Evaluating an early-stage startup is the discipline of judging a company's potential when there's almost nothing to measure yet. At seed and pre-seed there's no revenue history worth modelling, the product is half-built, and the plan on slide eleven will be wrong within a quarter. So a framework isn't a spreadsheet. It's a consistent set of questions you put to every deal, in the same order, so you're comparing companies on like terms rather than on how good the founder was in the room.
The version below is the one we keep coming back to: six lenses, weighted roughly in the order they matter at the earliest stage. It won't tell you whether to write the cheque - nothing should - but it stops you skipping the question that later turns out to have mattered most.
A framework doesn't make the decision. It stops you skipping the question you'll wish you'd asked.
Why use a framework at all?
Because pattern-matching on its own quietly betrays you. The deals that feel most exciting in a pitch - charismatic founder, big market, slick deck - are exactly the ones where it's easiest to skip the unglamorous checks. A framework forces the same diligence on the deal you love as on the one you're lukewarm about. Over a portfolio, that consistency is worth more than any single brilliant call. It also makes you honest after the fact: write down why you invested, and two years later you can learn from where your reasoning held and where it didn't. Memory alone won't give you that.
First lens: the team
At pre-seed and seed, the team is the asset. Everything else is a hypothesis they'll revise. So the real question isn't "is this a good idea" - it's whether these particular people can learn faster than they burn through the money.
Look for founder-market fit: some specific reason this team is unusually equipped to win this market - hard-won domain knowledge, a technical edge, a distribution channel competitors can't easily copy. Probe how they make decisions. How do they handle a question they don't know the answer to? Founders who say "we don't know yet, here's how we'll find out" tend to outlast the ones who have a confident answer for everything. And take references seriously, including back-channel ones. How someone treated a co-founder they fell out with tells you more than any rehearsed pitch.
Second lens: the market
A great team in a market that can't get big enough still can't return a fund. So you're testing two things: is the market large enough to matter, and is something changing that makes now the moment?
Be wary of the top-down "1% of a £50bn market" slide - it's almost always reverse-engineered to look fundable. Better to build it bottom-up: who exactly is the customer, how many are there, what would they realistically pay, how often. The more interesting question is the "why now" - a regulatory shift, a new platform, a change in buyer behaviour. Markets that have sat still for a decade rarely reward the company that finally turns up.
Third and fourth: product and traction
Product and traction travel together, because the only thing that validates a product this early is whether anyone outside the building wants it. You're not assessing whether it's finished; you're assessing whether it solves a problem people feel keenly enough to change their behaviour for.
Traction is the evidence. With little or no revenue, that means signals of genuine pull: retention curves that flatten rather than bleed, repeat usage, pilots customers pay to extend, waiting lists that actually convert. One paying customer who'd be upset if the product vanished is worth more than a thousand free sign-ups who'd never notice. Ask for the raw numbers behind any headline metric - "growing 40% month on month" means something very different from a base of ten than from a base of ten thousand.
Fifth lens: the deal itself
A good company can still be a poor investment if the terms are wrong. This is where you read the cap table, the valuation and the term sheet rather than the pitch.
Is the valuation in line with comparable UK rounds at this stage, or has it drifted up because money was easy when they last raised? Is the cap table clean, or already crowded with passive early holders who took too much? What do the rights attached to the shares actually give you - and what happens to your stake at the next round? If you're not yet fluent in liquidation preferences, option pools and anti-dilution language, that's worth fixing before you sign anything; the terms quietly decide how much of any exit actually reaches you.
Where SEIS and EIS fit
For UK investors, this is also where you check whether the company qualifies for the venture capital schemes, because the reliefs change the maths materially. The Seed Enterprise Investment Scheme (SEIS) offers 50% income tax relief on up to £200,000 per tax year; the Enterprise Investment Scheme (EIS) offers 30% on up to £1,000,000 a year. Both need a three-year hold, exempt the gains from capital gains tax if relief was claimed, and carry loss relief that softens the blow when a company fails - the realistic outcome for a fair share of any early-stage portfolio.
The check itself is structural: has the company secured HMRC advance assurance, and will it issue you an SEIS3 or EIS3 certificate so you can claim? The company-side conditions - gross assets, employee count, age and how much it can raise - sit with the company, and several thresholds changed on 6 April 2026; current figures are on gov.uk. Hold the line on what this lens is for, though: a weak company with advance assurance is still a weak company. The relief should improve a deal you'd consider anyway - never be the reason for it.
Sixth lens: the downside
The last question is the one optimistic investors skip: what does this look like if it goes wrong, and can I live with that? Early stage is a power-law game - most write-offs are total, and a few winners are meant to carry the rest. So size the cheque as money you can afford to lose entirely, and think about whether you can follow on if the company does well. Running out of room to back your own winners is one of the quieter first-timer mistakes.
Then write down the two or three things that would have to be true for this to work, and the one that worries you most. If that worry is something you can't get comfortable with, that's a complete answer - and a perfectly good reason to pass.
Evaluating a startup: quick answers
What is the most important thing to evaluate in an early-stage startup?
At the earliest stages, most experienced angels weight the team most heavily, because the product, the market positioning and even the business model will almost certainly change before anything works. There's little operating history to analyse, so the question becomes whether these specific founders can learn faster than they burn cash. Team is rarely the only factor, but at pre-seed and seed it's usually the first.
How long should evaluating an early-stage startup take?
A first read of the deck and a founder call can take an hour. Serious diligence on a seed deal - references, market work, reviewing the cap table, contracts and SEIS or EIS advance assurance - more often runs across one to three weeks. The point isn't speed; it's having a consistent set of questions you ask every time, so you're comparing deals on the same terms.
What are red flags when evaluating a startup?
Common warning signs include founders who can't explain how they'll spend the money, vague or evasive answers about competitors, a cap table already crowded with passive early holders, valuations far above comparable rounds, no clarity on SEIS or EIS eligibility, and metrics presented without the underlying numbers. None is automatically fatal, but each is a reason to ask harder questions before committing.
Should SEIS or EIS eligibility change how I evaluate a deal?
The reliefs change the maths but not the underlying business. SEIS offers 50% income tax relief and EIS 30%, which can soften losses and improve net returns for UK taxpayers - but a weak company with advance assurance is still a weak company. Treat eligibility as a structural check during diligence, not as the reason to invest. This is general information, not advice; confirm your own position with an FCA-regulated adviser.
How do I evaluate traction when a startup has almost no revenue?
With little or no revenue, you look for evidence of pull rather than headline figures: repeat usage, retention curves, signed letters of intent, waiting lists that convert, or pilots customers actually pay to extend. The question is whether anyone outside the founding team genuinely wants the product, and whether that want is growing without the company paying heavily to manufacture it.
The Carry is independent editorial journalism, not financial or investment advice. Nothing here is a recommendation to invest in, or avoid, any company or scheme. Tax treatment depends on your individual circumstances and the rules can change. Seek advice from an FCA-regulated adviser before making any investment decision.