The Handshake Deal Protocol
Paul Graham
14 March 2013

Silicon Valley runs on handshake deals. A handshake deal is a verbal commitment to a transaction. The actual transaction comes later, when documents are signed and money changes hands.

Why do we need handshake deals? Why not just wait till the actual transaction? Because things can happen fast in the startup world.

So both investors and founders need a way to reserve space in a transaction. Founders need it because creating documents and getting them signed would slow down their fundraising, and investors need it because if they had to wait for documents to get created and signed before they could commit, they'd miss out on the hotter deals.

Handshake deals are not unique to Silicon Valley of course. They tend to arise wherever trust is sufficiently high and speed is sufficiently important. Diamond dealers apparently use them a lot.

Unfortunately, things don't work as smoothly in Silicon Valley as among diamond dealers. This is not a closed community of pros who deal with one another day after day. Many participants in the funding market are noobs, and some are dishonest.

Every cycle we get reports of supposed handshake deals that fell through. Without video of the conversation it's hard for us to be sure whether there really was a deal and the investor welched, or there wasn't and the founders are just victims of their own wishful thinking.

The problem is compounded by the fact that some investors deliberately mislead startups about how interested they are in investing. Startups' prospects can change rapidly. If investors say no in a way that sounds like yes, they can essentially take a free option to invest. They haven't actually committed, so it costs them nothing, but if the startup turns out to be a hot one, they can retroactively claim that their almost-yes was an actual yes, and that the startup is morally obliged to let them invest.

Fortunately there is a way to fix most of these problems: to define a standard protocol for handshake deals. We're going to start using this within YC, and we hope it will spread to the rest of the startup community.

The protocol defines an offer as an amount to be invested, plus a valuation or valuation cap (or no cap), plus an optional discount. Here are some example offers:

  • $100k at $5 million pre-money.
  • $100k at a $5 million cap.
  • $100k uncapped.
  • $100k uncapped with a %10 discount.
According to the protocol, you have a handshake deal if and only if the following happens:
  1. The investor says "I'm in for <offer>."
  2. The startup says "Ok, you're in for <offer>."
  3. The startup sends the investor an email or text message saying "This is to confirm you're in for <offer>."
  4. The investor replies yes.
Unless and until this process is completed, there is no handshake deal. So it is in the interest of investors to complete the 4th step, because until they do the startup is under no obligation to take their money.

Since both parties will usually have mobile devices from which they can send such messages, they should ordinarily do it in person as the final step of the agreement. They should each regard it as suspicious if the other is unwilling to.

At the very least this protocol will tell us who's at fault if we get a report of a handshake deal falling through. But it should do more than that. A definite protocol that leaves a trail will both prevent founders from misleading themselves, and discourage investors from misleading them.

I don't think the offer has to specify the documents to be used. In practice this is rarely an issue. People either use one of the standard documents (for small investments) or negotiate in good faith (for large ones). Market terms are well enough understood that it should be easy to see who's at fault if one party is making difficulties about the terms, and that's all we ask from this protocol.

The protocol deliberately makes it impossible to say certain things. For example, an investor can't just say they'll invest $x, without specifying a valuation or cap. Investors who do that can escape their commitment later by claiming the price turned out to be too high. An offer to invest has to specify a valuation or cap, or no cap. Otherwise it's incompletely defined and thus not even an offer.

It also isn't possible to make a handshake deal on an offer to invest a range of money. Investors will sometimes try to make a deal to invest, say, $50k to $150k. If a startup agrees to that, they're obliged to save $150k of space but the investor is only obliged to invest $50k. An offer to invest a range of money is really two separate things: an offer to invest the bottom end of the range, plus an expression of interest in possibly investing more. So we suggest startups respond to each separately: do a handshake deal for the bottom end of the range, and respond politely to the investor's interest in investing more, but don't feel any obligation to take more money till the investor commits to investing it. Knowing they're guaranteed no more than the lower end of their range should sometimes cause investors to commit upfront to investing more. And if it doesn't, it would have been a mistake for the startup to rely on getting more.

Finally, it isn't possible to add conditions to a handshake deal. For example, there is no way for an investor to use this protocol to offer, as some investors try to do, to invest if other people will—e.g. to say that they'll invest as part of a larger round if you can find a lead. That sort of commitment is so worthless in practice that it's a mistake for startups either to rely on it, or to feel themselves bound by it. It's not even rightly considered an offer, but at best a lead (and one that will rapidly grow cold).

Startups and investors can of course make any sort of arrangement they want. But they don't have a handshake deal according to this protocol unless the terms are precise and unconditional.





Thanks to Sam Altman, Marc Andreessen, Paul Buchheit, Ron Conway, Ronny Conway, Chris Dixon, Ben Horowitz, Ash Patel, Geoff Ralston, Joshua Schachter, Harj Taggar, Albert Wenger, and Fred Wilson for reading drafts of this.




Apply | What We Do | Blog | People | Jobs | Quotes | FAQ | Contact | Lib | Legal