Last week I wrote about what I actually look for in a founder. That piece did well, partly because it cleared away a lot of mythology. Charisma is overrated. Confidence is cheap. Vision without execution is just theatre.
This week I want to be explicit about what I deliberately left out.
When I invest, I separate the human from the machine.
The founder matters enormously, but once I am comfortable that the person can think clearly, learn fast, and survive discomfort, I stop zooming in on them. From that point on, I am underwriting the start-up itself. Its structure. Its incentives. Its ability to compound without heroics.
This article is about that second part.
If you are a new angel, this is the bit you are probably underweighting.
If you are a founder, this is the bit investors are quietly judging while you talk about yourself.
I look for a structural reason the company should exist
Most pitch decks start with a problem statement. That is not enough.
I am looking for a structural reason the start-up should exist now, in this form, and at this scale. Something in the market that is misaligned, inefficient, or locked up by legacy incentives.
Good examples tend to have one or more of the following traits:
A large value pool that is real, visible, and currently captured by intermediaries
A market that remembers nothing from one transaction to the next, so value never compounds
A system designed for humans doing manual work that software can now absorb
A regulatory or pricing shift that changes behavior faster than incumbents can adapt
If the company disappears tomorrow, does the market snap back to normal, or does something genuinely break?
If the answer is “people would just go back to doing what they did before,” I lose interest quickly.
I care a lot about how value compounds, not just how revenue appears
Early revenue is nice. It is not the point.
What I look for is whether each action taken on the platform makes the next action cheaper, faster, or more valuable.
Compounding shows up in quiet places:
Data that persists across transactions instead of evaporating
Profiles, accounts, or assets that improve over time rather than reset
Workflows that get shorter and more automated with use
Trust or verification that, once earned, can be reused
A start-up that needs to reacquire, re-educate, and re-convince users every single time is a treadmill, not a business.
I want to see a flywheel, even if it is small and ugly at the start.
I want multiple revenue paths, even if only one is live
Single-line revenue models make me nervous, especially in marketplaces.
I am not looking for complexity. I am looking for optionality.
The strongest early companies usually sit in the middle of a lifecycle, not at the edge of a single event. They touch users before, during, and after a transaction. That creates natural surfaces for revenue that feel earned rather than bolted on.
When I hear:
“We will monetize later.”
What I want to know is:
Where exactly, and why that moment exists at all.
If you cannot point to at least three credible ways money could flow through the system, you are probably underestimating how hard scale will be.
I pay attention to unit economics long before scale economics
This is where new angels often get distracted by the wrong numbers.
I do not need perfect margins on day one. I do need to see a path to them that does not rely on miracles.
I look closely at:
What it actually costs to acquire supply and demand
Whether automation lowers costs over time or just flatlines them
Which parts of the business can reach software-like margins
Where low-margin revenue is strategic versus dangerous
If every dollar of growth permanently drags cost behind it, no amount of storytelling fixes that.
I care deeply about defensibility that emerges naturally
I am sceptical of moats that rely on slogans.
What works is defensibility that emerges as a by-product of use:
Verified data that competitors cannot scrape
Workflows embedded so deeply they are annoying to replace
Network effects that improve outcomes, not just vanity metrics
Local density or trust that takes time to replicate
If a competitor with a large budget could clone the product in six months and be ahead in a year, the start-up is not investable for me at seed.
Timing matters more than brilliance
Some ideas fail because they are bad. Many fail because they are early.
I want to see evidence that behavior is already shifting. Even slightly.
Small numbers matter here. A few thousand users behaving differently than the market used to is often more meaningful than a beautiful TAM slide.
When I see early traction that looks oddly efficient, cheap acquisition, fast response, organic sharing, I lean in. That usually means the market is already pulling, not being pushed.
What this means if you are a founder
If your pitch leans entirely on you, it is fragile.
Strong founders build companies that do not require them to be extraordinary every day. They design systems that work even when they are tired, distracted, or replaced.
If you want investors to trust you, show them a machine that makes sense.
A final note
If you read last week’s article and this one together, you might notice something uncomfortable.
Once I believe in the founder, I stop betting on the founder.
I start betting on the structure.
If this kind of thinking is useful to you, follow me here. I write for new angels and founders who want to understand how capital actually gets deployed, not how it is described in pitch decks.
What resonates for me here is the focus on where control actually sits — cost curve, contribution margin, and pricing power — rather than labels or narratives.
A lot of businesses get mis-underwritten because they borrow leverage from external platforms or assumed future cost declines instead of owning or shaping the underlying system. Once you apply that discipline, many “great stories” stop being investable very quickly.*
This I believe requires building an ecosystem.