Originally published here, reproduced below. Part 2 coming next, let me know what you think!

The way to get good fintech startup ideas? Not to try to think of good fintech startup ideas. My advice to entrepreneurs is to look for money problems rather than money solutions.

Most aspiring founders start by naming a sector. Payments. Capital markets. The office of the CFO. Then they go hunting for a product to build. In my experience, that’s backward. Starting from a sector is how you end up building a slightly worse version of something that already exists, because you’ve inherited the same map everyone else is looking at. The sector might tell you where to look, but it doesn’t tell you what to build.

Instead, start with the money and look at how it moves.

Money flows through a business. It comes in, it moves around, it goes out – and at every step someone records it, checks it, approves it, or carries it from one system to another. Wherever that movement slows, it’s a stong signal of friction. Perhaps it is a person re-keying numbers from one screen to another; or a reconciliation that runs once a day because that’s when someone has time; or two or more pieces of software that were never built to talk, with a human sitting in the gap between them.

If you want to find fintech startup ideas, follow the money until it hits a bottleneck – and then look hard at why it stopped.

Fintech startup tests

So far, so good. But sadly that’s not the end of the story. Not every bottleneck is a business.

The next stage is where I often see founders flounder. A point of friction is only worth building a company around if it passes two tests.

Test No. 1: Is a fix possible?

Can the tools we have today solve the problem you’ve identified? Three things have to be true:

• The work has to be a repeatable, automatable workflow, rather than a string of one-off judgment calls that look similar from a distance.

• It has to emerge from and throw off data, because without data there is nothing for software to learn from or act on.

• The problem must be contextual, by which I mean it must index on information about how the business actually runs: who hands what to whom, which system feeds which and what has to be true before the next step can happen.

If your problem involves workflow, data and context, you probably have something that is “possible” to solve. Especially where AI can be applied effectively.

Test No. 2: Would it be transformative?

Will solving this problem actually change anything? Will a customer’s business be meaningfully different afterward? Enough cost needs to come out, or enough headcount freed, or enough value delivered that paying you is an easy decision rather than a line-item negotiation.

One strong indicator that a bottleneck clears this bar is that the customer is already spending heavily on it. Perhaps the work is even outsourced. Better still: it’s a large line in the P&L, something the CFO already sees and most likely wants to reduce.

A CFO does not care whether they pay Acme Services or Acme AI. They care that there is a cost on their books – not attached to any of their own headcount – going out the door every month for a job they would rather not think about. Take a real bite out of that number, and you have a buyer. That is a very different starting position from trying to convince someone they have a problem they hadn’t noticed.

Both tests matter. A good fintech idea lives in the overlap, where both answers are yes. Many founders only check one of them.

What this looks like in practice

Let me make these tests more concrete with an example of a company I know well. Stacks makes software that handles the financial close, the monthly process of getting a company’s books right and signed off. Let’s apply both tests to this bottleneck.

Take possible first. The close is a repeatable workflow rather than a series of judgment calls. Doing the close gives you the full financial picture of the company, because you cannot close the books without touching all of its financial data. It also gives you context because you have to understand how every department and every tool feeds the numbers. Thus, here is the workflow, data and context, all sitting inside one task.

Now we come to transformative. The close is something every company does, every month – and almost nobody enjoys it. It eats senior finance time on a fixed cycle that never stops. Take the close off their plate, and you’ve potentially changed how the finance team spends its month.

By solving for the close, Stacks found a fintech idea in the key overlap between possible and transformative.

Finding a bottleneck you can solve

Which brings me to the long-term potential in finding a money bottleneck that is both possible and transformative: it usually points you towards the next bottleneck you can solve.

Once you have solved the close, for example, you are not just sitting on a close tool. You are sitting on the company’s financial data and a working map of how the business runs. That is exactly what you need to solve the next bottleneck in the flow, and the one after that.

This part is worth dwelling on because it is where the value in your fintech idea can compound.

A successful fintech company does not expand by solving one problem and then, one day, deciding to become a platform. You expand because solving one constraint hands you the data and the context to see and reach the next one. The pipes are connected. Clear a blockage in one place, and you can suddenly see further down the line than you could before, and you arrive at the next blockage already holding most of what you need to clear it.

The platform is a consequence of removing friction, not a plan plucked out of thin air.

Follow the money to where it slows and, at each bottleneck, ask the two questions. Can you solve it? Would solving it change the business enough to matter? Where both answers are yes, you have most likely found something worth building, and usually more than one thing, because bottlenecks often sit next to each other along the same flow.

Only then should you worry about the sector.

In my next article, I will walk through a handful of enterprises and apply the principles outlined here. We will look at real sources of friction, real areas where businesses may be able to create value, and where I think you might find compelling fintech startup ideas. Coming to you in your inbox tomorrow!

These are structured as notes to myself:

1. Lollapalooza is everything

This is the most important learning, and I’m obsessed with it

Charlie Munger coined the term ‘The Lollapalooza effect’. When I read it many years ago in Poor Charlie’s Almanack, I didn’t fully realise the extent of it. Many years later, as a full time investor, I now fully appreciate its importance in life and investing.

Simply put, the Lollapalooza effect is defined by Charlie Munger, in the Outstanding Investor Digest as follows:

”The most important thing to keep in mind is the idea that especially big forces often come out of these one hundred models. When several models combine, you get lollapalooza effects; this is when one, two, three or four forces are all operating in the same direction. And, frequently, you don’t get simple addition. It’s often like a critical mass in physics where you get a nuclear explosion if you get to a certain point of mass - and you don’t get anything much worth seeing if you don’t reach the mass. Sometimes the forces just add like ordinary quantities and sometimes they combine on a breakpoint or critical-mass basis”

Good VC’s find this Lollapalooza effect once in a while. Great ones find it consistently.

What does this mean though?

I think when you are starting out in venture, you might make mistakes that go to zero: Wrong founder, wrong market or insufficient traction. After a few years and deals, you start getting to a point where it’s hard, but still possible to get something terribly wrong. But you never stop needing to find that nuclear explosion. That moment of critical mass.

Most deals are additive. Strong team, decent market, early traction: they all combine predictably. Across the five deals I lead, and the 150 odd deals I have visibility over, I’ve noticed that the real Lollapalooza fund-returners are different. In these cases the market was ready, the founder was obsessed, distribution clicked, and the timing was right. Suddenly it’s no longer an additive deal. Instead it’s compounding into something no spreadsheet or model could have anticipated. It’s going critical.

Year 3 taught me to ask a hard question when I look at a deal: Are the forces in play here close to a threshold? Or will they just add up to “just fine”?

2. There is really no participation prize

I have been working my ass off.

In 2026, I have been in a ‘live deal’ situation every single week, and I have had maybe one weekend to myself so far. And for all that ass-offing work I have 0 deals to show in the last few months.

Which means I could have mooched about on a beach somewhere, nonstop since last summer, and made the same difference to my career. Namely: nada.

I’m exaggerating a bit of course. The incessant work does help in terms of building your pattern matching algorithm, improving seen rate, and learning. But realistically, in this business, only deals matter. And even then the great deals are the ones that really move the needle.

There is no prize for trying hard.

Talking of great deals:

3. How hard is it to get to the greats?

Turns out it’s really damn hard. Cole Rotman had an analysis on X in which he said that there were 22 investors globally who had ever led a Series A into a company that later IPO’ed for $5B+.

Even if you assumed he was off by a lot, that number is not more than 50 investors. If he was terribly wrong, it is probably 100. This excites me a lot. There are very few, if any, professions where I can get to the very top with such a binary switch. Where you go very big, or go… very home? You know what I mean.

At the same time, this also tells me something somewhat more sobering: It is very, very hard to tell if a company is going to go Lollapalooza at Series A.

But once you do, once you’ve spotted one great deal, it’s probably a lot easier to do it again. This is illustrated by the small number of investors in Rotman’s analysis: Many of those investors went on to spot several great deals. There are few winners, but they tend to keep winning.

There are things that can make a hard thing even harder. It is even harder if you are not in the Bay Area where the density is very high. It is hardest if you are not in a tier 1 Bay Area fund given the lack of visibility that entails over the fountainhead, if you will, of Tech.

This doesn’t mean it’s impossible, though. In fact I see this as an advantage for me given I can look where no one is looking. The Bay Area has a high density of Lollapalooza companies. It also has a high density of investors hunting for them. The ratio is brutal.

Outside the Bay Area, both are rarer. The ratio is more forgiving. My competition thins faster than the opportunity does. Which means, if you can develop the pattern recognition to spot a nuclear deal in a market that most investors seldom visit, the returns on being right are disproportionate.

Several people have done this very well and systemized it. (Firat Ileri at Hummingbird being a top example.) But I have to hustle harder, take more flights, and be willing to be a lot more flexible than other people. Most importantly, I might have to do deals that break the ‘pattern’ of what mainstream Bay Area investors look for, or think of. This pattern-breaking could be a combination of deal structuring, stage, type or any other variable.

4. Price really doesn’t matter post PMF… in the right markets.

But it will bite you in the ass if you get either PMF or the market wrong.

A couple of years ago I used to feel quite smug after declining companies on the basis of price, and smugger still when discussing them with other GP’s over drinks: ‘Ah yeah I saw that deal, what a ridiculous price it got done at eh?’

Turns out that was really stupid. Some of those companies are doing very well, and folks who led that deal are usually hoarse from laughing their way to the bank.

There were, of course, companies that were priced up and then struggled. This was primarily due to

a) PMF actually not being there and/or

b) market turning or not expanding.

This particular learning isn’t net new, but it’s one I’ve internalised over time.

There is more to price, than price.

5. The quieter your inbox, the harder the job.

It is a lot harder to be an investor in a fund with < 100% inbound.

6. It is better to get into the right company with lower ownership rather than not get into the company at all.

I used to think it was impossible to increase ownership in a company over time. I was wrong. And even if you can’t get a big enough slice, you still managed to get a great outcome. Getting in matters. You are free to disagree, but be ready to pick a fight. I take this lesson very personally.

Let’s play this out. Here’s what happens when you have an opportunity to invest in a great company without the kind of allocation you want:

1. You decline investing because ‘it is not a fund returner at available allocation’: this is actually not a great line of argument, because you are assuming that you can’t increase ownership over time. Congratulations, you now have 0% ownership of a great company. A story you don’t want to brag about, “I saw Series A of X company, and declined”, and no chance of being a great investor

2. You invest, but you cannot increase ownership: You still end up with a fantastic return, and a ring side view + logo creds of a fabulous company. If you invested in Stripe at Series A, that will count towards your funds track record. It will also help you spot what a great company journey looks like, which, as we established above, is a journey privy to only a handful of people. Additionally, with great companies, you often end up with a return that is beyond your wildest dreams. For example: A friend who invested into ElevenLabs’ pre seed didn’t get the allocation he wanted, but did it anyway. It returned his fund and much, much more.

There are a few arguments that people make against this, that, on the surface sound smart, but in reality are not so:

1. Kaushik, what about the option value of investing the same initial check into another company? The deal is the one that’s in front of you. You don’t know if you will ever see a company like that again. Every vintage, there are a handful of companies that matter. The odds of you seeing another great company within that fund vintage tend to 0

2. Kaushik but it is not a fund returner, and my fund has a strategy that every investment must be a fund returner. Ok, this is a legitimate argument. But it is predicated on the fact that there will be more than 1-2 ‘exceptional’ investments (non fund returner), that mess up the fund strategy and take you to a 3x’er upside. So it’s important that when you do this, you don’t do a spread of non fund returner investments, but keep it to true exceptions

7. IRL for conviction, research for confirmation

For most companies, in-person meetings and customer + founder references will create conviction. Desktop research will only confirm it, but is not going to increase conviction. Too much time on desktop research is a waste of time.

8. VC is an apprenticeship business at all levels

However, not everyone is lucky to have a legendary investor as their mentor. Which means that one needs to find a) several of these and b) be able to filter what to learn from whom. Additionally, given that the bar is actually quite binary (see pt 3 above), a great investor might actually just have been lucky once.

9. You can build founder relationships without investing, and this is a good idea

Founder relationships are everything. Even if not a ‘yes’ to invest, investing in the founder relationship itself helps a ton. For example, I couldn’t invest in a founder at seed, and Series A, for a number of reasons. Even though I liked them, and the deal, a lot. Later, I was able to help a growth round into the company from one of our sister funds.

Another founder who had a pleasant experience working with me is now one of my best deal sources.

10. Reputation compounds. Getting your first ‘good deal’ is hard, and then it gets slightly easier

Unless you are at a Tier 1 Bay Area platform, you are going to get shit deals initially. Then you find a decent deal. Then you find a good one. Then you start getting better deal flow. Ironically, this is a vicious and virtuous circle. Just that you have to go through both. Vicious in that you will only get terrible deal flow until you find a good one. But once you find a good one, it keeps getting better. Reputation compounds.

11. Time is a real variable

Time heals all, and time also reveals all. Sounds like Taylor Swift, but it’s true.

As I spend time in this industry, the companies I invested in are maturing. Many of them are raising up rounds. Others are going the other way. Time will eventually tell.

It’s also… kind of scary? Because a deal I do today is only going to ‘count’ in 3-5 years time. (Unless I get lucky.)

12. Fundraising is the highest and most complex form of enterprise sales

I’ve now spent a ton of time with LPs and fundraising. Fundraising is the absolute peak of enterprise sales.

(The Taylor Swift moment has passed.)

I don’t envy LP’s though. They have to pick a manager who, in the extreme, will be managing <1% of their capital. If the managers do exceptionally well, they might return >5x net. If they perform average, they will return 2.5-3x net.

This manager fills a hole in your portfolio. They need to maintain a team that is stable. They also need to have the right LP servicing mechanics, and the capital you give them is going to be a 10 year commitment.

Ultimately, I think the most important thing in this equation is trust. And I’ve come to learn that there are three ways to develop trust with an LP:

1. Consistent communication, whether the news is good or bad

2. Mastering your craft, and making sure they see it

3. And finally: no surprises

I have a lot more to add on the topic of LPs. But I will save it all for next year. 🙂

13. Keeping the operator lens sharp is hard

It has been 3.5 years since I built anything meaningful. AI has happened since then. Am I irrelevant? Is my operator-self redundant? The fear is real.

14. AI has changed everything

You can pass on a company, and in 3 years time it’s the next Lovable or ElevenLabs.

This is the kind of company trajectory that used to take a decade to show.

Also, it used to be said that Europe could not produce winners at American scale or speed. And yet, for every Harvey you have a Legora; for every Sierra you have a Parloa; and you have Lovable, Synthesia, ElevenLabs and several others. AI has played a huge part in levelling what seemed like eternally uneven playing fields.

It’s undeniable that everything we know from the last decade will be shattered by AI

15. Terrible advice. Everywhere you look

Choose whose advice you take. This note is no exception to the rule.

See you next year.

IDFA is what allowed the majority of the internet to be free. It allowed user identifiers to be sent to app publishers, which in turn allowed them to target and attribute better, resulting in strong outcomes for advertisers. Users on the internet, not unlike the physical world, are also different in terms of purchasing power. A high value user in the US, for example, allowed a number of users in emerging markets to access services on the internet ‘for free’, in a way subsidizing them. This worked great, because this allowed everyone to access the internet equally.

Post IDFA, we saw both targeting and attribution rates going down because of the scary ‘Ask app not to track’ modal by Apple. This resulted in dropped earnings for advertising reliant companies such as Meta, and while they have recovered, the reality is that targeting and attribution in a post IDFA world will remain more complex, and more expensive than the pre-IDFA world. Even if you don’t agree with that, it's fair to say that except big tech, others are not going to be able to achieve the same results. This means that now it is going to be harder for 2 people in a flat in Istanbul to create a chart topping game because the casual games funnel (monetise with ads first, introduce purchase later) will not work as well. Similarly, it is going to be harder for 2 entrepreneurs to start a Shopify shop and get it to a meaningful scale without a lot of capital.

IDFA was the ‘link’ between two pools of data, derived from behaviour, in the pre-AI internet.

In the post AI internet, over the next decade, we are going to see user interfaces change. Whether they will be prompts, or something else, I don’t know. What I do know is that AI has enabled human language interaction for the first time, replacing boxes, fields and derivatives from user behaviour on apps. For example, Perplexity or ChatGPT know a lot more about me than Google. Definitely more than Facebook. Why? Because on Google I am searching for ‘Vets near me’. That can signal to Google that I am looking for a vet. However, on Perplexity I search for ‘What does it mean if my dog is scratching himself constantly?’ Perplexity can actually point me towards Apoquel, which is more relevant than vets.

AI also has threaded queries. This means that while Google is a one way question, AI is a two way conversation. Therefore, Perplexity can deterministically identify more about me, than Google can do so.

I use Perplexity/ChatGPT as an example of a scaled AI first consumer app. There will be several more.

AI first consumer apps will be expensive in terms of infra costs. These costs will initially be subsidised by VC money, or by big tech pockets (happening today) but as the hype dies down, the market will ask how these will make money.

I think they will make money from ads. VC's don't like ad tech, primarily because the first image that comes to mind is tacky banner ads. However, there are elegant implementations of ads, such as Rewarded Video in gaming, which is a fair value exchange in terms of user value. So let's call these ways to monetise as monetisation units rather than ads which might be reductionist.

Maybe there is a way to implement these monetisation units in a way that is rewarding to both the user, the advertiser, and the app serving them? The lazy way is unlocked features/more queries/copilot access, but there is perhaps a more elegant implementation? I think that there will be a generational company to build in this space. The company will provide monetisation technology for AI first consumer apps. These will not be consumer apps owned by big tech, but rather everyone else. They will provide highly performing monetisation units, and connect to an advertiser marketplace (now commoditised so no need to build) and specialize in allowing AI first consumer companies to do what they do best (create apps) and take away the pain of monetization.

If this does not happen, then the internet is going to be highly unequal. That is not a great outcome for anyone. If you are also an AI absolutist, then that means that the number of users using AI will be lesser than the number of people using the internet. That, too, is not desirable.

P.S: Several people I’ve spoken to have asked ‘Why won’t users pay $1 to access X’. A majority of the world will not because 1) that will be spread across a number of apps, so it will $1 X number of apps and 2) I’ve lived in places where $1 discretionary is a large sum of money and can assure you that for a large percentage of humanity it continues to be so

I spent a lot of time building ad tech at Facebook - and I think a second coming of the industry is on its way. Maybe just in a different form!

I completed 2 years in VC last month. I’ve led a few deals, seen some uprounds, and navigated crises. I’ve also lost a bunch of deals. Here’s what I learned, written as notes to myself:

1. VC is a sales job and a sales job only. Our job on the investment side of things is to a) get access, b) get information and c) decide whether to invest. VCs will do anything for a and b - creating theses, intros, market maps, hosting dinners, reverse somersaults etc. However, only a few VCs end up in a position to run through the info, decide not to invest, and STILL leave with positive NPS.

   Several are terrible. They schmooze too much in the early stages, and when they say no, they burn bridges. It’s a fine balance to show interest without signaling commitment, a key part of the VC craft for long-term success.

2. Getting up rounds done is HARD, especially if not AI or flavour of the season. Founders will fall prey to VCs doing a) and b) above, think there’s a lot of interest in the company, and not do the dance. Or worse, start ‘picking’ on interest rather than offers, and limit top of the funnel. Being a founder is lonely af with no feedback, so founders mistake VC interest as interest to invest when it is interest to ‘know more’. The moment VCs know more, founders lose all leverage.

3. Damn, losing deals sucks. It’s the worst feeling. I lost a deal with an ex-colleague where I thought I had it ‘in the bag’, lost another because I sat by the sidelines too long, and lost another because of dilution concerns.

   Pay attention to losses that feel long-term bad. Those are the ones you truly believe in.

4. Trust your gut. I’ve met ~600 founders so far (we count). If my gut feeling isn’t developed after meeting so many founders, I shouldn’t be doing this job.

5. Some ‘great’ funds you see, read about, or listen to their partners on Acquired and 20VC are actually not great as a general rule. Some have had bad behaviour with founders, companies, or both. A great fund doesn’t necessarily mean great behaviour.

6. Most boards suck and are ineffective. I will die on this hill. It’s hard to simultaneously a) have a high-quality board b) have them physically and mentally present for 3-5 hours a quarter and c) have a free-flowing, high signal conversation. The board concept probably makes sense for a cobwebby old school company, but not for fast growing early/growth stage tech companies. Founders would be better off with 2-3 directors who actually give a shit and a whiteboard. IN PERSON!

7. It’s HARD to make big money (like actual $) in VC. You can either:

   • Invest at pre-seed, then invest in every round until C. You’re ‘flat-ish’ MOIC wise until Series C AND you are making the same investment decision everytime with negative NPS almost guaranteed if you say no. See below as a typical illustrative example:
     - 2M on 10M post pre seed, capital invested 2M
     - 5M on 30M post seed: pro rata of ~1M
     - 15 on 75 post Series A: pro rata of ~5M
     - 40 on 250 post Series B: pro rata of ~10M
     Total capital at work ~ 18M with a blended MOIC of ~2.75. You can run the math on this whichever way you want, but the meaningful uptick for capital at work only starts happening after this point

   • Wait until the A, and then assume that you will win it. Unless you’re already a heavy hitter, the best way to do this is to pre-empt. This means you will do this on incomplete/limited information, no data room (hey, I’m not raising, the founder says), and limited touch points with the founder. So you’re putting in a 15M-20M check at 75 post with v v high risk. Next round Series B, 40M on 250M with a contribution of 10M you end up with similar MOIC as above

   You are either taking multiple decisions or bundling up the risk in one high-stakes decision with a similar MOIC that you will report to your LPs. That’s hard!

8. Everyone is your friend until you are in contention for the same deal.

9. ‘We increase ownership in our winners’ sounds great mathematically but is nearly impossible to do practically. If a company is a winner, the next investor will push everyone else out, and it is nearly impossible to get super pro rata by the time you know the company is already a winner. Before it is a winner, well, why would you increase ownership? Better to get in with the ownership you want.

10. Just because someone’s invested in multiple winners doesn’t mean they’re smarter. Many VCs have a hindsight thesis and get a lot of shit wrong.

11. More shots on goal IS important. Being selective is good (you have to be), but knowing when to stop analyzing and pull the trigger is important. The more investments you make, the more exponentially likely you are to find winners versus linearly.

12. Network is EVERYTHING. A differentiated network leads to differentiated sourcing. Differentiated sourcing means you see deals before others, so you don’t have to worry about winning as much.

13. References on the founder are more important than the rest of the due diligence put together. Talk to more of the founders past associates than you spend time thinking about TAM. People rarely change behaviours, and more than one instance of bad behaviour is a habit. One founder I almost invested in referenced badly when I dug deep, and thank god I spent time doing that.

14. You can win on a) speed b) price c) brand d) personal reputation. There will always be someone better on c) and d), and different venture funds play different games, so b) is not a reliable strategy for me. It’s always better to be fast, as fast as possible.

15. Different venture funds play different games, so some deals don’t ‘make sense’. Some venture funds are looking to return 3x, others are happy even with a 1x, some are playing a completely different strategy and several do venture just as an experiment.

16. Don’t negotiate against yourself. Lowball offers usually lead to immediate rejection from founders, but also know where to draw the line on overpaying. For a first offer, expensive is fine, but ridiculous is not!

17. Conviction cannot be price dependent. A great company is a great company, irrespective of price. This is for a true venture outcome. Read this in conjunction with the earlier point.

18. Pricing is just what someone is willing to pay. So don’t look for logic in the other person’s price as long as you have your own logic for yours.

19. Don’t second guess yourself. If the founder says they will move mountains, they will move mountains. If the founder says they will make cars fly, they will make cars fly. If you fundamentally don’t believe that, then you lack conviction. I wanted to pull the trigger on a deal where I kept second guessing a bunch of ridiculous things the founder said he’d do - and guess what - he did most of them.

20. The only truth in VC is below:

    

1. Conviction: Writing an angel ticket of a few $k vs leading a round of a few $m is very very different, almost as different as playing street cricket vs one day internationals. Information as a VC at the earliest stages is not *that* much more than what an angel would have, and the best deals close really really quickly (even in a bad macro environment). You need to decide, and you need to decide quickly

2. Capacity: Realistically, I can sit on 8-9 boards at one time and actually show up. So every investment is 1/8 slots gone for the foreseeable future. Its a lot harder to take that decision than the capital investment decision

3. Consensus Risk: VC's see whole of market. Angels do not. A very interesting dynamic I observed is that you see a space heating up (e.g. office of the CFO) and pretty much most VC's independently converge upon the same 1-2 companies. I call this 'consensus risk'. These 1-2 companies raise a lot more capital at higher valuations from better funds vs the others. As an angel/part timer there is no way one sees whole of market, and unfortunately many founders seem to have little idea of this

4. Flight to Quality: These companies subsequently raise more capital/likely to raise more capital, and this results in lemming behaviour and 'hot' deals. As a VC, one way to win is by trying to win the hot deals, a better way in my opinion is to do the work and read the market to get to the hot ones before the others. This is of course easier said than done

5. Pre-seed/seed (first institutional round): is f*cking hard. I have nothing but respect for funds that specialise in this stage. The economics are hard, survival rates low, you need to put in a lot of work, and most importantly you need to really pound the pavement more than anyone else to source

6. Performance: Once you become an investor you start seeing performances of companies and funds. Often the companies you thought were 'great' from the outside do not offer great returns, while several unknown ones have returned capital well. Same goes for funds too!

7. Sourcing: I used to think before becoming a VC, that sourcing after a point is automated/don't need to do/you could get more junior team members to do it. Untrue, I think sourcing is like an SIP contribution, keep doing it so it compounds. Access is everything

8. Luck: There is a lot of luck involved in venture. Maybe there are all seeing fortune teller VCs, but as far as I've read/seen so far many companies were great investments because of a 'lucky' event which could be to do with the business, investment timing, or exit timing. They would have been good investments otherwise

9. Macro: Theoretically I used to think since VC is a long game (10+ years) the macro environment will have minimal effect. However, macro surprisingly affects valuations, ability to raise money and general optimism by a lot

10. TAM: TAM calculations are largely an intellectual exercise if the founder is exceptional. On the other hand if the founder is less than exceptional TAM calculations are the theoretical limit

11. Risk: Anecdotal, but I surprisingly see more deals get done with team risk (questionable team) rather than direction risk (what are you building). I still think it is better to back a strong team with questionable direction

12. Passing: It is easy and lazy to pass. It feels like work, but its actually lazy. Pulling the trigger is hard and is where reputations are built

13. Content expertise: I used to think its important to be deep in a space to make the right investment. However I find that there are several VC's who have little depth in a given space but are great investors. I need to figure out how as my operator default is to go deep

14. DD: Founders don't do nearly as much DD on VC's as much as VC's do so on founders. This is not great. VC’s do several DD calls for example, founders do 1-2

15. Fundraising: is a different skill to business/company building and sadly so. Some founders are great fundraisers, others are great business builders. The best ones are both. I think learning fundraising pays off dividends, especially until series B

16. Sales: A large part of VC is sales. So show up on time, stay present, run a CRM, try to fulfil the promise you made at pitch and don't be an asshole

17. Hive mind: VC's love to meet other VC's. Especially same stage VC's, which is counterintuitive. I don't fully understand why yet

18. Work: as a product person involves tangible, real things. You can do a lot of 'work' as a VC, pull long hours and not move the needle one bit. For example, you could spend a full year going to events, meeting VC’s (see 17 above), meeting some founders, some operators and not do a deal. On the other hand you could theoretically work 2 days in a year, do 2 deals and not do much else (presuming the deals work well)

19. Perception: Every deal looked great at some point and shitty at some point. Not something one should solve for!

20. 'Help': VC's don't build a company, founders do. As a former operator it is quite painful sometimes to 'influence' rather than 'do', but that's the tradeoff of the job!

21. How VC works: Very few founders seem to know how VC actually works - spending 30 mins on this might result in better understanding VC outcomes and behaviours. Well worth it!

I will revisit these and add more to this X thread as I learn more during my career in VC. Hit subscribe, and follow on Twitter.

If you’d like me to elaborate on any of the above points, let me know which one(s) and I’ll try to do so!

• What I am reading now: Flash Boys by Martin Lewis, Clear Thinking by Shane Parrish

• I recently was on this podcast to talk about the shift from operator to investor, and thought you might like it too - check it out and let me know!

I wanted to release 10 issues this year. Regrettably, I failed, but this is my attempt at getting #3 out before the year ends while keeping a high bar.

In my last issue, sent several months ago, I wrote about how to write a user manual for yourself. In this, I referenced a 'manager manual'. This will help set expectations with your manager (without making it awkward). This is, of course, assuming your manager is a good manager: one who cares about your progress and well being.

If you manage teams right now, I'd love to hear from you. Let me know if you'd find it useful if your team members wrote one of these!

Why should you write a manual for your manager?

The manager relationship often offers the greatest upside for learning and growth. Having been a manager and managed by several, I can say that the first 6 months are the most awkward. In this time, a good manager is:

• Understanding your working style

• Understand how to motivate you

• Share context with you

• Develop an understanding of how happy/excited you are to be doing what you’re doing

• Understand your career goals and how to get you there

You can make this infinitely easier for them by stating these preferences upfront. This is not to say that you don’t develop an emotional bond/relationship with them by going for coffees and beers. But, this document will help your relationship get to the parts that matter, and avoid the initial awkwardnesss. Here’s how you write one:

Section 1: Personal Background and working style

Write a section on yourself (the person, not the professional) and include details such as:

• Where did you grow up?

• Why did you study what you studied?

• What are your hobbies?

• What does your family look like?

Or anything that you may be comfortable with. The idea here is to break the ice and show that you’re willing to open up and be informal a bit.

For working style, I recommend taking one of the personality tests. One can argue that these are as accurate as horoscopes. They are, however, directionally right for me. For example, I discovered that I’m an extraverted thinker, ie, I think aloud. I set this context at the beginning and it has helped managers work with me before. For example, my most recent manager stopped asking for async comments. Instead, he started white boarding with me, so that I can showcase my thinking in the best way.

A few tests you can take are here: MBTI (I’m an ENTJ), Truity (has most credible alternatives)

Section 2: Things you’re good at, and things you want to work on

Its always helpful to outline a list of things you’re good at. These could be skills for your job, but also some that are not required for your job but you’re good at. For example, in my case, as a product person I’m good at:

• Influencing x-org

• Prioritisation

• Product Strategy

I am also good at the below things, which aren’t core to be a great product person:

• Bringing a below par performer back to above par

• Managing c-suite archetypes for large enterprise accounts

I’m not top quartile for the below, and have no current plans of improving that:

• Process management

Important here for you to reflect if you want to double down on your strengths or improve weaknesses. I, for one, am firmly in the former camp (many studies have shown this as well). I try to be as aware as possible of my weaknesses and delegate to someone else where possible.

Section 3: How should someone manage you?

Ideally, you want to make the first move in defining how someone should manage you. Questions I would answer are:

• Would you like a more directional or hands off style? I find it helpful to ask for a more directive managerial style in the first 6 months, and then hands off. I pull my manager in where required from that point onwards. From experience, the first few months of managing someone involves tiptoeing around ‘how much is too much’. This helps avoid that!

• How will you stay accountable? Its important that you flag to your manager your cadence and medium of updates. The single most annoying thing I find as a manager is surprises - no one likes them. Its your responsibility to flag how you will communicate updates and when.

• How do you want your manager to utilise 1:1’s? This is key - most managers use these for status updates. I find it helpful to set the agenda right at the beginning - and state that you’ll provide updates in a different forum. I like to use my 1:1’s for unblocking and career coaching - what about you?

There are several other things you can add to this section, but I hope this helps you make a start

Section 4: Why are you here and where do you want to get to?

Its important to call out why you’re at this current job. What are the skills you want to learn? Where do you feel you’ve reached in your career? Equally, it is important to flag to your manager where you want to get to.

When I first started managing teams, I used to assume that everyone wanted career growth. This is not true. For example, someone on my team had recently become a parent. For the immediate future, they wanted to stay at their current level. This was so that they could optimise for work life balance and spend precious time with their child.

Make sure you call out what’s important to you, if you can, prioritise it. For example, for me it is currently 1) Career Growth 2) Family time and 3) Craft improvement. This keeps changing at regular intervals, and I keep my manager posted on these changes.

Last, in this section avoid statements like ‘I want to get from Level X to X+1’ or ‘I want to move from IC to Manager’. Sure, those are good goals but why? You can frame the earlier statements as

• ‘I’d like to more platform work than product work, so that I understand how product attach works’

• ‘I enjoy helping people succeed and so I’d like to do more of that’

Both are inputs, which are great. Outputs are then a function and that’s on your manager to work on - set these expectations without ambiguity.

I hope you found this useful, I’d love for you to refer a friend or post on social media:

Share

My next set of newsletters will be on practical product skills and tools. Thank you for reading, and if you aren’t a subscriber, do consider subscribing using the button below:

Subscribe now

That’s an example of what could go into a user manual. But first, let’s talk about what a user manual is, and why do you need it? So a user manual is a short document, usually about three or four pages that will help people understand how to work best with you. 

Why do you need a user manual?

The initial phases of a work relationship, similar to a personal relationship, are awkward. Everyone is on their best behaviour, and you’re constantly trying to test the limits of what’s possible. Not only is that hard to do, but also adds unnecessary mental stress to an already stressful time - starting a new job or working with a new person.

As you get more senior, you’ll have to meet more and more people, and also have large teams. It takes a while to work comfortably with all of these people - with a user manual you make the other person comfortable, and bat from the front foot. This is more important in a hybrid world where someone might be remote and you’re not going to be able to read non verbal cues. I believe it takes much longer to get comfortable working with a new person remotely vs in person. A usermanual helps remove that awkwardness and takes the guesswork out of a new work relationship.

I wrote my first user manual at McKinsey - the Firm is famously known for its adherence to the Myers Briggs personality type (MBTI) and I thought it might be useful to expand on it so it isn’t as clinical. Now, user manuals are pretty common and I found quite a few of them at Stripe! Let’s get into what it takes to create one:

Section 1 (¼ page): your personal story

I like starting with a personal story so people know where you came from. For example, this is from my user manual:

‘I am from India, but have lived and worked abroad for the last decade in several countries. I am married and live in North London with my wife and miniature dachshund, Churro. I am generally fond of animals, and have a soft corner for anything to do with dogs or horses. If I was not in tech I would be a vet.’

This helps people get to know you as a person, rather than just a CV. It also helps people find common points. The more intersections/opportunities for common points you can provide, the quicker you can break the ice. People are always looking for conversation starters so you want to provide as many of them as possible here.

Section 2: (½ page): your career story

Quickly move to your career. You don’t need to explain your entire CV, but have short paragraphs for each ‘phase’ in your career. For example:

• I started my career at X, and learnt A, B and C. My most significant accomplishment here was to do <example>

• <2nd job> 

• And so on

If possible, add a bit here around what you liked about each job, and what you disliked. This helps people form a picture of your working style

Section 3 (½ page): How to work with you so you’re at your best self

First, take an MBTI (or similar) test. These tests are easily available on the internet and while you may not believe in them, it will equip you with the vocabulary that you may need to describe your working style.

In this section, you should talk about at least the following things:

• How you like to structure your day, especially if you have childcare etc. when you have blocked time out

• Your working preferences (WFH, Hybrid, office etc.)

• How you like to brainstorm/think. Some people (like me) are extraverted thinkers, which means they think as they speak and generally like to brainstorm ideas live. Others like to take the problem statement, ruminate and present a POV later. Which one are you?

• What kind of culture you’re used to - for example, Facebook was highly contextual and collegial. Stripe is written and more cerebral. Your colleagues in your first 6 months can help you settle into the new culture if you flag where you came from

• Your SLA’s on replying to things, and your primary communication channel (Email/Slack/Whatsapp)

Section 4 (½ page): Peer reviewed strengths and weaknesses

In Issue 1, we spoke about running a survey when you left your last organisation. You can use results from that, or use any data you may have collected from your performance reviews. Being open about this does not make you weaker, but in fact makes you stronger as you can be clear on what you want to improve on (and what you don’t). I firmly believe that it is better to double down on strengths than trying to solve all weaknesses and become perfect.

That’s it! These two pages will deliver outsized ROI, and have been instrumental in me settling down into several roles. Colleagues have appreciated the effort and openness, and I’ve been able to work better. A user manual is in addition to a manager manual, which you should make for your manager. What is that you ask? Something for future issues :)

Do you want my user manual to see what a full fledged one looks like? If so, share this on LinkedIn or Twitter, comment by tagging me and I’ll DM it to you on Linkedin. I’m also trying to grow to as many people as possible so I have incentives to continue writing - I’d greatly appreciate a share or forward to anyone who would find this useful!

See you in Issue 3, ‘How to write a manager manual’. If you don’t already, feel free to follow me on Twitter. Thanks for reading!

1. I want to share a lot of my learnings with people who might find it useful. I had a lot of asymmetric information in my early career - and growing up in a 90’s India was a handicap as I realised later. I want to make sure that other people don’t have the same handicap

2. I started a new job a few months ago, which requires me to write extensively and I want to get better at it

Without further ado, here is my first issue:

An algorithm to quit your job with grace

Why is this important? The tech world is super small, and word goes around. Additionally, this is a time of heightened anxiety (quitting + starting a new role) and one is susceptible to making mistakes in the absence of information. There are several articles about how to start a new job and your first 90 days (I think there’s even a book!) but there’s really nothing about your last 30 days. We spend half our lives or more at work, we develop very deep relationships there. And then when we leave, it’s like a breakup - so how do you stay friends with your ex? I’ve made several mistakes and stumbled through quitting three wonderful organisations (L’Oreal, McKinsey, Facebook/Meta) and have been able to hone in on an algorithm that works - I call it the ‘7 final steps’ algorithm. Following this has helped me stay great friends with folks in organisations I’ve left, and my ‘people equity’ remained untouched (with a couple of return offers, several co-investors in companies I’ve invested in, and even a couple that I’ve been able to bring with me to my new company!)

Step 1: Create a script

First, write down the script. You might be quitting because of a whole number of reasons. Maybe you were very happy or not. Irrespective, the golden rule here is that you’re not going to achieve anything out of leaving with a lot of venom. Your script should be some form of:

• I am quitting this job 

• In the last X years, I’ve done X and Y, and learnt A, B and C

• I am thankful for the opportunity to work with you

• However, I felt that I should move on because of D, E, F

• I am going to move to a company in <sector>, and in a role that does <role description>

Here, important to note that you should maintain the principle of least privilege, which means you really don’t need to tell people details of your move until you’ve actually moved:

• Do not reveal your next company’s name. You have nothing to gain from revealing that

• Do not reveal your new team, exact role or anything else such as salary

So something of the order of ‘I’m going to be the head of product at a web3 company’ should be enough. Some people will probe you. Politely decline, by saying ‘I’d rather not reveal where I’m going as I’m still finalising some details’

Step 2: Take what you can

No, I do not mean the snacks in the micro kitchen. Finish up your allowances (e.g. a fitness or commuting allowance), run through any training that you may have and anything that you might need from internal tools. For example, when I left McKinsey, I did all of the training on feedback, board strategy, exec comms that existed in McKinsey, primarily because I knew that nobody does that better than McKinsey. Also take this time to make sure that:

• You understand what your notice period is (and if you need to negotiate it)

• You have no stock vesting dates coming up. If you do, wait until the vest happens

• Remember to take your health insurance details, as they usually run 1-2 months after your job ends. You will however not have login details later so you’ll struggle if you don’t. In addition, account for this when you’re thinking about negotiating your notice period.

Step 3: Create your CRM

In every organization, you’ve built relationships and people equity over time. However, contacting these people outside of Slack is hard (sometimes impossible!). Add your colleagues on LinkedIn, and if their intranet has contact information note them down. 

I use Dex (not a plug) but Excel works well too!

—after this is point of no return, so assume that anything after this will probably be your last day—-

Step 4: Chat with your manager

Do not drop this as a surprise. Ideally they should have an inkling that you’re going to leave the job anyway. Or at least that you’ve been looking. If they are surprised then either you’ve not done a great job or your manager is not a great manager. 

When you’re telling your manager, definitely thank them. Irrespective of how your time working with them was, you probably learnt something. Walk them through your talking points and agree with them on:

• When you’ll send your resignation email

• When and how you’ll tell the team and cascade comms through the org. The more senior you are, your departure might create waves and might affect morale

• When will be your last date (use what you learnt in Point 2)

• What messaging to use for your team/others

This might result in one of three outcomes: 

1. your manager sees you as a threat to the company, in which case you go on garden leave and lose all access immediately. This might happen, for example, if you’re moving from Facebook to Bytedance

2. Your notice period gets shortened, and you get paid for it

3. You serve your entire notice period. In this case, definitely offer help to hire a replacement, especially if you had a good time at your company. It will only result in good things, and your team will be thankful to you

Step 5: Create a handover plan

Unless your access is revoked immediately, create a handover plan. I’d make two versions, one for your manager and the other for the person who will replace you. The first one should be pretty high level, with details of projects, who is responsible etc whereas the latter should ideally be more detailed with a lot more people context (more than anything else)

Step 6: Run a survey

I constantly try to get better at what I do. So when I left Facebook, I ran a survey across peers, managers and my team, in order to understand what I do well and what I don’t. I know this is what performance reviews should do - but there’s so much context there that’s hard to cut through. I found better insights when I ran this survey, the results of which I then included in my user manual and manager manual. What are these manuals, you ask? They’re on the agenda for future issues :)

Step 7: DO NOT do anything on LinkedIn

There is absolutely no reason to make an announcement on LinkedIn. There is also no upside to updating your job title the day you join. I never thought of this (and always have updated immediately) but going forward I won’t - there’s only downside in doing so. If your new job does not work out, you now have the burden of a 1-2 month stint at a company that you have to justify. Sadly, the world is always going to assume it’s your fault (even if it was the company’s). I know folks who’ve joined excellent companies in bad orgs, and have had to quit in 2-3 months because the environment was so toxic. Why do you want to make this public when you don’t need to? 

I hope you found this useful - any and all feedback is welcome (you’d be doing me a favour). I’m also trying to grow and go to as many people as possible so that I can become a better writer and I have more incentives. So if you can forward this newsletter, post on social media, and ask people to subscribe, I will be very grateful. Thanks and see you in Edition 2: ‘How to write a user manual’ - in your inbox in the next couple of weeks!