(Good) ideas are rare.
Why strong ideas still matter, and how to find them
Why strong ideas still matter, and how to find them
Disclaimer: Most of the reasoning here assumes you want to start a company that can scale as much as possible. If you have a bootstrapped business that already generates a healthy amount of cash flow, you’re probably doing better than most VC-backed founders. Aside from ego, it’s worth questioning whether you really want and need to play the VC game.
If you spend time around startups, you've probably heard some version of:
"Ideas are worthless. Execution is everything."
There's truth in it. Ideas are abundant. Everyone has them. But turning an idea into a real company, and then into something that lasts, is brutal. The work is messy, slow, and emotionally expensive.
And still, whenever we look at Uber or Facebook, it's hard not to think:
"That's an amazing idea, how did no one see it?"
Or worse:
"I had that idea years ago."
The uncomfortable reality is that many "good ideas" only look good in hindsight. Execution, iteration, distribution, and market shifts make the outcome feel obvious later. But that "obvious" outcome can take 5, 10, 20+ years to materialize.
So yes: an idea alone, with no execution, does not matter.
But the more interesting question is the one most people avoid:
If you do have a world-class executor behind it, does the idea still not matter?
My answer: sometimes it matters a lot. Sometimes it matters less. It depends on what kind of game you're playing.
A simple way to put it:
Execution is a multiplier. The idea sets the ceiling (and sometimes the slope).
Why I'm Writing This
I’ve been thinking about this obsessively for a personal reason: I try to be brutally self-critical, and one of my consistent weaknesses is coming up with truly strong startup ideas. So I’m writing this as a self-reflection and a form of self-advice.
I’m good at operating inside an existing direction. If the goal is clear, I can execute. But when it’s blank-page time, I’ve often struggled.
And I’ve seen firsthand why this matters: the same founders can look “stuck” for years on an idea that goes nowhere, then pivot and suddenly grow like crazy. In fact, that arc shows up across many iconic startups (Lovable, Cursor, Loom, Slack, etc.).
The team was world-class. They didn’t lack the ability to build or to sell. They were just playing the wrong game: the wrong market, the wrong timing, or the wrong wedge.
The hard part is you don’t know what’s true while you’re inside it. Are you early? Are you wrong? Are you just being an idiot? Sometimes it’s all three.
That’s part of why I like South Park Commons. They’ve been explicit about the importance of what they call “-1 to 0”: the pre-company phase where the primary job is not scaling execution, but finding a direction that avoids a low ceiling, a “local maximum” (source).
What I Mean by a "Good Idea"
When I say "good idea," I don't mean clever. I mean an idea that is structurally strong.
My rough definition:
- It is grounded in a real job-to-be-done.
- It targets underlying pain points that are stable over time.
- It has the potential for customer LTV that can expand dramatically, ideally without a hard cap.
I'm intentionally not centering TAM. TAM matters for venture scale, but people often use TAM as a crutch. It is easy to dress up weak ideas with big numbers.
Also, timing matters a lot. Timing has a real luck component. You can be "right" and still be early in a way that kills you.
Why VCs Say They Bet on Teams
Early-stage investors often say they bet on teams more than ideas. There's logic there.
A common pattern: early memos focus on founder quality, not the initial product. For example, Chamath Palihapitiya shared an investment memo from an early bet in Groq where the emphasis was clearly on Jonathan Ross (source).
Y Combinator is similar in spirit: accept strong teams, trust that they will iterate into the right thing.
That's valid.
But it hides a problem that shows up in the real world:
Most founders converge on the same ideas, because the same ideas look legible, fundable, and safe.
Paul Graham's essay on why smart people end up with bad ideas captures this well (essay).
The Trap: Obvious Ideas Feel Safe
Last month I attended the AI 2026 event in San Francisco. Vinod Khosla said something simple that I keep replaying:
Everyone is talking about AI. The opportunities are massive.
But the more interesting question is: what is nobody looking at that has nothing to do with AI?
You can swap "AI" with any hype cycle: crypto, web3, whatever.
This pairs well with a second mental model I think is underused. Jeff Bezos framed it like this:
Most people ask what's going to change in 10 years. The more important question is: what's not going to change?
The meta-point is not "copy Amazon." It's: if you build around what stays true, you can compound effort without constantly rebuilding the foundation.
What I Learned From 2,000+ Startup Applications
I've founded or supported 3 entrepreneurship programs and reviewed over 2,000 applications. You start seeing the same pattern quickly:
People repeat ideas.
A clear example: after ChatGPT launched and OpenAI, Claude, Grok, etc. opened APIs, there was a boom in customer support startups.
It makes perfect intuitive sense. If basic chatbots were helpful, an LLM chatbot should be much better.
The market is huge. And it is also crowded.
Once a category becomes obvious, it becomes saturated, fragmented, and expensive to win. The cost of entry goes up because distribution becomes the bottleneck.
There's also a second issue: many "obvious" ideas are incremental improvements that quietly assume the medium stays stable. But big tech shifts change the medium.
Ivan Zhao captured this in one line:
"Future is often difficult to predict because it always disguises itself as the past."
Source: Ivan Zhao.
You can absolutely succeed without being methodological. But if you want a repeatable way to avoid obvious ideas, you need some structure.
A Practical Framework for Finding Strong Ideas
I don't love over-optimizing "strategy." Pretty documents do not reflect the reality of being stuck, uncertain, and stressed.
But in hindsight, a framework is useful as a compass. Here’s what, today, I think is helpful.
I. Start with the Job-to-be-Done
If you want stronger startup ideas, don't start by asking "what should I build?"
Start by asking: "what is someone trying to get done?"
The job is the outcome the customer wants, independent of the tool.
Examples:
- I need to get from one place to another.
- I need to eat.
- I need to store files.
- I need to reduce risk.
- I need to look credible.
- I need to save time.
Then add context:
- B2C vs B2B vs B2G.
- Which part of the customer's "Maslow pyramid" it touches.
- Whether it remains true over time.
- How urgent it is (nice-to-have vs must-have).
The Bezos question belongs here: what will not change in 10 years?
And the personal version matters too: can I stay interested long enough to do this for years?
II. How Far Can LTV Extend?
Unless you're in a world of very few buyers and massive contracts, you usually want a model where the same customer pays you repeatedly.
This became popular with SaaS, but it exists everywhere. Loyalty programs in consumer brands are the same logic.
A model I personally dislike (and have operated): cohort-based courses.
Not because education is bad, but because LTV is often structurally capped unless you expand a catalog forever.
Compare:
- "I need to learn data analysis." Finite.
- "I need entertainment." Recurring and open-ended.
III. What Unfair Advantages Do You Have (or Can You Build)?
Unfair advantages exist:
- network
- credibility
- industry experience
- distribution channels
- audience
- access to talent or capital
Some people inherit them. But many can be built deliberately.
Still, if you can choose between:
- a market where you have leverage
- a market where you don't
you should usually pick leverage, especially leverage in distribution.
IV. Do You Need Capital, and Can You Access It?
This is where many "great" ideas die.
True moonshots often require more capital. And most of us do not have the combination of:
- world-class social capital (network + credibility)
- financial runway
- nearby wealthy backers
This is uncomfortable to say because founders should avoid victim mentality. But it is real.
Especially in emerging markets like Latin America, where fundraising from local funds almost automatically makes you uninvestable by Tier-1 funds because the dilution you take is too high.
In fact, I think most startups in Latin America aren’t really playing a VC game. Founders raise $1M seed rounds and then spend years killing themselves trying to play the VC game, when in reality, they never had a good shot.
Venture is simple in a brutal way: you usually need credibility or traction. Ideally both. If you lack one, the other must be strong.
So many founders end up choosing ideas that produce revenue faster, even if the ideas are less ambitious, because revenue buys them time and bargaining power.
That's one reason "AI slop" exists: it's often the shortest path to traction.
Founders Fund has a strong argument around choosing meaningful quests (essay).
V. Given All This, What Can You Actually Start Now?
After you think through:
- JTBD
- LTV potential
- unfair advantage
- capital constraints
you should end with a shortlist that is grounded in reality.
Then you add the part that feels vague but matters: fit.
Not "passion" as a motivational poster, but alignment with how you want to spend your time.
Example:
A niche software factory can be a real business with demand.
But if you don't want to run a consultative business, it doesn't matter how "good" it is.
What If You Find a Match?
Then the answer is boring and timeless:
Execute. Ship. Sell early. Iterate fast.
If the direction is strong, execution compounds.
What If You Don't?
This is where I am right now.
I think the two most rational options are:
- Join a company in a domain you care about, ideally well-funded, and use it to build unfair advantages.
- Start something lower-hanging that generates revenue and buys you time, learning, and optionality for the bigger thing.
Timing, the Variable You Can't Control
No one is a magician about timing.
I've seen founders raise millions from top funds (Lightspeed, Accel, a16z, YC) and still fail when the wave they built on passed.
I've also seen founders struggle for years and then, almost overnight, the market shifts and they look like geniuses.
So the best you can do is:
- anchor on what will not change (Bezos)
- keep challenging the "how" as technology evolves
- keep asking what the customer actually values, not what the founder wants to build
How Did Great Companies Get Built Without All This BS?
Yes, I get it, all of this can sound like bullshit, especially coming from someone who hasn't started and scaled a "startupeable" idea yet. Indeed, lol.
But as I said, this is a reflection based on six years of learnings and a lot of conversations. Still, I've wondered: how did some great companies, with incredible people who executed so well, get started?
I think it was a mix of naivete and a natural obsession with a personal problem.
If you study most big companies, you'll see a pattern: the founders were their first users. They were solving their own problems. As a result, they landed on a job-to-be-done they wanted to complete and, with luck (and obviously a ton of iteration and execution), it turned out many other people shared the same job-to-be-done.
For these cases, there are two things worth noting:
- That is a very small sample vs. the total population. Most founders who build stuff they find cool don't end up doing well.
- The outlier founders probably wouldn't be reading this because, if they were, they'd already be executing.
All of this framework is what you reach for when you want a more deliberate approach. It is not required to win, but it can help you avoid the most common traps.
This anecdote from Mozart reflects what I'm trying to say. The honest thing is that true outliers, like Mozart, are one in millions.
"This 21-year-old kid comes up to Mozart and asks, 'How do I write a symphony?'
Mozart says, 'You're too young.'
The guy goes, 'You were writing symphonies younger than I am now.'
And Mozart's like, 'Yeah, but I wasn't going around asking people how to do it.'"
For the rest of us, the goal is not to manufacture genius. It's to consistently search for ideas with strong structure, then execute hard enough for the structure to actually matter.