
Module 4 — Running the Process · Lesson 4.3
Negotiating the Term Sheet
Liquidation preference, anti-dilution, the board and pro rata — read against the NVCA models
What you'll learn
- Explain liquidation preference, participation and how they change payouts in a sale
- Distinguish broad-based weighted-average from full-ratchet anti-dilution using the model formula
- Read the board-composition and pro rata provisions of the model documents
- Prioritise negotiation using evidence on which terms investors treat as fixed
Founders tend to negotiate the valuation and skim the rest. That is backwards often enough to matter. A slightly lower price with clean terms can leave a founder better off in most outcomes than a higher price with a large preference or a full ratchet. The way to read a term sheet is clause by clause against the market standard, asking two questions of each: what does it pay out, and who does it put in control?
Liquidation preference and participation
The liquidation preference decides who is paid first, and how much, when the company is sold or wound up. In the NVCA model Certificate of Incorporation (updated October 2025), each series of preferred stock is entitled, before any payment to common stock, to a multiple of its original issue price — the model leaves the multiple as a blank to be negotiated — plus declared but unpaid dividends. The model treats a sale of the company as a deemed liquidation event, so the preference applies to an acquisition, the usual exit, not only to a shutdown.
The model offers two alternatives for what happens next. With non-participating preferred, the investor takes either its preference or what it would receive by converting to common, whichever is more. With participating preferred, the investor takes its preference and then also shares in what remains as if converted — downside protection plus upside.
The difference is large in a middling sale. Suppose investors put in 10 million dollars for 25 percent and the company sells for 30 million. With a 1x non-participating preference, they compare 10 million with 25 percent of 30 million (7.5 million) and take the 10 million; the common holders share 20 million. With 1x participating, they take 10 million and then 25 percent of the remaining 20 million, 15 million in all, leaving common 15 million. At a 300 million dollar exit the two barely differ. Participation bites exactly in the modest outcomes where founders most need the proceeds.
The 2016 survey of venture capitalists by Gompers and co-authors reports that a 1x preference is typical, that 2x or greater appeared in 19 percent of their investments, and that participation rights were used on average 53 percent of the time — less often in information technology and in California, where the authors found terms more founder-friendly overall.
Anti-dilution
Anti-dilution protects preferred holders if the company later sells shares at a lower price than they paid. It works by lowering the price at which their preferred converts into common, so they receive more common shares on conversion.
The NVCA model gives two versions. The standard one is broad-based weighted average. The model's formula is CP2 = CP1 × (A + B) ÷ (A + C), where CP1 is the conversion price before the new issue; A is the common outstanding before the issue, counted broadly to include options and convertible securities; B is the number of shares the new money would have bought at the old price; and C is the number of shares actually issued. A small down round moves the price a little; a large one moves it more, in proportion.
The alternative is full ratchet, which resets the conversion price to the new, lower price regardless of how few shares were sold at it. A tiny round at a low price can hand the earlier investors a large transfer of ownership from everyone else. The 2016 survey found weighted average to be the more standard choice and full ratchet used in 27 percent of investments. A founder seeing full ratchet should ask why, and expect to give something to remove it.
Control: the board and the voting agreement
Control terms often matter longer than economics. In the NVCA model Voting Agreement — most recently updated in June 2026 — stockholders agree to vote their shares so that the board is composed as the agreement specifies: one or more directors designated by named preferred investors for so long as they hold a threshold of shares, and one or more common directors — the model's examples include a director designated by key holders of common stock and a seat held by the chief executive. The same agreement can include a drag-along provision, under which holders agree to support a sale of the company approved by specified holders.
The arithmetic of the board is the arithmetic of control. A five-person board with two investor seats, two common seats and one independent is a different company from one with three investor seats. Founders should model the board not just after this round but after the next one, when another lead will usually want a seat.
Pro rata, information rights — and where to spend effort
The model Investors' Rights Agreement (October 2025) gives each Major Investor a right of first offer on new securities the company proposes to sell, letting them keep their percentage in future rounds — the pro rata right — subject to listed exceptions. It also sets the information rights covered in Lesson 3.2. Pro rata rights are close to universal: the survey found them in 81 percent of investments. Lesson 2.2 showed how they compound dilution for everyone else when several investors exercise them.
The survey also asked how flexible investors are on each term. The least negotiable, in order, were pro rata rights, liquidation preferences, anti-dilution protection, vesting, valuation and board control. They were more flexible on the option pool, participation rights, the investment amount, redemption rights and, most of all, dividends. That is a map for a founder's effort. Do not spend goodwill trying to remove pro rata or a 1x preference; the market treats them as standard. Spend it on the terms that are both costly and movable — participation, the size of the option pool, a full ratchet, an above-1x multiple — and on the board.
The usual cautions apply with force here. These documents are drafted by counsel, every blank is a negotiation, and this lesson is not legal advice. Its job is to make sure you know which blanks matter before you sit down.
Compare term sheets side by side, on the record
- 1
Store each term sheet with the round
Put every term sheet in the round's linked Drive folder, and move the investor's status on the deal in /deals to Term sheet with a next step dated to the response deadline.
- 2
Have the agent tabulate the terms
Attach the term sheets to a chat in /ai-agent and ask for a table: price, pre-money, preference multiple, participation, anti-dilution type, board seats, pro rata, option pool, and any departure from the NVCA model. Check every cell against the document yourself.
- 3
Model the payout, not just the price
Ask the agent to compute proceeds to common at three exit values — modest, expected and large — under each term sheet. The cleaner sheet often wins in the modest case.
Reading the sheet
- Proceeds to common at a modest exit
- What common holders receive at an exit near the post-money valuation, under each term sheet.
- Healthy signal: Compare across sheets. A Kavanah suggestion: if one sheet's price is higher but this number is lower, the price is not the better offer.
- Departures from the model
- Count of terms that differ from the NVCA model's standard alternatives (above-1x preference, participation, full ratchet, unusual control rights).
- Healthy signal: Each one should be understood and, ideally, traded for something. Several at once is a signal about the investor.
- Board composition after the next round
- Projected investor, common and independent seats after one more priced round.
- Healthy signal: Know it before signing. Control is easier to keep than to win back.
Key takeaways
- ·Liquidation preference and participation decide payouts in modest exits; 1x non-participating is the founder-friendly baseline.
- ·Broad-based weighted-average anti-dilution is standard; full ratchet transfers far more on a down round.
- ·The voting agreement fixes the board; model control after the next round, not just this one.
- ·Investors treat pro rata and 1x preferences as fixed; spend negotiating effort on participation, pool, ratchets and the board.
A signed term sheet is a plan to close, not a close. The last lesson of the module covers the mechanics between signature and money in the bank — including the Form D.
Sources
- 1.NVCA Model Certificate of Incorporation
National Venture Capital Association · NVCA Model Legal Documents (updated October 2025) · 2025
Liquidation preference as a negotiated multiple of the original issue price plus declared but unpaid dividends; deemed liquidation events; participating and non-participating alternatives; the broad-based weighted-average formula CP2 = CP1 × (A + B) ÷ (A + C) and the full-ratchet alternative.
- 2.NVCA Model Voting Agreement
National Venture Capital Association · NVCA Model Legal Documents (updated June 2026) · 2026
Board composition by designation (preferred directors, common directors including the CEO director) and the drag-along provision.
- 3.NVCA Model Investors' Rights Agreement
National Venture Capital Association · NVCA Model Legal Documents (updated October 2025) · 2025
Major Investors' right of first offer on new securities (pro rata) and its exceptions.
- 4.How Do Venture Capitalists Make Decisions?
Paul Gompers, William Gornall, Steven N. Kaplan, Ilya A. Strebulaev · NBER Working Paper 22587 · 2016
Frequency of terms (pro rata 81%, participation 53%, full ratchet 27%, 2x+ preference 19%); 1x typical; least negotiable terms in order; more founder-friendly terms in IT and California.