What VCs Really Negotiate Before Writing the Check (It's Not Valuation)

Deep thoughts on AI and aspirations —— ByteDance Deep Thinking Circle

A survey of 885 venture capitalists across 681 firms reveals a contradiction most founders miss: VCs talk valuation constantly, but what they actually obsess over before wiring the money is something else entirely.

This research, often called the “instruction manual for venture capital,” breaks early-stage investing into six stages: deal sourcing, screening, pricing, term negotiation, portfolio management, and exit. The most misunderstood stages are precisely the third and fourth—pricing and terms. Most founders think negotiation centers on valuation numbers. The data tells a different story.

Start with Counter-Intuitive Numbers

The most sobering number in this set is the last one. Fewer than one in ten investments generates 10x+ returns, over half exit through acquisition, and nearly a third go to zero. This isn’t the profile of a failed firm—it’s the normal distribution across the entire industry.

This distribution determines all VC behavior: given that the hit rate is extremely low, the real decision question for every investment boils down to one thing: if this turns out to be the one, how much do I retain, and how deep a hole can I survive.

Valuation Is the Visible Card, Control Is the Hole Card

Understanding the distribution above makes term negotiations transparent.

The survey identifies a useful pattern: VCs are least willing to compromise on four items—board control, anti-dilution rights, liquidation preference, and valuation. They’re relatively flexible on dividends, option pools, funding amounts, and redemption clauses. On the surface this looks like negotiation style differences, but it actually exposes VC pricing logic: what they’re really buying is decision rights in extreme scenarios.

Back to an industry where 93% believe they can beat the market but nine out of ten deals are mediocre—why control matters more than money becomes clear. Liquidation preference determines who gets paid first and how much when the company is sold; anti-dilution rights determine how early investors adjust their position when the next round prices down; board seats determine who approves major decisions like replacing the CEO. These three elements show no value in normal times and only cash out when things go wrong. And in the VC world, things going wrong is the norm.

So the correct way to read a term sheet is: valuation is the price of this funding round, terms are the power structure of this partnership. Founders who haggle over valuation numbers while making wholesale concessions on terms are essentially trading long-term power for short-term face. The reverse strategy usually works better: give face on valuation, hold every inch of ground on control and preference rights.

Here’s my own supplementary judgment for reference: what’s actually worth fighting over in valuation negotiations is the connection between the current valuation and the next round, not the absolute number. The survey also shows most VCs don’t even build financial models in early stages—the most common metrics are IRR and investment multiples. Valuation is more craft than calculation. Since their number is essentially made up, every point you insist on may cost you concessions elsewhere. Choose carefully what you pay elsewhere.

Where Deals Come From Determines Who You Are

Another widely misunderstood stage is deal sourcing. Only about 10% of deals come from founders proactively reaching out; over 60% come from networks and portfolio company referrals.

These numbers have harsher implications for founders than they appear: your carefully crafted cold emails and pitch decks belong to the minority in most funds’ deal flow. The path into the mainstream pipeline is network referral—which means your last funding round, last partner, last customer are all writing invisible recommendation letters for you.

The implications for investors are more direct. The survey shows 49% of respondents believe screening ability is their biggest value-add, higher than post-investment support and deal sourcing. This ranking is worth half-believing. I lean toward thinking screening truly is first, but that 49% number comes from VC self-assessment, and in the same sample 93% believe they can beat the market—suggesting this industry systematically overestimates itself. Real-world excess returns more likely come from a combination of screening, networks, and post-investment work; people just prefer crediting their own judgment.

Two actionable conclusions for founders. First, start building relationships 12 months before fundraising, not sending BPs 12 days before; portfolio company referrals and mutual friend endorsements work better than any materials. Second, actively use your VC’s resources after closing—87% of surveyed investors participate in portfolio company strategy, and around 70% help connect customers and next-round investors. If you don’t actively use these services, they remain on paper, and half of their impression of you in the next round comes from records of this time together.

How to Use This Research

Start with boundaries. This data describes the classic form of traditional equity VC, and the survey was completed before AI models changed fundraising dynamics. Today’s seed round valuation inversions, acqui-hires by giants, and investors’ upfront scrutiny of revenue quality fall outside this data’s coverage. Mechanism-wise it still holds (control logic, power law distribution, network-driven deal flow); number-wise it needs discounting.

Within these boundaries, it serves three groups differently.

Founders currently fundraising: shift negotiation focus from valuation to term structure, especially liquidation preference and board arrangements; and accept the reality that most exits happen through acquisition, so maintain “acquirability” from day one—not surrendering, but respecting that 53% number.

Individuals doing investment or angel work: don’t get infected by that collective 93% confidence. The most useful thing about this research is it proves this industry averages mediocre. Excess returns belong to those who admit most of their judgments will be wrong and therefore design protection for every bet as “possibly the wrong one.” Position sizing, terms, portfolio construction—all tools for pricing error.

Managers allocating resources inside enterprises: this data translates directly—manage new business lines as a venture portfolio, accept high failure rates, design stop-loss structures for tail risk, place bets on the few high-ceiling judgments, and don’t outsource “value-add” to process.

One final thought. What’s really at stake across the fundraising table was never valuation, but decision rights in extreme scenarios. Valuation only determines how much you split if you win; terms determine how much you lose if things go wrong. In an industry where nine out of ten deals are mediocre, figure out first how to survive the losing nine, then talk about winning the one.

Key points: fewer than one in ten investments generate 10x returns, 53% exit through acquisition—this distribution drives VCs to price extreme scenarios through terms; valuation is the visible card, control is the hole card, founders should shift negotiation focus from numbers to power structure; 60% of deal flow comes from networks, start building relationships a year before fundraising; data describes traditional VC form, numbers require discounting in the AI era.