The asymmetry
Email's cost structure is the root of the problem. Sending one message and sending a million cost almost the same. Reading them does not. The recipient pays in attention for every message, and the sender pays nothing for the attention they consume.
A deposit on first contact moves a small part of that cost back to the sender, and it scales with breadth:
| Sender | Messages to new contacts | Deposits at $5 |
|---|---|---|
| A person writing to someone they want to meet | 1 | $5, refunded if they get a reply |
| A focused partnership search | 20 | $100, much of it refunded |
| An outreach campaign | 50,000 | $250,000 |
Illustrative figures. Recipients choose their own price.
Nobody reads the message to decide whether it's worthy, and there's no identity check. The price does the sorting.
Why refunds matter
If deposits were kept, the system would punish exactly the people it should welcome: someone writing with a real reason. Refunding wanted messages makes the final cost zero for them. Donating unanswered deposits to charity, rather than letting recipients keep them, removes any incentive to ignore people. How donations work.
An old idea
In 1992 Cynthia Dwork and Moni Naor proposed making senders solve a small computational puzzle for each message, to combat junk mail. Adam Back's Hashcash, in 1997, turned that into a practical stamp. Computational stamps failed for one reason: attackers control botnets full of free computing power, so computation costs the honest sender more than the spammer.
Paid schemes followed. The best known, Goodmail's CertifiedEmail, was adopted by AOL and Yahoo in 2006 and closed in 2011. It charged bulk senders a fraction of a cent to bypass filters. The money went to the mail provider, recipients had no say, and the public saw it as paying for access to their inboxes. Other paid-to-reach services stayed niche because paying small amounts online was awkward and both sides had to adopt.
What's different now
The recipient can impose it alone
Earlier schemes needed senders and providers to adopt a standard first. A recipient-side deposit works on day one: the recipient's address holds unknown mail and replies with terms. No sender has to install anything.
Payment friction is falling
Card checkout is familiar to every human sender, and prepaid credit spreads one payment across many messages. For machines, the x402 standard revives the long-reserved HTTP 402 Payment Required status code, letting software pay inside a web request. That makes cent-level prices practical.
The agents are the ones who can pay easily
Payment rails built for AI agents make paying frictionless for software. That sounds like a flaw, but it isn't: the goal was never to block agents. An agent sending one message for a person with a real reason pays trivially. An agent sending fifty thousand pays fifty thousand times.
Honest limits
- Price doesn't prove a person was involved. Where that matters, combine it with attestation. Compared here.
- Card fees make very small deposits uneconomic today, which is why card deposits start at a few dollars until prepaid credit and machine payments arrive.
- Well-funded senders can pay. For them the deposit becomes a strong signal of intent, which is still useful to the recipient.