Email spam test: how to tell if your cold emails land in spam

You launched a campaign on a new cold email domain and the replies are thin. The question is always the same: is the offer weak, or is the mail landing in spam? An email spam test answers that, but only if you know what each test can and can't tell you. Here are four ways to check whether your cold emails land in spam, in the order we'd run them.

The free email spam test: your own inboxes

The cheapest email spam test is the one you can run in ten minutes. Create a few accounts you control: a personal Gmail, a Google Workspace inbox on a different domain, an Outlook.com account and a Microsoft 365 inbox if you have one. Then:

1. Send from the exact inbox that sends your campaigns, not from a different one.

2. Use the same copy, the same links and the same signature as the real campaign.

3. Check where each message landed: inbox, Promotions tab, spam, or nowhere.

4. Reply to a couple of them, then send again. Providers treat mail differently once there's a conversation.

Gmail and Microsoft often disagree, so test both. If a message goes to spam on a seed account you control, that's a real signal. If it lands fine, that is weaker evidence than it feels like, which brings us to the limits.

Placement tests and their limits

Placement tests (sometimes called an email deliverability test) work by sending your message to a set of seed addresses across providers and reporting where each one landed. They are useful for one thing: spotting that something is broken. A sudden jump from inbox to spam across every seed is a real alarm.

They are poor at telling you your actual inbox rate, for three reasons.

• Seeds aren't your prospects. A provider decides placement partly on how a specific recipient has treated similar mail before. A seed account with no history behaves differently from a VP whose inbox is full of cold email.

• The sample is small and fixed. A handful of accounts can't stand in for your whole list, and the same accounts see you again and again.

• Seed traffic can look odd. Sending to accounts that never open, click or reply is not a normal engagement pattern.

Use seed tests as a smoke alarm, not a thermometer. Email deliverability tools covers what the tooling does well and badly.

Read the headers

Headers show what the receiving server decided about your message. In Gmail, open the message, use the menu and choose Show original; Google's page on tracing an email with its full headers explains the steps. Look for the authentication results:

• spf=pass

• dkim=pass

• dmarc=pass

Anything other than pass on any of the three is your first fix, ahead of copy or timing. Also check that the domain in the passing results is your sending domain, not just some service's. SPF, DKIM and DMARC for cold email explains what each result means and how to repair a failure.

Headers won't tell you why a message went to spam, but they rule out the most common cause in one look.

Check Google Postmaster Tools

Google Postmaster Tools shows Gmail's view of a domain you verify: domain reputation, spam rate, authentication and delivery errors. Google's sender guidelines say to keep the spam rate below 0.1% and never let it reach 0.3%. If you can see that number, you have the single most useful deliverability metric for Gmail recipients.

Two cautions. Postmaster Tools may show little or nothing for a domain that sends low volume to Gmail, so an empty dashboard on a new domain isn't a clean bill of health. And it covers Gmail only. For Microsoft, Microsoft SNDS offers a similar view of IP reputation, though it is built around IP addresses rather than domains. For more on what reputation means, see domain reputation.

Read your reply and bounce patterns

Your real campaign data is the only test that uses your real prospects. Split it by recipient provider and look for lopsided results.

• Bounces with a reason attached. Spam blocks and policy failures often come back as bounces. Our bounce code library explains each code and whose problem it is. In SendingOS, every bounce is read and sorted into hard, soft or block with its reason and SMTP code, and the thread is labelled Bounced so you can filter to just those.

• One provider goes quiet. If replies from Gmail addresses keep arriving and replies from Outlook addresses stop, that points at Microsoft placement, not at your offer.

• Replies stop across the board after a change. A new list, a volume jump or a DNS edit in the days before is the likeliest cause.

• Out-of-office and auto-replies still arrive. Automated replies prove the message reached a mailbox, even if no human answered. If those vanish too, placement is a suspect.

There is no benchmark reply rate to compare against, so compare your campaign with itself: this week against last week, Gmail against Outlook, the old list against the new one.

Putting the tests together

No single test settles it. Run them as a sequence:

1. Seed test from the sending inbox to Gmail and Outlook.

2. Headers on every seed message: all three authentication checks pass.

3. Postmaster Tools, if the domain has the volume to show data.

4. Reply and bounce patterns by provider, week over week.

If the seeds land and the headers pass but replies are still low, the problem probably isn't spam. If seeds hit spam or authentication fails, fix that before you change a word of copy. How to stop emails going to spam lists the fixes in order, and the whole deliverability picture shows where this test fits.