A deck is a founder's argument for why their company deserves a meeting. It is not a data room, and it is not proof. Most early decks are short, written fast and full of numbers without definitions. Your job in the first read is not to decide whether to invest. It is to decide what you now know, what you would need to know, and whether the company is worth the next hour of your team's time.
Knowing how to evaluate a pitch deck consistently matters because volume is high and attention is scarce. In a survey of 885 venture capitalists, Gompers, Gornall, Kaplan and Strebulaev found that the average firm screened about 200 companies a year and closed about four. Most of those decks get a few minutes. A repeatable method makes those minutes fairer to founders and more useful to your partners.
Start with fit before you judge quality
Before you form an opinion about the company, check whether it is a company your fund can invest in. A brilliant Series B deck is a pass for a pre-seed fund, and saying so quickly is a kindness. Write your fund's hard rules down so every reader applies the same ones.
- Stage: is the round one you lead or join, and is the company at the stage you back?
- Sector and business model: is it inside the areas your thesis names, or a stated exception?
- Geography: can you legally and practically invest where the company is incorporated and operates?
- Check size and ownership: does the round size leave room for your typical cheque and target ownership?
- Timing: is the round open now, or is this an early conversation for later?
When the deck does not say, mark the answer as unknown rather than as a failure. A deck that leaves out the round size has not broken a rule. It has created a question. Our guide on how to write an investment thesis covers turning your mandate into rules you can check this way.
How to evaluate a pitch deck slide by slide
Most decks follow a familiar order. Sequoia's guide to writing a business plan lists company purpose, problem, solution, why now, market potential, competition or alternatives, business model, team, financials and vision. Use that order as a checklist of what a reader needs, not as a test the founder must pass. A deck that combines slides or skips one can still be strong.
| Part of the deck | What a strong version shows | Question when it is missing or vague |
|---|---|---|
| Purpose | One plain sentence on what the company does and for whom | Who is the first customer, and what do they use it for? |
| Problem | A specific customer, the pain, and how they deal with it today | How do customers solve this now, and what does that cost them? |
| Solution | What the product does differently, shown rather than claimed | What can a customer do with it today that they could not before? |
| Why now | A change in technology, regulation or behaviour that opens the door | Why could this not have been built five years ago? |
| Market | A bottom-up size built from customers and price, with assumptions | How many customers like your first ten exist, and what would each pay? |
| Competition | Direct and indirect alternatives, including doing nothing | Who do you lose deals to, and why? |
| Business model | Who pays, how much, how often, and the cost to serve them | What does one customer bring in, and what does it cost to win and keep them? |
| Traction | Metrics with definitions, periods and units | How is this number defined, and over what period? |
| Team | Why these people, with the experience this problem needs | Who does what, and what gap do you plan to hire for first? |
| The ask | Round size, instrument, what the money buys and the milestones it reaches | What will be true in 18 months that is not true today? |
Survey data suggests where to spend your attention. In the Gompers et al. study, 95% of venture firms named the management team as an important factor and 47% named it the most important, ahead of business model, product and market. That does not mean the team slide matters most. It means that every slide is also evidence about how the founders think.
Read for evidence: pull out every claim
The most useful habit in deck review is to separate what the founder states from what you can check. On your second read, list each material claim on its own line with the slide it came from. A claim is anything that would change your view if it turned out to be false.
- 1Write the claim in the founder's words, with the slide number, such as "20 enterprise customers (slide 7)".
- 2Add the definition the claim depends on. Does "customer" mean paying, signed, or in a pilot? Is "revenue" booked, recognised or annualised?
- 3Note the period and units: monthly or annual, gross or net, which currency, as of which date.
- 4Check consistency across slides. Does the revenue on the traction slide match the financials slide? Does the team size match the burn?
- 5Mark each claim as reported by the founder until you have something else. Most deck claims stay that way until diligence.
This list becomes the spine of everything that follows: the questions for the first call, the requests for documents, and the memo if the deal goes forward. Keeping the slide reference matters because decks change. When the founder sends version three, you still need to know where a number first appeared. Our guide on how to verify startup traction covers what to ask for once a claim matters.
Test the numbers without a spreadsheet
You can catch most problems in an early deck with simple arithmetic and a few sanity checks. You are not building a model. You are checking whether the numbers hang together.
- Market size: does the founder build it from customers and price, or quote one large top-down figure? A bottom-up estimate you can follow is worth more than a bigger number you cannot. The TAM, SAM and SOM entry explains the terms.
- Growth: if the deck shows monthly growth, check the starting point. Tripling from a very small base says less than steady growth from a real one.
- Unit economics: if price, gross margin and acquisition cost are shown, does a customer pay back what it cost to win them in a period that makes sense for the business?
- Runway: divide the cash the round raises by the monthly burn the plan implies. Does it reach the milestones the deck promises?
- Projections: treat forecasts as a view into the founder's assumptions, not as a prediction. Ask which assumption the plan is most sensitive to.
When something does not add up, write it down as a question rather than a conclusion. Early decks are often assembled under time pressure, and a mismatch is frequently a labelling error. It becomes a concern only when the explanation does not hold.
Common mistakes when reviewing decks
- Judging design instead of substance. Polished slides tell you the founder has access to a designer. Plain slides with clear numbers often tell you more.
- Treating missing information as a fail. A pre-seed deck without revenue is normal. Score what is there and list what is not.
- Reading only the slides you find interesting. Use the same checklist on every deck so the tenth deck of the day gets the same attention as the first.
- Letting one striking number carry the deck. Write down its definition before it shapes your opinion.
- Forgetting where facts came from. Without slide references, a claim from an old version can end up in a memo as if it were current.
- Forming a view on the founder from tone. Judge the reasoning and the evidence, not confidence, accent or presentation style. Our guide on bias in startup screening covers practical safeguards.
A written scorecard helps with most of these. Our deal screening scorecard gives one you can copy into any tool.
Where Delsivo fits
Delsivo is designed to do the first, mechanical part of this method for you, so your team spends its time on judgement. In the pilot release, a founder uploads a PDF deck through your branded portal. Delsivo extracts the slide text, legible charts and company facts, and every fact cites the slide it came from. Weak extraction is flagged rather than guessed. The deck analyzer page shows how that works.
The extracted facts are checked against your thesis rubric. Each criterion shows the rule, the value found, the source and the reasoning, with a result of Met, Mismatch or Unknown. Unknown turns into a question for the founder, and a mismatch is surfaced for review, never an automatic rejection. Claims go into a tracker labelled founder reported until something corroborates them. AI suggests and cites. Your analysts and partners decide what happens next.
The method above works just as well with a document, a spreadsheet and discipline. The tooling only changes how much of it you have to do by hand.
Sources
- Sequoia Capital: Writing a business plan (checked 6 November 2026)
- Gompers, Gornall, Kaplan and Strebulaev: How Do Venture Capitalists Make Decisions? (NBER Working Paper 22587, September 2016) (checked 6 November 2026)
Delsivo publishes this guide and makes software for venture funds, so we have a stake in the topic. Facts about other companies, research and regulations come from the sources listed, on the dates shown, and can change. Corrections are welcome through our contact page.