Two numbers get called the same thing, and confusing them is why a sender can watch a healthy dashboard while their campaign sits in the spam folder. One is reported by the platform that sent the mail. The other is decided by the mailbox provider that received it, and never reported back.
What does a delivery rate actually measure?
That a receiving server accepted the message.
When your platform hands a message to Gmail's servers, Gmail answers. A 250 means accepted. A 5xx means rejected outright. A 4xx means try again later. Delivery rate is the share of your messages that got the first answer rather than the other two, and it is the only one of the two numbers your sending platform can compute, because it is the only one that happens in a conversation the platform is part of.
Google reports the mirror image of this figure in Postmaster Tools, and its wording shows how narrow the measurement is:
The Delivery Errors dashboard displays the percent of all authenticated messages (SPF or DKIM) that were rejected or that temporarily failed, compared to all authenticated messages.
Rejected or temporarily failed. Nothing in that sentence is about a folder.
So what is deliverability?
Where the message went after it was accepted.
A mailbox provider takes delivery of a message and then decides what to do with it: inbox, spam folder, a tab, or a silent drop. That decision happens after the SMTP conversation has closed, on the provider's own infrastructure, using signals the sender cannot see. Bird's own documentation draws the line in one sentence:
Delivery rate tells you a receiving server accepted your mail. It does not tell you whether the message landed in the inbox or the spam folder.
So deliverability is the practical question, and delivery rate is the one that is easy to measure. That mismatch is the whole problem.
Why can my platform not just tell me?
Because nobody sends the answer back.
There is no protocol by which Gmail tells a sender "this one went to spam". The feedback that does exist is thinner than people expect and each piece measures something different:
| What you can see | What it actually tells you |
|---|---|
| Delivery rate and bounces | The receiving server accepted, refused, or deferred the message |
| Complaint rate | How often a recipient who saw the message pressed the report-spam button |
| Domain and IP reputation | The provider's running judgment of you, as a rating rather than a number |
| Opens and clicks | How recipients behaved, distorted by privacy proxies that fetch images |
Complaint rate is the one that misleads most often, because it is counted only where a recipient could act. Google is explicit that the figure covers messages "delivered to engaged recipient's Inbox and then marked as spam by the recipient", and warns that a heavily filtered sender sees a low number for the wrong reason:
If Gmail automatically sends a significant number of your messages to spam, the rate shown in the dashboard might seem low, because recipients get fewer of your messages in their Inbox.
A complaint rate that falls while placement gets worse is the same illusion as a delivery rate that stays at 99 percent while the campaign dies.
How do I measure the gap?
With a second source, because the sending side genuinely cannot supply one.
Two approaches exist, and they answer different halves of the question.
- Seed accounts hold real mailboxes at the major providers and record which folder a copy of your send lands in. That gives you placement per provider rather than one blended figure. Bird's Inbox Insights is this, run against your own sends.
- The provider's own dashboards report their judgment of you rather than your placement. Google Postmaster Tools publishes domain and IP reputation, the user-reported spam rate, and an authentication breakdown, all for mail to Gmail only.
Neither replaces the other. Seed data tells you where mail is landing today; reputation tells you which way it is heading.
Which number should I report?
Both, labelled honestly, and never delivery rate alone.
Delivery rate belongs on the dashboard because a fall in it is a real and urgent signal: something is being rejected. It just cannot carry the weight of a deliverability claim. A team that reports "99.4 percent delivered" as though it meant "99.4 percent reached the inbox" has quietly promised something nobody measured.
The practical framing is that delivery rate is a hygiene floor and deliverability is the outcome. If delivery rate drops you have a bounce, an authentication break, or a block to fix, and why emails go to spam walks the causes. If delivery rate is fine and engagement is falling, the problem is placement, and no amount of staring at the sending log will show it to you.
For the practices that move the second number rather than the first, how to improve email deliverability is the longer playbook.