Inbox placement test for email

How to read the results

Each test mailbox gets one result. A found result tells you where the marked message was in that mailbox; no result in time does not prove the message was rejected.

• Inbox

  Found in the inbox. Gmail’s Promotions and Updates tabs and Outlook’s Other tab are inbox categories, so they count as Inbox, not Spam. ^[3]

• Spam / junk

  Found in the spam or junk folder of that test mailbox.

• Other folder

  Found, but in a folder that is neither the inbox nor spam.

• Pending

  The test is still inside its results window and that mailbox has no result yet.

• Not observed

  The mailbox was fully checked after the window closed and the marked message was not found. This does not prove the message was rejected.

• Reader unavailable

  Sky could not complete a check of that mailbox after the window, so there is no result for it.

• Ambiguous

  A matching message was found but could not be tied reliably to one test or folder, so it is not counted as Inbox or Spam.

Two inbox rates, read together

Against all assigned mailboxes: Inbox results divided by every test mailbox in the test, including ones with no result. Against found results: Inbox results divided only by mailboxes where the message was found in a folder (Inbox, Spam / junk or Other folder). The first is stricter; the second ignores missing results. Neither predicts what every subscriber will see.

What a seed test can and cannot tell you

A seed test sends a marked message to test addresses at mailbox providers and records where those specific copies were found. It is useful to compare providers, or to check a campaign version before a wider send. ^{[1] [10]}

It is a sample, not a miniature copy of your list. Test mailboxes do not have each real subscriber’s history with your brand, and providers weigh many signals beyond one message. Twilio notes that seed accounts lack the engagement history used in filtering; M3AAWG describes sending-domain history as one factor alongside content and IP reputation. ^{[1] [9]}

Recipient behaviour matters too. Google says spam reports from users can lower a domain’s reputation, and Gmail learns tab preferences from what recipients do. Microsoft says SNDS’s aggregate filter results do not map directly to Inbox or Junk for each person, because user settings can change how mail is handled. ^{[2] [3] [4]}

With a small number of test mailboxes, one result can move a percentage a lot. A seed test also cannot explain the cause on its own. Use it alongside provider data and your ESP or MTA logs. ^{[1] [4] [5]}

When to run a test

• Before an important campaign. Check the exact version you plan to send, as one part of a wider pre-send review. ^[10]
• After authentication or DNS changes. Compare folders, then check authentication and delivery-error data in Google Postmaster Tools. ^[5]
• After an ESP, IP or domain move. A new sending setup may need time to build history; read the test next to your logs and reputation data. ^[9]
• After a big change to content or sending pattern. Use a test to spot a difference, not as proof of what caused it. ^{[1] [9]}
• When one provider’s results drop. Compare test results with that provider’s complaint, reputation and SMTP response data. ^{[4] [5] [6] [8]}

Use it with these signals

• Google Postmaster Tools. Gmail spam rate, authentication and delivery errors. It covers mail to personal Gmail accounts, is not real time and can leave out low-volume days. ^[5]
• Microsoft SNDS. IP-level activity, aggregate filter results and complaint reports for IPs you control. Microsoft says the filter result is not direct Inbox or Junk placement. ^[4]
• Yahoo Complaint Feedback Loop. For enrolled DKIM domains, Yahoo reports when a recipient marks mail as spam. Suppress that recipient and review what they were sent. ^[7]
• Bounce and deferral logs. Keep full SMTP responses. Yahoo separates temporary 4XX errors from permanent 5XX errors and lists causes including complaints, authentication and invalid recipients. ^[8]

Who this is for