If you send a test email from a new mailbox and it lands in spam, it is natural to suspect your setup. In most cases we review, the setup is fine and the cause is the test itself: the wording of the email, the number of messages sent, or the type of inbox that received it. This article shows how to run a fair test, how to read the result, and what to change first.
Use a neutral email. The fastest way to separate "my mailboxes" from "my email copy" is to send a short, plain-text message with nothing a filter can react to:
No links, images, attachments or tracking.
No signature, phone number or postal address.
No prices, amounts, offers or sales wording.
A natural subject line and one or two sentences, for example a simple question.
In our experience, a neutral email from healthy mailboxes usually lands in the inbox 90 to 100% of the time. A sales-style email can score far lower from the same mailboxes on the same day. In one case, a campaign-style email reached 0 of 13 Microsoft 365 test inboxes, while a neutral email sent from the same mailboxes to the same seed list reached 90%.
This neutral email is a diagnostic, not a template for your campaigns. Your real emails can still include a short signature and an unsubscribe option as your compliance rules require. The test only helps you find which elements to add back, and how.
Change one thing at a time. Keep the same sender mailboxes and the same recipients, and change only the email. If you change the copy, the recipients and the tool at once, you cannot tell which one made the difference.
Send enough messages, from every mailbox you care about. One email to your own Gmail address shows how that one inbox treated that one message. It does not show how your mailboxes perform overall. In one example, a customer sent a single plain-text message from each of 10 mailboxes to one Gmail account. Seven landed in spam and three in the inbox, even though all ten passed SPF, DKIM and DMARC. A Warmforge placement test with the same text, sent to many inboxes, reached 94% and 86% inbox.
Match the recipients to your real audience. Personal inboxes (Gmail, Outlook.com, Yahoo) filter more strictly than business inboxes, especially for senders with little history. Our mailboxes are built for B2B outreach, so test against business inboxes, or run the Warmforge placement test, which is designed for this. See Placement Test on Warmforge.
Test after warm-up has had time to work. Mailboxes and domains set up within the last day or two have no sending history. It is common for the first messages to land in spam at Gmail or Yahoo, even when authentication passes. Keep warm-up running for a couple of weeks, then test.
The Warmforge Placement Test groups results into four bands: Best (95–100%), Good (90–94%), Not Good (70–89%) and Bad (0–69%). These bands describe a result, but they do not tell you how many messages it is based on. Use the points below to judge how much a single result means.
One message is an anecdote. Mixed results from a handful of messages are normal. If you get 3 inbox and 7 spam from single sends to one recipient, repeat the test with more recipients before drawing a conclusion.
Each seed moves the score a lot. In a test with 5 Google and 5 Microsoft inboxes per mailbox, a single message changes that mailbox's result by 10 points, and one inbox type by 20 points. A difference between 70% and 90% on one mailbox can be a single message.
Look at Google and Microsoft separately. Google and Microsoft filter differently, so a mailbox can look healthy at one and weak at the other. Averages hide this.
"Missing" is not the same as "spam". A message that is missing from the test inbox has not reached it yet or was not found there. Check the result again later, because delays are common, and note the status next to each message.
Check authentication, but do not stop there. SPF, DKIM and DMARC passing means your DNS is correct. It does not guarantee the inbox, because providers also weigh reputation, sending history, engagement and content.
A test is an estimate. No tool can show exactly where every future email will land, so treat a result as a guide, not a guarantee.
Business and personal inboxes differ. Business inboxes are usually more forgiving than personal ones, so a good result at business inboxes does not promise the same at personal inboxes.
Look at the trend, not one run. Repeat the test after a few days of warm-up. Mailboxes that improve between tests are usually just building reputation.
Do not judge by a test tool you have not checked. Seed lists and recipient types differ between tools. Look at which recipients were used and whether they were personal or business inboxes before comparing results from different tools.
When a neutral test reaches the inbox but your real email does not, the cause is almost always the content. Work through this list in order, changing one item per test:
Signature. Keep a simple signature in your real emails, but when you diagnose a problem, remove it entirely and add it back line by line. Phone numbers, postal addresses, logos and social links are common triggers.
Links and tracking. Remove links, shortened URLs and tracking, then add back only what you need.
Subject line. Avoid money, guarantees, urgency, ALL CAPS and emojis.
Body text. Remove prices, amounts and strong offers, and keep emails short and personal.
Repetition. Identical text sent to many people can be filtered quickly. Use spintax and personalization so that messages differ from one another. You can also check your copy with the Salesforge Spam Checker.
Contact quality. Verify your leads before sending, and keep daily volume low and gradual when you start.
If the neutral email also lands in spam on both Google and Microsoft across several mailboxes, even after a couple of weeks of warm-up, contact us (see below).
Contact support if:
a neutral plain-text email still lands in spam after warm-up has run for about two weeks;
one mailbox or domain performs much worse than the others in repeated tests;
real campaigns land in spam, but the same message sent from the same mailbox directly does not; or
messages are rejected or deferred and you have the error text.
To help us look into it quickly, please send:
the affected mailbox addresses and domains;
the test report link or ID, with the date and time of the test;
the exact email as sent, including subject, signature and any links;
how you sent it (campaign platform or webmail), and the recipients used, with the provider and whether they were personal or business inboxes;
when warm-up started, which tool runs it, and whether it has run every day;
whether you have sent real campaigns, and your daily volume per mailbox;
the full original headers or Message-ID of one message that landed in spam or junk, and any bounce message text.
With these details, we can run our own placement test and tell you whether the issue is the email, the recipients, or something we need to look into.