Jason Lemkin on Seed Investing, Due Diligence and the 2023 Venture Market

Jason Lemkin tells Harry Stebbings that the conventional advice to new venture investors is not to rush into major deals in their first year. He ignored it, and says his first five investments included Pipedrive, Algolia, Talkdesk, Greenhouse and Salesloft. The valuations and future IPOs he mentions are his 2023 account, not all realized exits.

Lemkin advises a new investor with a strong network or brand to use that advantage early rather than wait automatically. Established colleagues can help with ownership and investment discipline. He argues that an early, durable winner can shape an investor’s reputation for years, though his examples cannot establish that early success is typical.

His contrasting advice for someone joining another firm is to remember that the fund belongs to its managers. He says he did not take that lesson calmly enough: a new partner may not control the management company or fully understand its economics. His claimed lifetime returns reflect his own assessment.

Lemkin credits part of his early success to having time to hunt for founders without existing boards or distressed investments competing for attention. He also stayed within a narrow range of products, stages and founders he understood, including search, contact-center software and sales tools.

Knowing his preferred deal type helps determine how much risk and valuation he can accept. Lemkin warns against relying too heavily on current public-company comparisons: markets can change. He describes sizing checks against the fund, rather than treating a single comparable as a permanent ceiling.

His two preferred categories were earlier seed companies with roughly 10 to 20 customers and later seed businesses approaching $1 million in annualized revenue. He aimed for double-digit ownership, using smaller checks in the former and up to $4 million or $5 million in the latter. Rising prices have made the later strategy harder for his fund, he says.

Within a reasonable range, Lemkin cares more about getting the target ownership in a strong company than negotiating a small change in valuation. Stebbings pushes back that a move from a $12 million to a $25 million valuation can materially increase the check and reduce diversification for a small fund.

In his assessment, seed was vibrant in 2023 despite talk of a downturn. Lemkin sees more founders and executives with liquidity writing small checks and creating fragmented rounds. He says such small allocations seldom fit his fund’s return model, although he occasionally makes them for other reasons.

Lemkin expects well-connected founders to continue assembling rounds from many angels. He recounts advising a founder with $3 million in commitments to close instead of waiting indefinitely for a lead. If institutional capital is available on good terms, he says, combining it with individual investors can also work.

Stebbings notes that multistage funds have also moved toward seed as later-stage activity slowed. Lemkin expects some of that interest to fade: a very large fund needs substantial ownership in large outcomes, and tiny early checks cannot substitute for deploying meaningful capital into winners. Staying active can nevertheless keep investors connected to founders.

Lemkin suggests seed investors look beyond prominent insiders for founders who may have less access to networks and more reasonable prices. He argues that the most sought-after seed deals can be too expensive for small funds, even if their prices reflect competition. High valuations can also make a subsequent round difficult.

Stebbings says he has heard investors claim they never invest from cold inbound. Lemkin disagrees, citing his own experience and a discussion with other investors who said they had found strong investments that way. Preferences for a short email or a detailed deck differ, and he suspects raw inbound is less useful at later stages.

Lemkin does not accept that seed investing must return exclusively to unusually risky frontier projects. He believes small seed funds may struggle to earn attractive returns in the hottest rounds and must find promising founders earlier. His conclusion is a strategy judgment, not evidence that every popular deal will fail.

He recounts a founder offered an investment at a $450 million valuation who reconsidered after hearing what would be required to generate a fivefold net return after dilution. Lemkin uses the anecdote to argue that a high price has consequences for future outcomes; it is one founder’s experience.

Asked about crowded AI categories, Lemkin says markets are periodically reinvented. He preferred to wait longer in crowded categories, citing Pipedrive in CRM, while Algolia had fewer direct rivals when he invested. In AI sales tools he would like to see a company reach $1 million to $2 million in revenue before choosing, but prices can make waiting costly.

On software spending, Lemkin believes businesses had passed the worst of a period of budget cuts. He describes customers first optimizing cloud workflows, then removing little-used applications, and finally substituting features already included in other products. These are his observations from 2023, not a universal measure of the market.

He points to company results and management commentary as signs that spending was stabilizing, and suggests AI workloads could encourage new investment. He does not expect a rapid return to 2021 conditions. His optimism about an eventual rebound and 2024 IPOs was a forecast at the time.

Lemkin calls the Series A and B environment difficult, even though investors with established funds still had capital. In his view, founders’ expectations of a fast, highly priced round often conflicted with investors’ renewed discipline. He says some companies would fail, while others could build traction and raise later.

Stebbings describes portfolio companies with years of runway but no product-market fit after raising preemptive rounds. Lemkin says the outcome is uncertain and advises founders who abandon their original plan to offer remaining money back, not necessarily to return it unilaterally. He cites Slack’s predecessor as an example of investors choosing to back a pivot.

Lemkin believes some founders no longer appreciate the obligations attached to venture funding. He recalls a young founder expecting a much higher valuation within months and another questioning why an investor cared about unused cash. These anecdotes underpin his opinion, not a claim about all founders.

Stebbings prefers older or repeat founders, saying experience can prevent costly mistakes even at a higher price. Lemkin prefers taking a chance on first-time founders where he can contribute operating experience. He argues that introductions to sales leaders and other hires can reduce some, though not all, mistakes.

They debate whether a good founder should need an investor. Lemkin distinguishes dependency from accelerating a capable outsider’s access to a network. A founder arriving in a new ecosystem cannot recreate years of relationships immediately; Stebbings responds that access exists on a spectrum rather than in two neat categories.

Lemkin says his own spreadsheet showed around 90% of his ordinary investments making money, versus about half of the pre-revenue repeat-founder investments he analyzed. Those figures describe his stated experience and include his forward-looking judgment; they are not industry-wide realized-return statistics. Higher prices and an extra period before revenue make the latter harder for his fund.

His biggest regret is that he did not build a larger team as the SaaS market expanded. Earlier, with fewer investors and startups, he found it easier to see opportunities. He chose to remain a founder-first investor rather than manage a much larger organization, despite encouragement from limited partners.

The size of a fund follows from team and deployment choices, Lemkin and Stebbings argue. A larger team needs enough capital to support partners and reserves, but a $500 million fund also requires substantial exits to generate its desired return. Their arithmetic is an illustration, not a promised outcome.

Lemkin sees the dislocation in Series A as a potential investment opportunity. He says capital remains available for sought-after seed companies, while Series A investors were more disciplined outside AI. His rule of thumb is that each successive round should be harder, unlike the unusually easy escalation he recalls in 2021.

On diligence, Lemkin recalls a technical expert warning an investor against a company, after which the investor proceeded anyway. He generalizes from his experience that venture diligence can become confirmatory once a decision is made. The unnamed allegation and outcome are his account and cannot be independently established from the interview.

Lemkin asserts that every VC has fraud in a portfolio and predicts that some problems may never become public. He gives examples such as overstated recurring revenue or margins, while noting that lies are not always crimes. His sweeping assertion is unverified; misstatements, legal fraud and the evidence required to establish either are not interchangeable.

Stebbings argues that investors should confront founders who misrepresent figures rather than protect their popularity. Lemkin says difficult conversations have sometimes hurt his standing with founders but, in his account, helped companies. He also recognizes that leaving a toxic situation can be a rational use of an investor’s time.

For growth-stage SaaS in 2023, Lemkin describes an active but selective market. He suggests efficient companies with $30 million to $60 million in annual recurring revenue could receive offers around 10 to 15 times ARR, with exceptions. These are his observations and rough ranges, not guaranteed terms.

Some founders may defer a primary round if they are near break-even or do not need the cash, while earlier high valuations can make a new round difficult. Lemkin says growth investors are also seeking secondary purchases from early holders in his portfolio. A secondary transaction transfers existing shares rather than necessarily funding the company.

Some microfunds sell secondaries to return cash, he says, but his own largest limited partners told him they preferred a potentially higher eventual return over a discounted sale. Stebbings notes that liquidity needs vary: endowments and foundations can have ongoing obligations. Lemkin acknowledges his sample is limited.

One LP suggested reopening an earlier fund, subject to the necessary consent, so more capital could participate in a potentially strong portfolio. Lemkin says the conversation reminded him to distinguish gross from net returns and to understand each LP’s objectives. A fund’s deployment strategy and actual results remain separate questions.

Lemkin illustrates why a dramatic company-level multiple is not necessarily a fund-level triumph: a 10% stake in a $500 million exit returns roughly $50 million before relevant adjustments, or one times a $50 million fund. The example omits details such as dilution, fees and expenses; a one-times gross return is not the same as a net profit.

Asked whether fundraising was exceptionally hard, he says it was tougher than the 2020–2022 boom but easier than when he started. He describes three LPs with different approaches: one reducing managers, another changing strategy and a third adding new managers. His anecdotes show different behavior, not a representative survey.

Stebbings expects pressure at both the largest funds and inexperienced emerging managers, with some advantage for established mid-sized funds. Lemkin says enormous sovereign pools may value the ability to deploy large amounts at returns that suit their own objectives. Both are describing possible fundraising dynamics, not guaranteed fund performance.

They argue that LP turnover can reflect a change in the type and size of investors a fund seeks. Lemkin also expects large amounts of capital to flow into companies before public listings, though the amounts depend on market conditions. On Tiger and SoftBank, he says he is not close enough to know their futures.

Lemkin says he has become more skeptical about distributed work and some technology employees’ effort since the boom. He makes broad, critical claims about the current generation of workers and their hours, drawing on his own experience; the interview does not establish those claims as workforce-wide facts.

For new entrants, he suggests a difficult market can reward people who bring a fresh perspective and adapt. He sees some workers and salespeople as still anchored to unusually favorable 2021 conditions. This is his interpretation rather than a measured account of individual performance.

In the final round Lemkin predicts that the second half of 2024 will bring roughly one notable IPO a week, provided conditions hold. He thinks large listings such as Stripe or Databricks could restore confidence, and acknowledges that worsening markets would change the picture. The speakers jokingly discuss a bet before thanking each other; the prediction was not an established outcome.