Venture Capital interview preparation

Questions

100

Traced to a firm

31

Firms

12

Updated

September 2026

Asked at

All firmsGeneral Atlantic9Insight Partners7Silver Lake6Vista Equity Partners4Bessemer Venture Partners3ACAccel2Advent International2Battery Ventures2Andreessen Horowitz1Coatue Management1Sequoia Capital1WPWarburg Pincus1

Topic

All topicsSourcing and deal flow5Market sizing and estimation8Founders and teams5Unit economics and cohorts11Term sheets12Cap table and dilution7Early-stage valuation7Portfolio construction6Board and governance4Down rounds and secondaries4Exits and liquidity4Fund economics5Sector theses and markets6India venture market6Fit and motivation10

Level

AnyCoreIntermediateHard

Type

AnyTechnicalFitCaseMarket viewBrainteaser

1. 051How do you value a company on ARR multiples, and when does that break?Early-stage valuationIntermediatetechnicalGrowth equitySaaS-focused funds

   Say this

   Enterprise value divided by annual recurring revenue, benchmarked against public comparables and recent private rounds, then adjusted for growth and retention. It breaks the moment the revenue is not actually recurring, and it breaks completely as a cross-sectional comparison because the multiple is a function of growth.

   Then walk it

   1. The mechanics: take ARR, not trailing revenue — the annualised value of contracted subscriptions at a point in time. Apply a multiple from public SaaS comparables, discount it for private illiquidity and scale, then adjust up or down for growth and net retention.
   2. Growth is what the multiple is really pricing. Public software has historically traded at roughly 4 to 8 times forward revenue for 20 percent growers and 12 to 20 times for 40 percent-plus growers with good retention. Those bands have moved violently — in 2021 the top decile traded above 30 times, and the same companies traded under 10 times eighteen months later with unchanged fundamentals.
   3. That volatility is the first thing that breaks it: the multiple is a market-sentiment variable, so an ARR multiple set at the top of a cycle is not a valuation, it is a timestamp.
   4. Second break: the recurring claim. Usage-based revenue, services revenue dressed up as subscription, annual contracts with no auto-renewal, or a revenue base where the top ten customers are on pilots — none of these are ARR, and companies routinely present them as such. Ask for contracted, auto-renewing revenue only.
   5. Third break: it ignores the cost of getting the revenue. Two companies at $20m of ARR growing 50 percent are the same on this metric and completely different if one has a burn multiple of 1.2 and the other 4. That is why the multiple always has to be read against efficiency and retention.
   6. So how I would actually use it: as a sanity check and a market-clearing reference, never as the primary output. The primary work is the growth durability and retention case, and the multiple is what translates that into a price the market will accept this quarter.

   Where candidates lose it

   Quoting a multiple band without saying that it is a function of growth and retention, and without acknowledging how far those bands moved between 2021 and 2023. Also accepting the company's ARR definition — interrogating what is actually recurring is most of the real work.

   Expect next

   • What would you accept as ARR and what would you strip out?
   • Would you pay a higher multiple for 40 percent growth or 130 percent net retention?
   • How do you value a usage-based pricing model?
2. 052What discount rate would you use for a Series A company, and can you defend it?Early-stage valuationHardtechnicalEarly-stage VC

   Say this

   Practitioners use 30 to 50 percent at Series A, and honestly I cannot defend a specific number inside that band. What I can defend is the logic: the rate has to reflect the probability of total loss, and it is doing the job that a proper probability-weighted scenario model should be doing instead.

   Then walk it

   1. The conventional ladder: seed 50 to 80 percent, Series A 40 to 60, Series B 30 to 50, growth stage 20 to 35, late-stage pre-IPO 15 to 25. Those come from practitioner convention and rough realised-return data, not from CAPM.
   2. Why not CAPM: there is no observable beta for a private company with no revenue, the cash flows are not a range around a central case, and the risk is overwhelmingly idiosyncratic rather than systematic. CAPM would give you something like 12 percent, which is absurd here.
   3. What the high rate is actually doing: it is a crude substitute for the probability of zero. A 50 percent discount rate applied to a success-case forecast is another way of saying most of these companies fail.
   4. Which is why the better technique is to separate the two: forecast the success case explicitly, then probability-weight it, and discount at something closer to a normal equity rate. A 60 percent chance of failure plus a 20 percent discount rate is far more defensible and far more debatable than a single 50 percent rate with a hockey stick behind it.
   5. The practical consequence to name: at these rates, cash flows beyond year seven are worth almost nothing, so any early-stage DCF is essentially a bet on a terminal value. Discounting $100m of year-ten value at 45 percent gives you about $3m. The output is whatever you assume the terminal value is.
   6. So my honest answer is that I would not run a DCF at Series A. I would use the venture method and comparable round pricing, and I would keep the discount rate discussion for a growth-stage asset where the cash flows are real enough to discount.

   Where candidates lose it

   Producing a confident single number with a CAPM build-up behind it. An experienced interviewer will take that apart in two questions. The strong answer gives the convention, explains what the rate is standing in for, and proposes the probability-weighted alternative — then says plainly that a DCF is the wrong tool at this stage.

   Expect next

   • So would you ever run a DCF on an early-stage company?
   • How would you probability-weight the scenarios instead?
   • What rate would you use for a growth-stage company with $80m of ARR?
3. 053What is an LBO?Early-stage valuationCorephone / first roundAdvent InternationalTechnology, Media and Telecom · Palo Alto · 2020

   Say this

   Buying a company using mostly borrowed money, where the target's own cash flows service the debt. You put in a slice of equity, pay down debt over the hold with the company's cash generation, then sell — and your return comes from deleveraging, EBITDA growth and any multiple expansion.

   Then walk it

   1. Structure: a typical deal is 40 to 60 percent equity today, the rest debt, secured against the target's assets and cash flows. The debt sits on the company, not the fund.
   2. The three return drivers, and you should always name all three. Deleveraging: every rupee of debt repaid converts directly into equity value at a constant enterprise value. EBITDA growth: revenue growth plus margin improvement. Multiple expansion: exiting at a higher multiple than you paid, which is the least controllable and the one you should never underwrite.
   3. Worked version. Buy at 10x $50m of EBITDA, so $500m, with $200m equity and $300m debt. Five years later EBITDA is $75m, debt is down to $150m. Exit at 10x is $750m, less $150m of debt, so $600m of equity on $200m in. That is 3x, roughly a 25 percent IRR, with no multiple expansion assumed at all.
   4. What makes a good LBO candidate: stable, predictable cash flows, low capital intensity, a defensible market position, hard assets or contracted revenue to lend against, and an identifiable operational improvement. Cyclical, capex-heavy or pre-profit businesses are bad candidates.
   5. Why a venture investor should know this: it is increasingly the exit path. Software companies at $50m of ARR with slowing growth and positive cash flow are now bought by software-focused buyout funds rather than IPO'ing, so the LBO maths sets the floor price for a lot of venture-backed companies. Vista and Thoma Bravo have taken dozens of them private.
   6. The obvious limitation: leverage magnifies both directions. The same structure that turns a 10 percent EBITDA gain into a 30 percent equity gain turns a modest miss into a covenant breach, and it is the reason buyout funds will not touch a business whose revenue can fall 30 percent in a quarter.

   Where candidates lose it

   Describing the debt and stopping. Name the three return drivers explicitly — deleveraging, EBITDA growth, multiple expansion — because that is what the question is checking. And in a venture or growth interview, connect it to the exit path for software companies, which is why they are asking a VC candidate about LBOs at all.

   Expect next

   • What makes a good LBO candidate?
   • Which of the three return drivers matters most?
   • Why does an early-stage investor need to understand this?

   Reported by candidates at Advent International (Technology, Media and Telecom, Palo Alto, 2020). Source: Wall Street Oasis.

4. 054An oil company loses forty million dollars of market capitalisation because of litigation, then sells an asset to pay for it. Is the share price drop justified?Early-stage valuationHardsuperdaySilver LakeTechnology, Media and Telecom · San Francisco · 2022

   Say this

   A $40m drop is justified only if the expected after-tax cash cost of the litigation is about $40m and nothing else changed. The asset sale is a separate question: if the asset was sold at fair value, the sale itself destroys no value and the share price should not move again for it.

   Then walk it

   1. First, price the liability properly. What matters is the probability-weighted, after-tax, present value of the cash outflow, plus any legal costs, less insurance recovery. A $40m headline settlement at a 25 percent tax rate and 70 percent probability is closer to $21m of economic cost.
   2. Second, ask whether the litigation revealed something. If it signals an ongoing practice that will generate more claims, or a regulatory exposure across the asset base, the drop should exceed the direct cost — the market is repricing future cash flows, not just paying a fine. That is usually the real answer for litigation-driven drops.
   3. Third, the asset sale. Selling an asset at fair value is value-neutral: you swap an asset for cash of equal value. Enterprise value falls by the asset's value, cash rises, equity value is unchanged.
   4. But sold at a discount, which is what a forced seller does, it is value-destructive twice over — once for the discount and once for the loss of an asset that may have been worth more inside the portfolio than to the buyer. A distressed sale to fund a settlement is a classic way a $40m problem becomes a $60m one.
   5. Then the tax detail worth mentioning for an oil asset: a sale can trigger a large gain against a low tax basis, so the after-tax proceeds can be materially less than the headline price, and the company may need to sell more than $40m of assets to net $40m.
   6. So the structured answer is: justified if the drop equals the after-tax expected cost and the litigation is genuinely one-off. Understated if it signals a systemic problem. Overstated if the market priced the headline number rather than the probability-weighted after-tax figure, which markets frequently do on litigation news.

   Where candidates lose it

   Answering yes or no. This is a framework question and the only wrong answer is an unconditional one. The two things you must separate are the cost of the liability and the information content of the litigation, and you must state that a fair-value asset sale is value-neutral while a forced one is not.

   Expect next

   • What if the asset was sold at a 20 percent discount to fair value?
   • How would you price the litigation if the outcome is binary?
   • Does the asset sale change enterprise value or equity value?

   Reported by candidates at Silver Lake (Technology, Media and Telecom, San Francisco, 2022). Source: Wall Street Oasis.

5. 055Would you rather buy a low quality business at a great price, or a high quality business at an okay price?Early-stage valuationHardsuperdayCoatue ManagementTechnology, Media and Telecom · New York · 2023

   Say this

   High quality at an okay price, and in venture that is barely a choice. A great business reinvests at high returns so time works for you. In a cheap bad business, intrinsic value erodes while you hold it and your entire return depends on a re-rating arriving quickly.

   Then walk it

   1. The compounding argument: a business earning 30 percent on incremental capital that can reinvest most of its cash flow converges your return on that reinvestment rate over a long hold, and a sensible entry multiple becomes second-order.
   2. The reverse for a low-return business: every year you hold, value is decaying, so you are renting a re-rating rather than owning a compounder. Get the timing wrong and a cheap asset stays cheap and gets cheaper.
   3. Horizon decides it, and say that explicitly. Over ten years, quality wins almost regardless of entry price. Over six months with a hard catalyst, the cheap asset can be the better risk-reward — that is an event-driven trade, not an investment philosophy.
   4. Why this is close to a non-question in venture specifically: entry price on the winner is nearly irrelevant to fund returns. If one company returns the fund 30 times, paying 30 percent more at entry turns 30x into 23x, which barely registers next to missing it. The cheap mediocre company returns 2x at best and consumes a partner's time for eight years.
   5. The honest counterargument, which you must give: 'high quality' is often just a description of a stock that already worked, and paying any price for quality is precisely how people lost money in the 2021 vintage. Quality at an okay price is fine; quality at any price is how you write down a fund.
   6. So my answer: quality with a valuation discipline. The error that permanently destroys capital is owning a declining business. The error of overpaying for a good one is usually survivable, given time.

   Where candidates lose it

   Giving the textbook Buffett answer with no acknowledgement of horizon or of the risk of overpaying for quality. The 2021 crossover vintage is the obvious counterexample and a good interviewer will raise it, so raise it yourself. And in a venture seat, connect it to the power law — that is the version of the answer that fits the seat you are sitting in.

   Expect next

   • When does the cheap asset win?
   • How do you avoid overpaying for quality in a hot market?
   • What does the power law do to this trade-off?

   Reported by candidates at Coatue Management (Technology, Media and Telecom, New York, 2023). Source: Wall Street Oasis.

6. 056Explain the power law and what it actually means for how you vote in a partners' meeting.Portfolio constructionIntermediatetechnicalEarly-stage VCSeed funds

   Say this

   Venture returns are not normally distributed — a small number of investments produce most of the return, and roughly half return less than the capital invested. It means the only question that matters in a partners' meeting is whether this company could return the fund, not whether it is likely to lose money.

   Then walk it

   1. The shape: across a typical early-stage portfolio, something like 50 to 60 percent of investments return less than 1x, a middle band returns 1 to 3x, and one or two return 10x or more and produce the majority of the fund's gains.
   2. So the asymmetry is total. The downside of any single investment is capped at 1x of a small cheque; the upside is unbounded. That means the cost of a false negative — passing on the outlier — is enormously higher than the cost of a false positive.
   3. Which changes the question you ask. Not 'what is the probability this works' but 'if it works, is it big enough to return the fund?' A company with a 60 percent chance of becoming a $200m business is a worse portfolio decision than one with a 10 percent chance of becoming a $10bn business.
   4. This is why consensus voting is dangerous in venture. The companies that produce outlier returns are usually the ones that divide the partnership, because by definition consensus ideas are priced. A firm where every investment is unanimous is probably screening out its best decisions.
   5. It also dictates reserves. If one company will produce most of the return, the right behaviour is to concentrate follow-on capital into the names that are working and stop funding the middle. The hardest discipline in the job is refusing to feed a decent company that will never be an outlier.
   6. The limitation worth naming: the power law is a description of outcomes, not a licence for recklessness. It gets misused to justify paying any price for anything with a big story. The constraint is still that the portfolio has to be constructed so that one outlier is enough — which means enough shots, and enough ownership in each.

   Where candidates lose it

   Reciting 'one investment returns the fund' as a slogan. The interviewer wants the consequences: how it changes the question you ask in diligence, why it argues against consensus decisions, and what it implies for reserves. And you should name the misuse of it, because 'power law' has become the standard excuse for undisciplined pricing.

   Expect next

   • How many investments does a fund need for the power law to work?
   • So would you back a company the whole partnership disliked?
   • What does this mean for how you allocate reserves?
7. 057How many investments should a hundred million dollar seed fund make, and how much do you reserve?Portfolio constructionHardtechnicalSeed fundsIndian venture capital

   Say this

   Roughly thirty to thirty-five companies with about half the fund reserved for follow-ons. So call it $45m of initial cheques averaging $1.4m for 10 to 15 percent ownership, $45m of reserves, and $10m for fees and expenses over the fund's life.

   Then walk it

   1. Start from the return requirement and work back. A $100m fund needs $300m gross to return 3x net-ish to LPs. If one company produces $200m of that, I need to own enough of it: a $2bn exit with 10 percent retained ownership gives $200m. So the entry ownership target has to survive dilution to 10 percent.
   2. That fixes ownership at entry around 12 to 15 percent, because three later rounds will roughly halve it unless I follow on. Ownership target, not cheque size, is the primary constraint.
   3. Then portfolio size. Too few names and you may simply not own an outlier; too many and you cannot own enough of each or spend time on them. Thirty to thirty-five is the conventional band for seed, and the maths behind it is that at roughly a 1-in-20 hit rate for a fund-returner you want at least twenty-five shots.
   4. Reserves: 50 percent is the standard split at seed and it is the single most consequential construction decision. A fund that deploys 80 percent into initial cheques gets crushed in the winners, because the Series B and C are where the ownership is defended.
   5. Fees drag, and you should mention it because it catches people out. A 2 percent management fee over ten years is roughly 20 percent of committed capital, though most funds step it down. So the investable capital out of $100m is $80m to $85m, not $100m, and every portfolio-construction number has to be built off the investable figure.
   6. The India-specific adjustment: at seed in India, cheque sizes of $1m to $3m buy meaningfully more ownership than the same cheque in the Bay Area, so the same $100m fund can run a slightly more concentrated book at higher ownership. The offsetting constraint is exit scale — fewer billion-dollar outcomes means the fund-returner has to come from a smaller pool of candidates.

   Where candidates lose it

   Giving a portfolio count with no arithmetic behind it. Build it from the fund-return requirement through ownership target to cheque size — that sequence is the answer. And forgetting the fee drag, which makes every construction number 15 to 20 percent tighter than the headline fund size suggests.

   Expect next

   • What if you could only make ten investments?
   • How would that change for a $500m multi-stage fund?
   • How do you decide which companies get the reserves?
8. 058When do you decide not to follow on?Portfolio constructionHardsuperdayEarly-stage VCSeed funds

   Say this

   When I would not make the investment cold at the new price. That is the only test, and applying it honestly is hard because I am anchored on my entry price and on not wanting to signal doubt. Sunk cost and signalling are the two forces pushing every follow-on decision the wrong way.

   Then walk it

   1. The discipline: re-underwrite the company from scratch at the new price as if a stranger brought it to me. If I would pass, I pass, and my existing position is irrelevant to that judgement.
   2. The specific triggers for not following. The team has changed in a way that removes the reason I invested. The market turned out to be structurally smaller than underwritten. The metrics are fine but the shape is wrong — growing revenue with deteriorating retention. Or the price now requires an exit outcome I do not believe in.
   3. The uncomfortable one: the company is doing fine and will probably return 2 to 3x, but it will never return the fund. In a power-law portfolio that capital is better spent defending the position in a potential outlier. Passing on a healthy company is the hardest call in the job and it is usually right.
   4. Signalling risk is real and you should address it rather than pretend it is not. If an existing investor does not participate, incoming investors read it as information, and it can genuinely make the round harder for the founder. So the decision has to be communicated early, directly to the founder, with a clear reason — never by going quiet.
   5. What I would do to make it cleaner: agree the reserve policy in advance at the portfolio level, so the decision is a framework being applied rather than a verdict on the company. And where I can, offer to introduce other investors, which is the honest version of support when I am not writing the cheque.
   6. One structural caveat: a fund at the end of its investment period with no dry powder has no choice, and everyone in the market knows it. That is a fund-construction failure showing up as a portfolio decision, which is exactly why reserves are set at the start.

   Where candidates lose it

   Answering only on the company's merits and ignoring signalling risk. It is the thing that makes this decision genuinely difficult, and interviewers want to hear you handle the founder conversation. Also failing to mention the hardest case — the perfectly decent company that cannot return the fund.

   Expect next

   • How do you have that conversation with the founder?
   • What is signalling risk from a multi-stage fund?
   • Would you ever follow on just to protect the signal?
9. 059Would you rather own twenty percent of a five hundred million dollar outcome or five percent of a five billion dollar outcome?Portfolio constructionIntermediatetechnicalEarly-stage VCSeed funds

   Say this

   The second: $250m against $100m. But the real answer is that it depends on cheque size and the probability of each, because the two positions are not bought for the same price and not with the same likelihood.

   Then walk it

   1. The arithmetic first, fast: 20 percent of $500m is $100m. 5 percent of $5bn is $250m. The billion-dollar outcome wins by 2.5 times even with a quarter of the ownership.
   2. That is the core lesson of venture and why ownership discipline can be overrated: outcome size dominates ownership. A partner who insists on 20 percent and therefore passes on the companies that will not sell 20 percent is optimising the wrong variable.
   3. But then the cost side, which is what makes it a real question. To hold 20 percent of a $500m company you probably invested $3m at seed and defended it. To hold 5 percent of a $5bn company you may have put in $2m at seed and been diluted, or $50m at Series D. The multiple on invested capital could favour either.
   4. And probability, which is the part candidates skip: the $500m outcome is perhaps ten times more likely than the $5bn one. On expected value the two can be close, and for a small fund the $500m outcome may be perfectly sufficient while for a $2bn fund it is noise.
   5. Which is the real point: the answer is determined by fund size. A $50m fund is made whole by the $500m exit. A $1bn fund needs the $5bn one, which is why large funds structurally cannot invest in companies with $500m ceilings, regardless of how good those companies are.
   6. So my answer: the $5bn outcome, and I would say the interesting version of the question is not which I prefer but what fund size makes each one the right target.

   Where candidates lose it

   Doing the arithmetic and stopping. It takes five seconds and is not what is being tested. The content is in tying it to fund size and to the probability of each outcome — that is what turns a mental-maths question into a portfolio-construction answer.

   Expect next

   • How does fund size change your answer?
   • What ownership do you actually target at seed, and why?
   • If the $5bn outcome is ten times less likely, which do you pick?
10. 060What makes your investment philosophy different and better from others'?Portfolio constructionHardsuperdayGeneral AtlanticGrowth Equity · New York · 2022

    Say this

    State something narrow enough to be wrong, then say what it costs you. A philosophy that excludes nothing is not a philosophy. And be careful with 'better' — the defensible claim is that it is a genuine edge in a specific slice of the market, not that it dominates everyone else's.

    Then walk it

    1. Pick a real lane and say it in one sentence. Something like: I look for businesses where the distribution channel is the moat rather than the product, because product advantages in software now decay in eighteen months and channel advantages compound.
    2. Then say what it makes you pass on, which is the part that proves it is real. That philosophy means passing on most pure-technology plays and most companies whose pitch is a model or a feature. Naming the exclusion is what makes it falsifiable.
    3. Then the edge claim, carefully. 'Better' in investing means one of three things: better information, better judgement, or better access. Only the first and third are checkable, so I would argue from those — a specific network, a specific operating background, a specific market where I see things earlier.
    4. Ground it in one concrete instance. A company you looked at, what the consensus view was, what you saw that was different, and what happened. A real example beats any amount of framework.
    5. Then connect it to the firm, because in a growth-equity interview this question is partly 'do you understand what we do'. If they run concentrated growth rounds with an operating team attached, a philosophy built on post-investment value creation fits; one built on early-stage pattern recognition does not.
    6. And be honest about the limit: my philosophy would have missed some of the best companies of the last decade, and here is the category it would have missed. That admission is what makes the whole answer credible rather than promotional.

    Where candidates lose it

    A philosophy so broad it excludes nothing — 'I look for great teams in large markets' is what everyone says and therefore says nothing. The second trap is the word 'better': claiming superiority over a firm's existing approach in their own office is a bad trade. Argue for a specific edge, name what it costs you, and say what it would have missed.

    Expect next

    • What would that philosophy have made you miss?
    • Give me a specific company where it produced a different answer from consensus.
    • How does it fit with what we do here?

    Reported by candidates at General Atlantic (Growth Equity, New York, 2022). Source: Wall Street Oasis.

Firm tags come from public, anonymous candidate reports on Wall Street Oasis: strong signal, not sworn testimony. Firms are named as the places a question was reported, not as partners of Fin Maverick. Answers are written for this page to show how to think out loud; they are not scripts to recite.