Inbox Placement Seed Tests: How to Run One and Read It Honestly

An inbox placement seed test sends one real campaign to a list of monitored mailboxes ("seeds") at Gmail, Outlook, Yahoo and other providers, then reports where each copy landed: inbox, a tab, spam, or nowhere. It is the fastest way to see whether a provider is filtering a specific message. It measures what happened to those test accounts, not to your recipients, so read it as a signal rather than a score.

Most of the value comes from running it properly and knowing which conclusions it can support. This guide covers both. For diagnosing Gmail spam placement once a test shows a problem, see why emails go to spam in Gmail.

How a seed test works

The testing tool owns a set of real mailboxes, gives you their addresses and a test ID, and reads the mailboxes after you send. GlockApps' tutorial describes the flow: the tool provides "the unique id string and the test list", you "paste the id in any place in the message Subject line or body", and you send the message to the seed list. You can send through your own email service provider, or let GlockApps send it from a connected mailbox.

The ID is how the tool matches seed copies to your test. GlockApps warns that if it is missing, you get an "empty report that says 'Waiting for data…'"

Results come back as four rates in GlockApps' report: inbox rate, tabs rate, spam rate and missing rate. Other tools use similar categories under different names.

Running a test that means something

A seed test is only useful if the message is identical to what your real recipients get, sent the same way. Every variable you change makes the result describe a different email.

1. Send from the real platform, account and IP. A test sent from your personal inbox measures your personal inbox's reputation.
2. Use the real From domain and the same authentication. Check SPF, DKIM and DMARC first with the deliverability audit, so you are not testing a known configuration fault.
3. Use the real content, including links and tracking. Links are part of what filters inspect. If your campaign uses a tracking domain, the test should too. See custom tracking domains.
4. Put the test ID where the tool asks, and keep the rest of the subject line as it will be sent.
5. Send seeds in the normal flow, not as a separate blast at a quiet hour from a different sending pool.
6. Test more than once. One run is one data point. Repeat with the same message on another day before rewriting a campaign around one result.

For cold email

If you send from several inboxes, test from the inbox you are worried about, not your best-performing one. A test from one mailbox tells you about that mailbox, its IP and its domain. If you rotate across many inboxes and domains, run a test from each domain rather than assuming one result covers them all.

Reading the results

Read by provider first, then by category, and treat "missing" as the most urgent result. An overall inbox percentage blends very different mailbox systems into one number that matches none of them.

Result	What it usually means	What to check
Inbox	That provider accepted and filed it as wanted mail for an account with no history with you	Nothing urgent. Compare with real engagement data
Tab (Gmail Promotions, Updates and so on)	Accepted and not spam, but classified as a category	Whether this matters for your mail type. See the Gmail Promotions tab
Spam	Accepted, then filtered	Authentication, the reputation of your domain and IP, link domains, then content
Missing	Never arrived in the inbox or spam folder	Your platform's bounce and deferral logs for that provider

A few patterns are worth recognising:

• Spam at one provider only points to that provider's view of your reputation. If it is Gmail, Postmaster Tools has the reputation side of the story; see reading Google Postmaster Tools.
• Spam everywhere points at something every filter can see: missing or failing authentication, a blocklisted domain or link, or a sending IP with no reverse DNS. Check your domain and sending IP against the major blocklists.
• Missing at one provider is usually a rejection. Find the SMTP response in your logs, because it states the reason.
• Inbox on seeds while real results are falling is the case seed tests cannot explain, which is the next section.

What a seed test cannot tell you

Seed mailboxes behave like brand-new recipients, and real recipients do not. Google said in 2015, on the official Gmail blog, that "not all inboxes are alike. So while your neighbor may love weekly email newsletters, you may loathe them. With advances in machine learning, the spam filter can now reflect these individual preferences."

If a filter weighs how each person treats your mail, an account that never opens, replies to, or rescues your messages from spam is not a stand-in for one that does. That cuts both ways:

• Engaged recipients may see better placement than the seeds. People who reply and move your mail out of spam are teaching their own filter something the seed accounts never will.
• Disengaged or complaining recipients may see worse. A seed that lands in the inbox says nothing about how the filter treats people who keep marking you as spam.

Other limits follow from how the test is built:

• The sample is small. A seed list covers a handful of accounts per provider. Treat a swing of one or two seeds as noise.
• Business mailboxes vary per organisation. A Microsoft 365 or Google Workspace seed reflects that tenant's settings. Your recipients' companies may run different policies or third-party gateways that the seed account does not.
• It tests one message, once. It cannot show reputation trends over weeks, which is where most placement problems actually live.
• It does not test your list. Placement depends partly on how many of your real recipients bounce, complain or never engage. A seed list contains none of the dead addresses or spam traps your real list may contain.

Where seed tests fit with other signals

Use seed tests to check a specific message or change, and use provider data and your own results for the ongoing picture.

Question	Best evidence
Did my new DNS setup or template break placement?	Seed test before and after
Is Gmail's view of my domain getting worse?	Google Postmaster Tools, over weeks
Are real people receiving and reading this?	Replies, clicks and conversions by provider
Are my bounces coming from bad addresses?	Your platform's bounce logs, and list verification before sending

The last row is the part a seed test never touches. If real campaigns are bouncing, verifying the list first removes addresses that no longer exist and flags catch-all domains as risky; the free email checker shows the result for a single address. A clean seed test on top of a list full of dead addresses is reassurance you have not earned.

Short version

1. Send the test exactly like the real campaign, with the test ID where the tool asks.
2. Read results per provider, and investigate "missing" first.
3. Repeat before acting on a single run.
4. Do not read inbox placement on seeds as placement for your list. Gmail personalises filtering, and seeds have no history with you.
5. Pair seed tests with Postmaster Tools and real engagement data for the trend, and keep the list itself clean.