths and inclinations in business
3️⃣ Domain expertise matters (we would have known these things if we were already in edtech) but can be acquired fast (through good experimentation)
4️⃣ Doing things that don't scale. We did the right MVP but we could have been more persistent on the ops piece and seen it as a necessary part to get to productised scale.
5️⃣ Education is about outcomes, very few people learn something as a pursuit of craft or excellence.
Idea 2: Fan clubs for small and mid creators
I’d realised the advantage of domain expertise from the first idea I worked on (link below), so this time I started with something I knew much more strongly.
From my experience at ShareChat, I knew that it was tough for small and mid-sized creators to generate revenue. And from my chatrooms experience, I’d known what a great power community can be in terms of monetisation.
Hence decided to combine both to enable smaller creators to earn even before they become famous. We were specifically interested in solving this for skill-based creators (mostly from YT/IG) e.g. musicians, dancers, artists etc.
We talked to a lot of these folks and received good intent in terms of solving this problem. Hence, we built out an initial version and started running some communities with these creators.
We quickly realised a few things:
➡ The effort required for these creators to run and maintain these communities was much higher than they anticipated
➡ This led to a lot of creators prioritising brand collabs for monetisation, which gave them a large reward with one video, rather than running a paid community (almost) daily
➡ Interestingly, one other way creators looked at this was - “I’d rather spend my extra time making one extra video in the hope that it goes viral than spend time on the community”
Some of these insights also came down to status vs wealth driven societies. 👯 💰
In status driven societies, fame is prioritised over money (which is what we saw) vs our learnings on US-based creators who looked at monetisation much sooner (hence the success of Gumroad, Kajabi etc).
My personal learnings here:
1️⃣ We had domain expertise to know this problem existed and how it could be solved, but we should have been much more rigorous in terms of the urgency of our target creators to solve this problem. In a way, we over-indexed on our domain expertise but under-indexed on the problem validation rigor.
2️⃣ Once we had low intent from these creators, we also generalised this phenomenon across all types of creators. In hindsight, we would have been more successful with Finance creators but that would still be a temporary win imo.
Idea 3: AI copilot for product analytics
Data analysis was of course a regular part of work, and I wanted our product team to be much more efficient then. There were a couple of Product Analysts on the team and they were perpetually overloaded.
The idea soon evolved to a virtual Product Analyst sitting within Slack, who will not just write SQL, but with enough context also answer why things went wrong (upto a level).
This was July 2022, so pre-GPT days. Today, it’s called Analytics Copilot.
Analytics is a brutal industry and I realised it the moment I started researching on this. I spoke to product teams in other companies, and even CEOs of analytics companies, including those who had shut down and also those who had exited. And everyone unanimously echoed the sentiment of how hard distribution was in this industry.
Distribution is anyway generally hard, so that feedback was taken with a pinch of salt 😅
However, feedback from most users was to have something like this in their existing tools and not move to another tool (which was also expected).
This meant cracking distribution and a heavily integrated product at the same time.
And this was with a full-time job on the side. 😥
As the mountain of challenge in front of me seemed to grow bigger in size, my motivation started dropping. And there was no one in it together to keep the motivation going.
And as anyone who has gone to the gym knows, having a gym buddy is the biggest consistency hack. 🏋♂️
This idea didn’t even go to MVP and I believe two reasons played a big part in it:
1️⃣ Not having a co-conspirator in the journey (My need, can be different for different people)
2️⃣ Me working on it part-time (had some circular reasoning on when to leave)
The point of sharing these examples is to illustrate that just the problem aspect of startup building has so many variables and dimensions that are overlooked before people jump in. And we haven’t even touched team, solution etc.