What’s the difference between delivery, acceptance, and inboxing?

These three words get used interchangeably in vendor dashboards, and that sloppiness costs senders money. They mean different things, and only one of them tells you whether the email got read.

Delivery means the sending attempt did not bounce. The receiving server did not slam the door with a 5xx permanent failure or a 4xx temporary failure. That is it. "Delivery" tells you the SMTP conversation finished without an error, nothing more. Most ESPs report this as their headline number because it is the easiest one to make look good.

Acceptance is the stricter version. The receiving Mail Transfer Agent returned a 2xx response code, usually 250, which in RFC 5321 means "Requested mail action okay, completed." The receiving server has taken custody of the message and is now responsible for what happens next. Acceptance is a real handshake. Delivery is just "nobody yelled."

Inboxing is what you actually care about. The message landed in the primary inbox tab where humans see it, not in Spam, not in Promotions, not in Updates, not in some quarantine folder Microsoft invented last quarter. Inboxing is the only one of the three that correlates with opens, clicks, replies, and revenue.

Here is the trap. You can hit 99% delivery and 99% acceptance and still inbox at 40%. The receiving server accepts the message with a clean 250 OK, then its filters decide internally to route it to Spam. From the SMTP log on your side, that looks identical to a message that hit the inbox. Your ESP cannot see folder placement because folder placement happens after the handoff, inside the mailbox provider's filtering pipeline. This is why ESP dashboards lie by omission. See what "delivered" actually means in ESP reports for the longer version of that argument.

A concrete example. You send 10,000 emails to a Gmail list. Your ESP reports 9,950 delivered (99.5%). Of those, maybe 9,900 were accepted with a 250 response. Of those 9,900, Gmail's filter routed 4,000 to the Promotions tab, 1,500 to Spam, and 4,400 to the primary inbox. Your real inbox placement rate is 44%, not 99.5%. The gap between those two numbers is where most senders lose revenue without knowing it.

How do you measure the gap? You cannot get it from the SMTP log. You need either seedlist testing (sending to a panel of monitored addresses across providers and reading where each one landed) or panel data from a provider with real user inboxes. Both have limits. Seedlist testing gives you a directional read, not a per-recipient truth, because placement can vary across recipients at the same provider based on individual engagement history.

Why this matters in practice. If you are reporting deliverability to a CFO or a CMO using the ESP's "delivered" number, you are reporting a vanity metric. Gmail, Yahoo, and Microsoft all run separate filtering layers after acceptance, and the factors driving that decision (sender reputation, engagement, authentication, content) sit outside the SMTP handshake entirely. A campaign with 99.5% delivery and 30% inbox placement is a failing campaign dressed up as a success.

The shorthand we use with clients. Delivery is whether the post office took the letter. Acceptance is whether they signed for it. Inboxing is whether the recipient actually opened the envelope instead of dropping it in the junk pile by the door. Only the third one pays.