Delivery investigations draw on evidence from receiving mailboxes as well as the sending system. A sender alone cannot observe every recipient's destination folder.
What does delivery rate measure?
Delivery rate measures the share of messages accepted by receiving servers under the reporting system's counting rules.
At the end of a successful SMTP message transfer, a 250 reply confirms acceptance. A 4xx reply indicates a temporary failure. A 5xx reply indicates a permanent failure for that request.
For example, if a report counts 990 accepted messages out of 1,000 attempted messages, its delivery rate is 99 percent. That calculation does not say where the accepted messages landed.
Temporary failures can later become deliveries or final failures. Compare reports over the same observation window so pending retries do not distort the result.
What does deliverability measure beyond acceptance?
Deliverability concerns whether mail reaches the intended inbox rather than being blocked, lost or placed in spam.
A provider's acceptance response does not identify the destination folder. Inbox placement is therefore a separate measurement from the response recorded during SMTP delivery.
A Promotions tab is part of Gmail's inbox. It is a different placement from the spam folder, even when the recipient expected the Primary tab.
Acceptance can also be followed by a delivery failure. Keep subsequent bounce events when assessing a message's final outcome.
Why can a high delivery rate hide poor placement?
The receiver can accept most messages. It can then filter many of those accepted messages into spam.
Suppose the server accepts 990 of 1,000 messages. If 300 accepted messages reach spam, the delivery rate remains 99 percent. Only 690 reached an inbox in this example.
Delivery rate records the receiver taking responsibility for messages. Placement records where accepted mail went. Reporting the first as the second misstates what was measured.
Which signals can explain the gap?
Delivery errors, authentication results, complaints and mailbox placement each reveal a different part of the problem.
| Signal | What it can establish | What it cannot establish |
|---|---|---|
| SMTP acceptance | The receiver accepted the transfer | Inbox placement |
| Bounce or deferral | A permanent or temporary delivery failure | The outcome of unrelated accepted messages |
| Recipient complaint | A reported spam action | Every automatically filtered message |
| Authentication result | Whether the checked identity passed | Whether the recipient wanted the message |
| Seed placement | Where a test mailbox's copy landed | Placement for every real recipient |
Google's dashboard guidance warns that complaint rates can look low when automatic filtering reduces inbox deliveries. Low complaints alone therefore do not establish healthy placement.
Opens and clicks are supporting signals. Image loading and automated link checks can produce events without a person reading the message.
How do you measure placement with Bird?
You use Inbox Insights to compare measured mailbox placement with your own sending results.
A seed account is a test mailbox whose folder can be inspected. A seed test sends your message to those accounts and reports the observed placement by provider.
Panel measurements observe placement in a group of real mailboxes. Bird's placement view separates seed and panel measurements from your delivery, bounce and engagement data. Missing placement data does not mean zero inbox delivery.
The Inbox Insights guide explains domain tracking and test setup. Select provider coverage relevant to your recipients because an unmeasured provider has no placement result.
A seed test is a sample. Its mailbox history and behavior may differ from your subscribers', so report its scope with the result.
Which numbers should you report?
Report delivery rate and measured inbox placement separately, with their coverage and observation periods. They measure different stages and can cover different messages.
Investigate a delivery-rate decline through rejection and deferral evidence. If acceptance remains high but placement falls, examine authentication, recipient expectations and provider filtering signals.
Falling opens alone do not prove a placement problem. Confirm the change with other evidence before attributing it to spam filtering.
- Use delivery rate to report server acceptance and identify transfer failures.
- Use measured placement to report where sampled mailbox copies landed.
- Use provider diagnostics and complaints to investigate causes without treating them as placement measurements.
In short
Acceptance does not identify the folder.
A receiving server can accept a message and place it in spam.
Report delivery and placement separately.
Label each rate with its denominator, provider coverage and measurement period.
Placement measurements sample mailboxes.
Seed accounts show where their copies landed, without proving placement for every recipient.
Provider reports explain different failures.
Complaint and authentication reports complement delivery logs and measured placement.