Can we invest or not, we won't: a nine-year review of a VC after four cycles
Interviewee: Jocy, IOSG founder
By Joe Zhou, Foresight News
After nine years of investing, we slowly realized one thing: the hardest question to answer is actually not “what kind of founder can succeed”, but rather -- assuming the right track is chosen and there is no problem with the direction, why did some highly qualified founders not survive in the end?
Half of the answer lies with the founder, and the other half on the direction he chose and when to enter the market. After four cycles, certain models began to resurface over and over — although each founder's story was different, every segment of the market context was different.
But one conclusion is becoming increasingly clear: successful founders are all wonderful, while failed founders are surprisingly similar.
In nine years, hundreds of projects have been invested, and we have seen the rise and fall of too many Web3 entrepreneurs. Behind every failed investment is the cost of real money — millions for less, tens of millions for more.
To prevent the same mistakes from happening again, I built a “failed founders database”. The goal is simple: prevent yourself from stepping on the same river continuously.
There is never a “try again” option in the capital market, but we can choose to turn other people's pits into our own signpost. Looking at these failures, we are not only trying to improve our own hit rate, but we also hope to help more entrepreneurs take fewer detours.
6 portraits of failed founders
I have a habit of individually reviewing deals with each colleague every quarter; every half year, the whole team makes deep alignment; at the end of the year, I pull out a list of all the successful and failed projects I have submitted in the past.
The first half of 2026 has just come to an end, and taking advantage of this replay, we have added and summarized a “Portrait of a Failed Founder”. With this database, we hope to be able to turn the pits we've stepped on into muscle memory and avoid those deadly reefs in advance.
After four cycles of trial and error, these failure patterns gradually became clear — although every founder's story is different and every market context is different, the logic behind it is strikingly similar.
Before we expand, though, we need to be clear: the following models fall into two categories. The first category is at the level of the founder's characteristics, relating to a person's emotions, resilience, judgment, and self-perception; the other category is at the level of project structure, relating to structural choices such as token design and capital strategy. The former is about people; the latter is about things.
Founder Trait Categories
This type of problem is rooted in the founder's own personality, mentality, and internal drive. They have nothing to do with technology, nothing to do with the racetrack, but they are often the first to kill the project's culprit.
Type 1: Emotionally unstable
This is the most lethal category. When a project is retracted 80%, the community is being attacked intensively, and there has been no progress for 3 months, how the founder reacts almost determines whether the project can survive.
Failed founders get bogged down in emotions at these times — repeatedly proving themselves on Twitter, breaking out in internal conflicts with co-founders, and connecting with users in community groups. Under the same kind of pressure, the successful founders were already breaking down the problem and working on Plan B in the first week.
Actually, you don't have to wait until the retracement is visible. You can measure it before you invest: politely push back (question) his core assumptions to see how he responds. Some founders will seriously debate with you; they should stick to it, and the amendments will hold steady throughout the process; others will immediately defend or even fight back when they are being respected and questioned. The problem described by the latter was revealed earlier and more reliable than how he performed during the 80% retracement.
Type 2: Lack of Hunger/ Retreat
This category is easily overlooked because it's not “conspicuous.”
If the founder has a soft enough safety net behind him — whether it's family wealth, a high-paying retreat from a big company, or a “it doesn't matter if you can't do it” mentality — his choices in the darkest hours tend to deviate from the optimal solution. Starting a business is a place of death. Without the four words “full commitment”, it is difficult to get through the cycle.
We once discussed a project at IC, and there were huge differences in internal voting that day. It's a team where both Paradigm and a16z are willing to bet. The founder's family background is very good. He is also an LP for several US mega funds (top venture capital institutions over 5 billion US dollars), and these mega funds are also willing to support them. Looking at the investor structure (lineup) alone, this is one of the most beautiful deals we've ever seen.
But that day our dispute about this project continued until 1 a.m. In the end I voted veto (no).
The reason is, I think what the founder wanted to do was too much to challenge humanity. His idea is to go to the African market to become a crypto bank, which requires setting up a local promotion and execution team of hundreds of people. However, the founder himself grew up in the US and China. To do this, he had to actually move to Africa, start a business and live there for a long time, and stay on the front line. At the meeting, he repeatedly emphasized his determination, saying that he would go deep into the local market in Africa irrevocably.
But it was this kind of team that “seemed to get everything right” — with top institutional endorsements, perfect investor structure, and impeccable determination from the founders — we finally hit the pause button. As a result, although this project later succeeded TGE, it was far from their initial vision of becoming an “African neobank.”
This matter and the “execution machine” make sense. A team can get a perfect score in every dimension where it can be scored: organization, structure, history, determination, plan. However, the most critical thing in starting a business is often exactly the one that cannot be graded — is there really a perfect fit between this person and the thing he wants to do. Ticking all of the boxes is the easiest way for people to overlook this problem.
We went back to it many times later: the question was never that he wasn't good enough; it was that he was good at everything, so much so that we almost forgot to ask the only important question -- a person who grew up in the US and China, would, or would like to actually push the next five years of his life to the forefront of Africa?
Type 3: Uncontrolled Ego
This kind of outward expression is often a “well-decorated execution machine” or a “professor-type founder.”
Let's talk about execution machines first. The OKR system is particularly sophisticated, and the deck works like the McKinsey Report, the founder who lists “execution” as the number one advantage — in our data, this type of person has a high amount of financing, but the number of investors they can attract later is small, and their exit performance is poor. Because they are good at providing optimal solutions to known problems, the most common thing in the crypto industry is that the foundation changes. A well-decorated house is much more brittle than a rough house.
But let's add: Whether the execution machine is an issue depends on the racetrack.
In the proven mainstream direction, distribution, recruitment, and repeated execution are the winners and losers; a well-decorated executive founder is probably the best choice. The problem is only in an emerging, non-consensual direction — where you need people who are more imaginative and dare to break through the blurred zone. So this isn't an iron law, but a founder-market fit judgment.
Let's talk about professor-type founders. Their technical understanding is usually the deepest in the audience and is worthy of respect. However, we will be observing in particular on two issues: one is whether they really understand business, and whether they are willing to compromise for the implementation of the business; the other is whether they are coachable, willing to learn, or unwilling to change.
When a professor sees himself as a teacher and VC as a student, the project usually gets stuck in place. Technical depth cannot be equated with product judgment, let alone commercial execution.
We have also invested in founders with a deep technical background and a strong sense of business. The key is not his education, but whether he uses technology as a means, a business implementation for his purpose, and whether he is coachable.
There's also an even more hidden layer: retreat.
Big companies and people with academic backgrounds often have good fallbacks (setbacks and alternatives). Once the project starts to sink, it's easier for them to return to their comfortable factory or academic path. That doesn't mean they're weak founders, but it probably means they're less hungry they have no way back and have to prove themselves. We value that kind of “if you lose, there's really no place to retreat.”
Finally, there is another type of path-dependent type — a direct copy of the style of play by the winner of the previous cycle from a major manufacturer. We call this kind of thing “do this cycle's thing using the method of the previous cycle.” Dai Wuson recently made a similar observation: “It's hard to beat Byte in Byte's rules.” By the same token, the winners of the previous era are the easiest to lose to the next.
Project structure class
This type of question is about how the founders understood the underlying structure of the project — what exactly is a token, how to design a capital strategy, and whether they have experienced the cruelty of the cycle firsthand.
Fourth category: Token priority, not product priority
This is unique to Crypto, and it's also the most dangerous category.
It's different from the previous categories — the problem is not the founder's personality, but his choice of project structure. But this choice itself would in turn reveal what exactly he sees as his true core.
The typical example is that revenue and equity are kept in an independent company entity. Tokens are only used as financing tools, and token holders have no right to claim actual business cash flow.
We believe that whether the token is a financing tool or a product framework determines whether this founder can go through the cycle.
The criteria for judging are simple: if the token returns to zero tomorrow, does this project have any value? If the answer is no, then Token is all he has, and the product is just its packaging.
Type 5: No Day 1 Exit Thesis
This is the principle our team has always emphasized — “Exit before Entry.”
If a founder doesn't even say how he plans to withdraw after 3 years on Day 1 (being acquired, Token Liquidity Withdrawal, or the company's own IPO), then his financing narrative in front of investors will always be distorted.
Rather than having to figure out how to retreat in the future on Day 1, the founder must understand the capital strategy and the ranking of milestones: What will this round of financing prove? What stats will unlock the next round? How will the return path for investors appear in the future? Early projects are often emerging, and it is probably impossible to say whether the final exit method is mergers and acquisitions, token liquidity, or IPOs — but “what is this round for and what to catch up with the next round” must be clearly thought out.
Founders who fail often say, “We are financing for a bigger vision.” Successful founders will say, “I'm merging this round today so I can catch the next round after 18 months. The indicator for the next round is XX.”
The last dimension
The first five categories have a common undertone — they are all red flags (note: Red Flag means danger or warning signal in an investment context).
More specifically:
Founder Trait Red Flags: Emotionally Unstable, Lack of Hunger/ Backward, Uncontrolled Ego
Project structure red flag: Token priority, no clear capital strategy
But the sixth category is not the same. It's not a red flag, it's a pricing issue.
Class 6: Type that has not gone through a complete cycle
Crypto is a full cycle of 3 to 4 years.
A founder has never experienced at least one complete bull or bear firsthand, and he seriously underestimated his vulnerability during his first bear market. It's not a question of ability; it's a matter of experience—if you haven't seen it, you don't know what that kind of stress feels like.
Our one has become a hard sizing policy: early teams without full cycle experience will be limited to an initial investment of less than $250,000.
The criteria are also simple: What are you doing in 2018 and 2022?
However, this category is different from the first five.
The first five categories are red flags. Their role is to help us identify “who should avoid them.” The sixth category is not a red flag; it has to answer another question: “Who can I bet on, and how much?”
Strictly speaking, lack of full cycle experience alone doesn't constitute a veto — it's more of a pricing factor. People who have experienced complete bulls and bears are often better at managing fluctuations, coping with community pressure, and the mentality of downturn; however, there are always exceptional geniuses who have experienced cycles.
Therefore, our approach is not to give up directly, but to hedge with sizing: early teams with no full cycle experience limit their initial investment amount to less than $250,000, and wait until they see evidence of stronger execution before adding.
Reversing the image of failure is someone we like
The purpose of listing failure images is not to label people, but to help us know more clearly for ourselves: in turn, what kind of people are worth betting on.
First category: obsession with the problem.
The best founders aren't interested in a problem; they're engulfed by it. He thought through the boundary situation, user behavior, how competitors would react, and what the second-order consequences would be. He's not pitching (demo) a product with you; he's living with that problem. This is the hardest to disguise and the strongest positive sign in a reference call — you can sense if a person is actually spending 24 hours with what he wants to do.
Type 2: Secondary Entrepreneurship + Non-Consensual Foresight
I particularly value second-time entrepreneurs who have experienced failure.
The failure mentioned here was a setback at the project level and I understood the reason; it was not a fatal flaw at the personality level mentioned earlier. The two are completely different.
Failure is nothing; the key is whether you can figure out where you are after failure.
More importantly, he must have his own non-consensus thesis — not the kind of person who follows Twitter popularity and second-hand information, but someone who really thinks independently and dares to make anti-consensus judgments.
The third category: Good communication + controlled ego
Communication skills are worth mentioning separately because they are so critical. A founder needs to explain complex ideas clearly — to users, investors, partners, employees, and the community. We've seen too many tech genius founders who write beautiful code but can't speak clearly. In the end, the project came to a state of affairs: as soon as there were no people in the team who could communicate with the outside world, the whole project went haywire.
As for ego, it's more subtle than you might think.
What we want is not an easy one
“Low ego”. The benefits of a low ego are coachability and a willingness to listen to feedback; but a founder who wants to be number one, wants to prove himself, and be able to withstand adversity requires a little ego as fuel. What's really dangerous is an uncontrolled ego — rewriting stories when performance is poor, putting yourself on the right side forever, and turning a blind eye to contrary evidence.
So the keyword isn't “low ego,” but “controlled ego”: ambitious, but not delusional.
The fourth category: no evasion, no limits, and strong willpower
The crypto industry has been exposed to the public spotlight and high pressure for many years. If you don't have the willpower to get the bottom of it, you won't be able to carry through the cycle at all, and you'll probably be crushed halfway. We have a core framework within us called the “Key Question”: the essence of early investment is not to stick to a thesis, but to continuously iterate the prior and posterior tests of each key question.
Put simply, it's Bayesian (Bayesian reasoning) thinking — that is, constantly updating one's judgments and beliefs based on existing information (prior probabilities) and new evidence (newly observed data), rather than sticking to an immutable conclusion. You can have strong opinions, but don't be kidnapped by your own opinions — if the reason changes, your judgment will have to change with it.
Category 5: Three Hard Metrics in the AI Era: Global Perspectives, Agencies, and Taste
Since its inception, Crypto has been the most globalized technological ecosystem — capital, talent, and community, flowing around the world in real time. In an increasingly fragmented world, being the founder of a global business from day one is itself a scarce commodity.
Let's look at AI again. It can solve problems within the distribution, but only people can ask original problems outside the distribution. So we're looking at two things: Agency (the ability to take the initiative to break the game) and Taste (aesthetics and judgment). Both are getting more expensive in the AI era.
Only when the founder's creativity and imagination are proven first can AI be his amplifier, not a lifebuoy.
Three life-saving tips for entrepreneurs: the cost of issuing coins far exceeds imagination; tickets are one million dollars
We have an internal habit of reviewing with extreme honesty, or even a bit cruel. You'll ask yourself: Why was this decision made at the time? Where exactly did that fatal mistake come from? If I try again, how will it change?
When doing post-investment management a while ago, we brought together the founders of all the projects that were invested in a meeting and gave three harsh but life-saving suggestions:
First, cash flow is far more important than the story. Projects that can survive this round are definitely not TVL or MAU, but real money and cash flow.
Second, don't issue coins for the sake of issuing coins; tokens are actually a heavy liability.
New tokens were issued on a large scale in the last cycle, and according to our internal statistics, the breakout ratio was over 80%.
Therefore, we recommend the project to be funded: if you can't send it, don't send it; if you can post late, send it later. why? Because of the hidden costs after issuing coins, they are far heavier than most people think.
We have calculated an account: the hidden costs after issuing coins far exceed most people's imagination, including market makers, liquidity, compliance, maintenance of exchange relationships, etc. In this cycle, this is a debt in the millions of dollars. If not even this amount of capital has been secured in the past few years, coins cannot be issued at all.
Third, be afraid of mobility.
Sell when you're best, buy when you're at your worst. The project party's valuation at the time of financing today determines what kind of performance it needs to deliver in the next three years to sustain the next round. If you can't take it, you shouldn't finance this money. Furthermore, when the token has the best liquidity, you should sell it decisively, and when it falls below the issue price, you should buy it back to support your agreement.
In today's market, many people are turning to AI, and many are fleeing Web3. Founders need encouragement, practitioners need support—everyone needs to gather a ray of light. Therefore, we will continue to export research and judgment in this industry and give the most realistic suggestions.
What do we think of the founders? Borrowed three frames from Zhang Yiming
The above criteria were not made out of thin air. We borrowed a lot of external references, of which I personally was deeply influenced by Zhang Yiming.
He had an analogy I always remember: empathy is the foundation, logic and tools are middle level, imagination is the sky. Corresponds to investments:
Foundation — Empathy: Can we treat people as ends rather than tools? Can you get along well with the team, attract top co-founders, and show real leadership? This is what we call “emotional stability, low neuropathy.”
Middle level - logic and tools: Can you make good use of tools and think structurally?
Sky — imagination: Can you see things that “may exist but haven't appeared yet”?
One of Zhang Yiming's favorite questions to ask during an interview: “What important things do you think differently from most people?” This question is used to test whether the other person is a person with independent thinking habits, not just a “mainstream media repeater.”
More than half of them couldn't answer. We also often use this question when making reference calls — if a founder can't say three things that are different from what he agreed on, he probably won't be able to come up with a non-consensus thesis.
In addition to this, Zhang Yiming also values two points. We are also learning from it: First, a strong sense of curiosity and hunger — willing to spend time imagining things that “may exist but don't yet exist”, rather than staying in proven fields to make minor improvements. The second is the ability to have a long thought chain — the ability to deduce a problem to the end on its own without external feedback. Corresponding to the crypto industry, is whether you can thoroughly think through a thesis on your own in 18 months without user feedback.
Write at the end
What we have summed up over nine years is not how to find the best founders, but how to ignore it.
But at the end of the day, these methodologies are just tools. We have an iron law within us: even the founding partner can't let the project go through.
If you can vote or not, you won't.
It sounds simple, but that's the whole secret to getting through the cycle.
Crypto's foundation is renewed every three years, and what allows you to get through the cycle is not judging once or twice, but whether you can press the “no throw” button over and over again.
Twitter:https://twitter.com/BitpushNewsCN
Compare the TG exchange group:https://t.me/BitPushCommunity
Compare TG subscriptions:https://t.me/bitpush
