How to evaluate an early-stage company before you invest

How to evaluate an early-stage company before you invest

An early-stage investment comes down to three questions. Can this team build something many customers will pay for? Is the market big enough to matter? Do the price and terms give you a fair share of the outcome? This guide shows angels and early-stage funds how to test each, with a scoring table and red flags.

Risk warning. Early-stage investments are high risk and illiquid. Most young companies fail, shares usually cannot be sold for years, and you can lose all the money you put in. Invest only what you can afford to lose, and spread it across several companies.

What the evidence says about returns and evaluation

Angel returns are highly skewed. In the largest US study of angels in groups, Wiltbank and Boeker (2007), supported by the Kauffman Foundation and the Angel Capital Education Foundation, 52% of 1,137 exits returned less than the capital invested, while the 7% that returned more than ten times produced 75% of all cash returned. A UK survey of 158 angels by NESTA and the British Business Angels Association (2009) found the same shape: 56% of exits failed to return capital, and the 9% returning more than ten times produced nearly 80% of the positive cash flows.

So judge each company on whether it could be one of the few large outcomes, not just whether it will survive, and build a portfolio.

How you evaluate also seems to matter:

• Due diligence time. In the Kauffman data the median angel spent 20 hours on due diligence; exits where investors spent more than the median returned 5.9 times the money overall, against 1.1 times for the rest. NESTA found that even around 20 hours of due diligence was linked with fewer failures.
• Industry expertise. Multiples were twice as high when angels invested in an industry they knew (Kauffman), and NESTA found such investments failed significantly less often.
• The team. In a survey of 885 venture capitalists (Gompers, Gornall, Kaplan and Strebulaev, NBER, 2016), 95% of firms named the management team as an important selection factor and 47% the most important, ahead of business model, product and market. The average firm spent 118 hours on due diligence and called 10 references per deal.

These are associations, not proof of cause, but they point one way: invest where you know the sector, and check the claims. The Angel Capital Association's due-diligence guidance (2007) adds that risks compound: if seven things must go right and each has a 90% chance, all seven succeed only about 48% of the time.

1. The team

Founder–market fit

Ask why these people, for this problem, now. Strong fit shows as direct experience of the problem, access to customers or data outsiders lack, and an insight you could not have read in a report.

A track record helps but guarantees nothing. Gompers, Kovner, Lerner and Scharfstein (NBER, 2006) estimated that venture-backed founders whose previous company went public had a 30% chance of success next time, against 18% for first-time founders and 20% for founders whose previous company failed (success meaning an IPO or a filing for one).

Completeness

Map the skills needed over the period this round funds (building, selling, operations, finance) against the people who have them. Gaps are normal if the founders know them and have a plan. Solo founding is increasingly common: Carta's Founder Ownership Report 2026 found about 36% of startups founded on its platform in 2025 had a solo founder, though two-person teams were the most common among companies that closed a round. Treat a solo founder as a question about coverage and key-person risk, not an automatic no. Check the equity split, founder vesting and whether every founder is full-time.

References and background

• Take the references offered, and find a few yourself through your network, respecting the founder's confidentiality. Ask each the same questions: strengths, struggles, handling of bad news, would they work with the founder again?
• Check the public record. Companies House gives free access to UK officers, filed documents, charges and insolvency information, including the founders' previous companies. In the US, a company selling shares under Regulation D must file a Form D notice with the SEC within 15 days after the first sale.
• The ACA guidance recommends a legal background check for lawsuits, tax liabilities and convictions.

2. The market: size it bottom-up, then ask "why now"

Top-down figures ("1% of a £40 billion market") tell you little. Build the number from the bottom: who exactly the first customer is, how many exist (from a source you can check), what each pays based on prices actually charged, and what share the company could realistically win in five to seven years.

Timing matters as much as size. Ask what has changed (a regulation, a cost, a behaviour, a platform) to make this possible now, and what happens if a larger company notices. In the Gompers survey, 18% of venture capitalists rated timing and luck together as the most important factor in their successes.

3. Product and technology

• See it work. Use the product yourself or watch a customer use it, not a scripted demo.
• Evidence of pain. The ACA guidance warns against products customers merely ought to like: check they feel the problem and have budget. It suggests two established customers confirming they buy, or will buy, as a reasonable hurdle.
• IP ownership. Confirm that code, designs and patents belong to the company, not a founder, former employer, university or contractor, with signed assignments. If patents matter, check filing dates against any public disclosure.
• Dependencies. Reliance on one supplier or data source carries that supplier's risk; our supplier due-diligence checklists help.

4. Traction and the quality of revenue

Early revenue is strong evidence only if you know exactly what it is. Ask for monthly figures behind every metric and a written definition of each. The SEC's 2020 guidance on key performance indicators, written for listed companies, sets a useful standard: a clear definition and calculation, why the metric is useful, how management uses it, and disclosure when the calculation changes.

Metric	Definition
MRR / ARR	Monthly recurring revenue from active, contracted subscriptions; annual recurring revenue is MRR × 12. Excludes one-off fees, services and unpaid pilots
Gross margin	(Revenue − direct cost of delivering it) ÷ revenue
Gross revenue retention (GRR)	Recurring revenue kept from a group of customers after a year, net of cancellations and downgrades, ignoring upsells; never above 100%
Net revenue retention (NRR)	The same, including upsells and price rises; above 100% means existing customers grow
Customer concentration	Share of revenue from the largest customer and the top five
Customer acquisition cost (CAC)	Sales and marketing spend in a period ÷ new customers won
CAC payback	CAC ÷ (monthly revenue per new customer × gross margin), in months
Lifetime value (LTV)	Monthly gross profit per customer ÷ monthly churn rate; a guess until there are years of data
Burn multiple	Net cash burned ÷ net new ARR over the same period

For example (an illustration): customers paying £100,000 of ARR a year ago now pay £110,000, after £15,000 of upsells and £5,000 of cancellations and downgrades. NRR is 110%; GRR is 95%.

Then test the numbers. Ask for a cohort table (revenue grouped by the month customers started), which shows retention better than any single figure. Call customers yourself, including one who left. Reconcile the deck to management accounts or bank statements. Pilots, letters of intent and pipeline are signals, not revenue.

Before revenue, look for substitutes: paid pilots, deposits, usage, regulatory milestones. In the Kauffman study 45% of companies had no revenue when angels invested.

5. Competition and defensibility

Every company has competitors, even if it is a spreadsheet or doing nothing; "no competitors" usually means nobody looked. Ask who solves the problem today and why customers would switch, what stops a larger company copying the product, and what gets stronger as the company grows: proprietary data, network effects, switching costs, a licence, patents, distribution. At seed, defensibility is often thin; ask what would make it real within the period this round funds.

6. Business model

• Who pays, and how: subscription, usage, transaction fees, licences, hardware plus service.
• Sales motion against price: a £5,000-a-year product cannot carry a long enterprise sales cycle.
• Margin at scale: do costs fall with volume, or rise in step with revenue?
• Capital intensity: how many more rounds before the company funds itself? Each one dilutes you. Does it need a licence or approval first?

7. The round itself

Use of funds and runway

Ask what milestone this money buys and whether it makes the next round possible. Runway (cash ÷ monthly net burn, planned hires included) must leave time to raise again. In the UK, the median gap between rounds for seed-stage companies rose to 14.4 months in 2025 from 12.4 months in 2024, according to the British Business Bank's Small Business Equity Tracker 2026 (Beauhurst data). In the US, the PitchBook-NVCA Venture Monitor (Q2 2026) reports that time between rounds has compressed, especially for AI companies, and that companies which raised at elevated prices need fast growth or risk a down round.

Valuation relative to stage

• UK: the median seed-stage pre-money valuation was £3.2 million in 2025 and the median seed deal £0.6 million, both records, according to the British Business Bank tracker (which uses Beauhurst's stage definitions, not round names).
• US: the PitchBook-NVCA Venture Monitor (Q2 2026) reports median pre-money valuations above their 2021 highs at every series, more than double 2021 levels at pre-seed and seed.
• Sector: Carta's State of Private Markets Q1 2026 put the median Series A valuation for a foundation-model AI startup at about $300 million, against $55 million for a non-AI startup. Compare like with like.
• Headlines: PitchBook-NVCA notes that a headline post-money valuation reflects the top price in a round, not what every investor paid.

Then work back from the exit. The Gompers survey found few venture capitalists use discounted cash flow; by far the commonest measure is the multiple of invested capital. Ask what exit, after future dilution, returns your target multiple, and how often companies in the sector sell for that much.

Terms

• Instrument. Many early rounds use a SAFE or convertible note that converts at a later priced round. In Y Combinator's post-money SAFE, the valuation cap is the highest valuation at which the SAFE converts, and ownership equals the investment divided by the post-money cap; a discount instead gives a lower price than the next round pays. With neither, you have no price protection.
• Priced rounds. Read the liquidation preference, anti-dilution, pre-emption, board and consent rights, founder vesting and leaver terms.
• Standard documents. The NVCA model legal documents are the usual US starting point. UK Private Capital (formerly the BVCA) publishes UK model documents drafted for Series A and states they are not appropriate for seed rounds, so seed terms vary more and need reading line by line.
• UK tax reliefs. If you rely on SEIS or EIS relief, ask whether the company has HMRC advance assurance. HMRC says it does not confirm that an individual investor qualifies and is not an endorsement or an indication of performance. See our SEIS and EIS listings.
• Missing terms. PitchBook-NVCA observes that undisclosed terms often coincide with structured, down or flat rounds.

Cap table health

Ask for a fully diluted cap table and a pro forma showing how every SAFE and note converts. Check:

• Founder ownership. Carta's Founder Ownership Report 2026 found the median founding team held about 56% of fully diluted equity by seed and 36% by Series A (rounds from 2021 to 2025).
• Stacked instruments. Carta's State of Pre-Seed Q2 2026 notes that US pre-seed deals above $2.5 million typically stack ten or more instruments. Many differently capped SAFEs can dilute more than the headline suggests.
• Dead equity held by departed co-founders or advisers, an option pool too small for planned hires, and equity promised but not documented.

Equity management software makes these checks easier.

Worked example (illustration only; invented figures)

• A company raises £750,000 at a £3 million pre-money valuation (£3.75 million post-money). You invest £25,000, buying £25,000 ÷ £3,750,000 = 0.67%.
• Average net burn of £50,000 a month gives 15 months of runway. If the next raise takes about six months, the milestone must be reached within about nine.
• At Series A it raises £4 million at £12 million pre-money (£16 million post). New investors own 25%, so your stake becomes 0.67% × 0.75 = 0.50%.
• Later rounds dilute you by a further 30%: 0.50% × 0.70 = 0.35%.
• A £60 million sale returns 0.35% × £60 million = £210,000, 8.4 times your money, before later investors' liquidation preferences, which are paid first.
• Ten times (£250,000) needs a sale of about £71 million. If companies in this sector rarely sell for that, the price is too high for the risk, however good the team.

8. The founders' own reporting

How founders report now previews how they will report to you. Ask for recent investor updates. Good reporting arrives on a regular cycle; shows cash, burn, runway and the same KPIs with unchanged definitions; reports misses as plainly as wins; and makes specific requests.

Check your rights. The NVCA model Investors' Rights Agreement (October 2025) provides for annual financial statements within 180 days of the year end, quarterly statements and a quarterly capitalisation statement within 45 days, optional monthly statements within 30 days, and the approved annual budget, but only for "Major Investors" above a threshold set in the deal. A small cheque may carry none, so ask for a side letter or follow a lead who shares reports.

Scoring framework and checklist

Score each area 1 to 5 on written evidence, not impressions. "Gate" rows are pass or fail: a fail ends the process. Weight the scored rows to suit your approach; the NBER survey suggests early-stage investors weight the team most.

Area	Evidence to ask for	Type
Integrity and background	References, Companies House or SEC filings, background checks	Gate
Founder–market fit	Career history, customer access, specific insight	Score 1–5
Team completeness	Hiring plan, advisers, equity split, vesting	Score 1–5
Market size and timing	Bottom-up customer counts and prices; what changed	Score 1–5
Product and technology	Your own use, customer calls, technical review	Score 1–5
IP ownership	Assignments, patent filings, licences	Gate
Traction and revenue quality	Monthly data, cohort table, reconciliation to accounts	Score 1–5
Unit economics	Gross margin, retention, CAC payback, burn multiple	Score 1–5
Competition and defensibility	Competitor map, reasons customers switch	Score 1–5
Business model	Pricing, sales motion, cost structure	Score 1–5
Use of funds and runway	Budget and milestone timeline	Score 1–5
Valuation and terms	Term sheet or SAFE, comparable rounds, exit maths	Score 1–5
Cap table	Fully diluted cap table, conversion pro forma	Gate
Reporting	Past investor updates, information rights	Score 1–5

Then write down the three things that must be true for the investment to work, and how you will know within a year.

Red flags

None is automatically fatal, but each needs a good answer:

• Founders who will not let you speak to customers or former colleagues.
• Metric definitions that change between decks, or ARR that includes pilots, one-off fees or unsigned contracts.
• One customer providing most of the revenue without a long contract, or late filings at Companies House.
• "No competitors", or a market sized only from the top down.
• IP held by a founder, former employer or university without a signed assignment or licence.
• A departed co-founder with a large stake, or SAFEs and equity promises that surface late.
• Runway shorter than the time needed to reach the milestone and raise again.
• A valuation justified only by headline rounds in a hotter sector, or terms the founders cannot explain.
• Pressure to decide quickly, an "exclusive" offer or an unsolicited approach, which are among the FCA's warning signs of investment scams.

Where to find companies and what to read next

• Browse raises. On TrustList's Companies raising board, founders publish what they are raising and at what stage, anonymously until they agree to an introduction. Investors who claim their firm's listing see amounts and use of funds and can request introductions. Traction figures are the founder's own unless marked as checked.
• Diligence. Finding early-stage deals and doing due diligence on them covers sourcing; Angel investing explained covers portfolio size and tax reliefs.
• What founders prepare. Our pitch-deck and data-room readiness checklists show what a prepared company should hand you; Funding rounds explained covers round structure.
• Invest with others. Groups pool expertise. See angel networks, the UK Business Angels Association and the Angel Capital Association.
• Other markets. The principles travel, but company law and tax incentives differ: see our guides for India, the UAE, Saudi Arabia and Pakistan.

About this guide

General information only, not financial, legal or tax advice. TrustList does not arrange or advise on investments and takes no part in any round. Sources checked on 26 September 2026; each figure applies to the period stated.