How to Evaluate a Startup Before the Metrics Exist

Brian Nichols is the co-founder of Angel Squad, a community where you’ll learn how to angel invest and get a chance to invest as little as $1k into Hustle Fund’s top performing early-stage startups.

Pre-seed startups rarely give you neat revenue, retention, or customer acquisition data. You still have evidence to evaluate. It appears in the founder’s customer knowledge, experiments, decisions, and rate of learning. A disciplined first pass separates those signals from a polished pitch, then identifies what deserves deeper diligence.

Start with a stage-appropriate evidence chain

At pre-seed, almost every slide is a claim. The deck may claim that a problem is painful, a market is huge, or a team has an edge. Your job is to trace each important claim through four steps:

1. Claim: What does the founder believe?
2. Evidence: What have they observed or produced?
3. Test: What did they do to challenge the belief?
4. Result: What changed in the product, market view, or plan?

Imagine a founder who says hospital finance teams lose hours reconciling vendor invoices. Eight interviews showing the same broken workflow are useful evidence. A manual pilot with two finance teams is stronger. A paid pilot that gets renewed is stronger again.

This chain lets you evaluate progress without pretending a three-month-old company should look like a Series A business. It also stops charisma from filling gaps in the record.

A survey of 885 institutional VCs found that investors considered the management team more important than product, technology, and other business characteristics. We agree that the team matters enormously at pre-seed. We still want to see that team interact with reality. Founder confidence is a trait. Changed decisions, shipped work, and customer proof are evidence.

Evaluation only reduces avoidable uncertainty. It cannot make a private startup investment safe. The SEC notes that early-stage private placements may involve limited information, illiquidity, and a high risk of loss.

Startup evaluation and startup valuation are different decisions

Evaluating a startup asks whether the company could become valuable and whether the team can make progress with the resources available. Valuation asks what price investors are being offered today.

Keep those decisions separate. A strong startup can be a poor investment at an unreasonable price. An attractive price cannot rescue a weak business. At pre-seed, projected cash flows are especially fragile, so assess the underlying company first. Then judge the round, ownership, and deal fit on their own merits.

Six questions for evaluating a pre-seed startup

The same six questions work across software, consumer, marketplace, and many other early-stage businesses. The proof you accept should change with the company’s stage and business model.

1. Why is this team equipped to solve the problem?

Founder-market fit can come from industry experience, lived experience, technical ability, customer access, or a combination of them. Ask:

• What did you see that outsiders usually miss?
• Which part of the problem have you experienced directly?
• What have you built, sold, or learned that reduces the execution risk?
• How do the co-founders divide decisions, and where have they disagreed?
• What belief about the company has changed in the past month?

Look past pedigree. A Census-based study of high-growth founders found that the mean founder age among the fastest-growing 1 in 1,000 new ventures was 45, and experience in the startup’s specific industry predicted much higher success rates. The signal is relevant capability, not a familiar school, employer, or founder stereotype.

Our co-founder and GP Eric Bahn describes hustle as “great execution meets high velocity.” In a pitch, that looks like a founder who can show what they shipped, what they learned, and how quickly one led to the other.

Strong signal: The team names a specific insight, shows work already done, and explains a decision that changed because of new evidence.

Weak signal: The answer leans on credentials, passion, or a personal story without showing why this team can reach customers and execute.

2. Who is the customer, exactly?

“Women ages 20 to 45” is a census checkbox. It does not tell you who feels the problem, what they do today, or why they would pay.

A useful customer description sounds more like this: “A commercial banker who earns six figures, eats lunch at her desk, uses these three financial apps, and currently handles this task with a spreadsheet and two email threads.” You should be able to picture her workday. The founder should know the customer well enough to play her in a musical. Seriously.

Ask the founder to walk through the last five customer conversations:

• What triggered the problem?
• How does the customer solve it now?
• How often does it happen, and what does it cost in time, money, or risk?
• Who uses the product, who pays, and who can block the purchase?
• Which interview contradicted the founder’s original view?

Specificity is useful because it exposes whether the founder has done the work. A vague persona can hide a vague problem. A precise workflow gives you something to test in customer calls.

3. What proof exists before revenue?

No revenue does not mean no traction. It means the proof appears in earlier forms. Read the evidence in roughly this order, while accounting for the business model:

• Paid proof: pre-orders, deposits, paid pilots, or paid contracts.
• Behavioral proof: repeated product use, referrals, completed workflows, or customers contributing meaningful time or data.
• Product proof: a functioning prototype, a successful technical test, or a manual service that delivers the proposed outcome.
• Discovery proof: detailed interview notes, a consistent problem pattern, or a waitlist with a clear source and target persona.

Treat nonbinding letters of intent as interest, several rungs below cash or repeated use. Treat raw waitlist size the same way. A list of 5,000 untargeted email addresses can carry less signal than five target customers returning to an ugly prototype every week.

Match proof to the company. A deep-tech team may need a lab result or engineering milestone before it can sell. A marketplace should show how it will attract both sides and where it can begin with concentrated liquidity. A business-to-business software company can often run a manual pilot long before the full product exists.

Then ask what the founder will test next. Early proof matters most when it feeds the next decision.

4. How will you find the first 10 customers?

This question tests distribution thinking before scaled customer acquisition metrics exist.

A weak answer names a menu of channels: paid ads, partnerships, content, and referrals. A stronger answer names one reachable group, one action, and one result: “I gave 10 referral codes to commuters at this train station, three signed up, and two completed onboarding.”

An even stronger founder is already running small experiments. They might say, “We spent $20 on direct mail, drove 40 people to the landing page, and eight joined the waitlist. Next we are calling those eight to learn which message resonated.”

Those economics probably will not hold at scale. That is fine. The experiment shows initiative, measurement, and a tighter next step.

Our co-founder and GP Elizabeth Yin puts it bluntly: “The best founders I’ve ever backed have all been incredibly metrics-driven.” At pre-seed, metrics-driven founders may have tiny numbers. What matters is whether they know the numbers, understand what caused them, and use them to choose the next test.

5. Can a narrow wedge become a large company?

A huge total addressable market slide tells you little by itself. Start with the wedge: one customer, one painful workflow, and one reason this company can win.

Ask:

• How many target customers fit that exact profile?
• What do they spend on the current solution or lose to the current problem?
• Which change in technology, regulation, cost, or behavior makes the timing favorable?
• What alternative will the startup displace first?
• If the wedge works, which adjacent customer or workflow comes next?

Competition is evidence that buyers spend money on the problem. It also raises the bar for differentiation and distribution. A credible founder can explain where incumbents are strong, why customers still struggle, and what narrow opening the startup can exploit.

Bottom-up market reasoning is more useful than a percentage of a giant industry total. The founder should be able to connect reachable customers, plausible pricing, and an expansion path. You can challenge each assumption separately.

6. Do the business model and round make sense together?

You do not need a perfect spreadsheet. You do need a basic picture of how money moves through the company.

Ask who pays, what they pay for, which direct costs come with each sale, how the company expects to acquire customers, and how long cash must support the plan. For a pre-revenue company, these answers will contain assumptions. Strong founders label them as assumptions and point to the experiment that will replace each one with data.

Then connect the raise to a milestone. “We are raising $1 million” is incomplete. A useful answer explains what the capital buys, which major risk it should reduce, and what evidence the company expects to have before it needs more money.

Finally, assess deal fit separately. Consider the proposed valuation, round structure, existing ownership, capital intensity, and likely need for future funding. A company can pass the business test and still fall outside your thesis, check size, risk tolerance, or price discipline.

Score evidence strength, not pitch quality

Numeric scorecards can create fake precision. A 7.4 out of 10 founder does not exist. Use a short decision memo that forces you to name the evidence and the risk:

• Thesis: Why could this become a meaningful company?
• Strongest evidence: Which observed fact most supports that view?
• Weakest assumption: Which belief has the least support?
• Likely failure mode: What could kill the company first?
• Disconfirming evidence: What fact would make you pass?
• Next milestone: What should become true over the next 6 to 12 months?
• Deal fit: Why does this round fit your investing strategy?

Finish with one of three outcomes:

1. Advance to diligence: The core claim has credible evidence, and the biggest open risk can be investigated.
2. Wait for proof: The idea remains interesting, but a specific milestone needs to happen first.
3. Pass: The evidence contradicts the thesis, trust has broken down, or the opportunity does not fit your strategy.

This shared language is useful inside an investing group. In our Angel Squad community, members can compare how they weighted the same founder answers and deal evidence instead of trading vague reactions to a pitch.

Red flags that survive a great pitch

Some weaknesses become clearer when you apply the evidence chain:

• The customer stays broad after follow-up questions.
• Five-year projections are presented as proof of current demand.
• The founder lists many acquisition channels and has tested none.
• Customer counts, revenue, or dates change between the deck and conversation.
• The team cannot name a belief that changed after customer feedback.
• Co-founders give conflicting answers about roles, ownership, or priorities.
• Competitors are dismissed as stupid, irrelevant, or nonexistent.
• Questions about burn, runway, or the use of funds produce defensiveness instead of a clear answer.
• Pressure to move immediately is used to shut down reasonable diligence.

One weak area does not automatically end the conversation. Dishonesty does. For the rest, decide whether the gap is learnable, whether the founder sees it, and whether the current round gives the team enough time to address it.

Raise the evidence bar as the startup matures

The framework stays consistent. The acceptable proof changes.

At pre-seed, you may rely on customer interviews, prototype use, pilot conversion, and the speed of product iteration. At seed, ask for retention, sales-cycle, pricing, gross-margin, and acquisition data where the business model supports it. By Series A, cohort behavior, repeatable growth, unit economics, and burn efficiency should carry far more weight than a founder’s story.

This prevents two common mistakes. The first is rejecting a very early company because it lacks mature metrics. The second is accepting anecdotes from a company old enough to have real operating data.

What comes after the first call

The six questions qualify a startup for deeper work. Before investing, corroborate the important claims through customer conversations, product review, founder and employee references, financial records, ownership records, and the actual financing materials. Our pre-seed due diligence framework goes deeper on that process.

Write the memo before social proof and round momentum change your memory of the pitch. Then compare what you believed with what the company later did. That feedback loop is how an investing framework becomes an actual investment thesis.

If you want to practice this process on curated early-stage deal flow and learn alongside other operators and investors, apply to Angel Squad.