Suppose a fund put $1 million into a seed round at a $20 million post-money valuation, for 5 percent, and the company now raises a $20 million Series A at $100 million post. The new shares dilute the seed stake to 4 percent, so the fund’s pro rata, the purchase that keeps it at 5 percent, is $1 million for 1 percent. To keep the arithmetic readable I will ignore dilution from any later round and all liquidation preferences, which a real model would include. On those assumptions the seed check returns five times its cost at a $125 million exit, while the $1 million follow-on needs a $500 million exit to do the same.
A blended cost basis hides that gap. The combined position, $2 million for 5 percent, returns five times its cost at a $200 million exit, and at that exit the new money has returned two times. A memo that reports one multiple for the whole position lends the second check the first check’s price.
The question for the new money is whether it beats the next-best use of the same dollars. In a fund of fixed size, a follow-on comes out of reserves that could back another holding’s round or, while the fund is still investing, a new company. I would write the follow-on memo against that alternative, as if for a partner with no position in the company. The original memo and two years of board work belong in it as evidence about the company and the team, which is where an insider’s real advantage lies. They do not change the price of the new shares.
The allocation carries information too. When new investors compete for a round, the company has reason to cut insiders back; when no outside lead appears, it may ask them for their full pro rata or more. That is one reason I would set the conditions for spending reserves at the first check, before anyone knows how the company will do: how much is held back, and what must be true to release it. An outside lead setting the price is a condition worth writing down. A round without one then becomes its own decision, made with the extra scrutiny that easy availability calls for.
The hard case is signaling. If insiders pass, the company may struggle to raise, and an existing investor that still believes in it can reasonably invest partly to prevent that. I think that is a legitimate choice, and a different one from underwriting. The memo should call the check support, size it from reserves set aside for that purpose, and keep it apart from the record of what the fund chose on the merits, so that later the fund can tell which kind of decision paid off.
Revised October 2026: this version adds the signaling case, in which an existing investor may invest as support, provided the memo records it that way.