A seed test is an email deliverability test in which you send a campaign to a controlled list of test inboxes, called seed addresses, at different mailbox providers. It shows whether the message reaches the inbox, spam folder, a Gmail tab such as Promotions, or fails to appear before you send to your real subscribers.

Seed test definition in plain language

Think of a seed test as a preflight check for an email campaign. Rather than immediately sending a newsletter, product announcement, password-reset template, or lifecycle message to thousands of recipients, you send an identical copy to a purpose-built set of inboxes.

Those inboxes are distributed across providers such as Gmail, Yahoo Mail, Outlook, iCloud Mail, and business-hosted domains. A seed-testing service monitors the test inboxes and reports where the message appeared. The result is commonly called an inbox placement result.

A seed test answers a practical question that ordinary delivery logs cannot answer: did the email merely reach the receiving provider, or did it reach a visible inbox location? An SMTP delivery event can mean that a provider accepted the message, even if it later routes that message to spam. Seed testing helps distinguish acceptance from likely inbox placement.

It is especially useful before consequential sends, including:

  • A high-volume newsletter or promotional campaign
  • A major product launch
  • A domain or IP migration
  • A new sending-domain rollout
  • A campaign with a new template, link domain, or content style
  • A re-engagement campaign aimed at inactive subscribers
  • A seasonal send where timing and visibility matter

Seed testing is not a guarantee that every subscriber will receive an email in the inbox. Mailbox providers make filtering decisions using many recipient-specific signals, including prior engagement, individual spam reports, user-level rules, and historical interactions. Still, a seed test provides an early, controlled signal that can reveal broad authentication, reputation, configuration, or content issues before a full send magnifies them.

Why a seed test matters for email deliverability

Email deliverability is not the same as email delivery. Delivery usually means a receiving mail server accepted a message. Deliverability is broader: it concerns whether the message successfully reaches a place where the intended recipient is likely to see it, usually the inbox rather than spam.

That difference matters because a campaign can report a very high delivery rate while producing disappointing opens, clicks, purchases, or product usage. If messages are accepted but filtered into spam, delivery logs may not show the entire problem. A seed test adds an inbox-placement perspective.

Catch systemic problems before a campaign scales

The most valuable seed-test findings are usually the obvious failures that affect many recipients at once. For example, a test can show that messages from a newly configured domain are landing in spam at multiple providers. It can also expose an accidental change to the visible From address, a missing authentication record, an unapproved link-tracking domain, or a malformed email header.

Finding one of these issues before a send can prevent a poor campaign outcome and reduce the risk of generating negative engagement signals. A large send that causes a spike in spam complaints, deletes without reads, or disengagement can make subsequent messages harder to place well.

Identify provider-specific placement differences

A campaign may perform differently at different mailbox providers. Gmail, Yahoo Mail, Microsoft consumer mailboxes, Apple-hosted mailboxes, and corporate gateways do not all use the same filtering systems or the same reputation data.

For example, a seed test could show these patterns:

  • Gmail seeds receive the message in Promotions.
  • Yahoo Mail seeds receive it in the inbox.
  • Outlook seeds receive it in junk.
  • A corporate test inbox rejects or quarantines it because of a security gateway rule.

That pattern does not prove that every user at each provider will receive the same result. It does, however, narrow the investigation. If only one provider is affected, the team can examine that provider's authentication, reputation, content, complaint, and infrastructure signals more closely instead of changing everything at once.

Protect campaign performance

Campaign success depends on visibility. A well-designed offer or update cannot perform if a substantial portion of the audience does not see it. By testing placement before a full send, teams can decide whether to proceed, delay, alter the message, segment the audience, or fix infrastructure first.

For transactional email, seed tests are useful when a template or delivery setup changes. A receipt, password reset, verification email, or account alert should be highly reliable because recipients expect it and may be unable to complete a workflow without it. For marketing email, seed testing can help protect revenue-generating sends, but it should be combined with audience-quality and engagement analysis.

What a seed test measures

A seed test is primarily an inbox-placement measurement, not a single universal score. Different tools can present results differently, but the underlying outcomes normally include inbox, spam or junk, tab placement, missing, and pending.

Common seed-test outcomes

Inbox: The test message appears in the main inbox or a provider-defined inbox location. At Gmail, some tools further distinguish between Primary, Promotions, Social, and other tabs.

Spam or junk: The message arrived at the mailbox but was filtered to the spam or junk folder. This is often the most urgent negative result because the recipient is less likely to see it.

Missing: The seed-testing system did not find the message in the expected time window. Missing can indicate delayed delivery, a provider rejection, a routing issue, filtering behavior that prevents visibility, or a test-matching issue. It should be investigated rather than automatically treated as a hard bounce.

Pending: The testing system is still waiting for the mailbox provider or seed inbox to report an outcome. Email delivery can be asynchronous, especially during temporary deferrals or provider delays.

Authentication results: A testing report may show whether SPF, DKIM, and DMARC checks passed. These checks are separate from placement but strongly relevant to it.

Spam-filter findings: Some tools also inspect message content, HTML, links, and headers for conditions that could increase spam-filter risk. This is helpful diagnostic information, but it is not a substitute for actual placement observations.

Inbox placement rate calculation

A seed test can be expressed as an inbox placement rate. The simplest calculation is:

Inbox placement rate = (number of seed messages placed in inbox ÷ number of completed seed results) × 100

The denominator should be clearly defined. Many teams exclude pending results while a test is still running. Whether missing messages are included depends on the reporting convention, but excluding them without documenting the choice can make a result look better than it is.

Worked numeric example

Suppose a campaign is sent to 100 seed addresses. After the test completes, the results are:

  • 72 messages in inboxes
  • 18 messages in spam or junk
  • 8 messages missing
  • 2 messages still pending

If you calculate a provisional rate using the 98 completed results, the inbox placement rate is:

72 ÷ 98 × 100 = 73.47%

Rounded, the campaign has a 73.5% inbox placement rate among completed seed results.

The spam placement rate is:

18 ÷ 98 × 100 = 18.37%

Rounded, that is an 18.4% spam placement rate.

The missing rate, when measured against all 100 seeds, is:

8 ÷ 100 × 100 = 8%

This result should not be interpreted as “73.5% of all subscribers will see the campaign in their inbox.” It is a directional sample of placement across the tested seed accounts. The stronger conclusion is operational: a campaign with nearly one in five completed seed results in spam deserves investigation before a major send.

What a seed test does not measure

A seed test is valuable precisely because it is controlled, but that control creates limitations. Seed inboxes are not a replica of a live subscriber list.

It does not reproduce real subscriber engagement

Mailbox providers use engagement signals. Depending on the provider and recipient, those signals can include whether people open mail, move it from spam to inbox, mark it as spam, delete it without reading, reply, add a sender to contacts, or repeatedly ignore it.

Seed accounts generally do not behave like a normal audience. They may have limited engagement histories, and they do not make purchases, use a product, or interact naturally with every campaign. Therefore, seed placement should be treated as a warning system and diagnostic input, not a forecast with mathematical certainty.

It does not replace delivery, complaint, and engagement data

A healthy deliverability program uses multiple signals together:

  • Seed-test inbox and spam placement
  • SMTP deliveries, deferrals, and permanent failures
  • Hard-bounce and soft-bounce trends
  • Spam-complaint rates
  • Opens, clicks, conversions, and downstream product activity
  • Domain and IP reputation indicators where available
  • Authentication and DMARC reporting
  • Audience growth, consent, and unsubscribe trends

For example, a campaign could place well in seed inboxes but underperform with real recipients because the audience was not expecting it or the content was irrelevant. Conversely, a seed test might show mixed placement while a carefully engaged recipient segment still performs well. The right interpretation combines the seed test with real campaign evidence.

It cannot make a bad list safe

Seed testing is not a workaround for weak permission practices. If a sender buys a list, sends to stale addresses, repeatedly targets unengaged contacts, or makes it difficult to unsubscribe, no content tweak can reliably solve the underlying problem.

Mailbox providers want recipients to receive mail they asked for and value. Yahoo's sender guidance emphasizes sending timely, relevant email to active and engaged recipients and avoiding purchased lists. Seed testing can reveal risk, but list quality and recipient consent remain foundational.

How seed testing works in practice

A typical seed-test workflow has four stages: create or obtain a seed list, send an authentic campaign copy, match the received message to the test, and review placement results.

1. Build a representative seed list

A seed list is a set of monitored test addresses. A useful list includes mailbox providers that matter to your audience. If 60% of the active list uses Gmail, Gmail deserves meaningful representation. If the business sells to companies, include corporate domains or a service that tests common enterprise filtering environments where possible.

A representative seed list may be segmented by:

  • Mailbox provider
  • Country or region
  • Consumer versus business mailbox type
  • Provider-specific variations, such as Gmail tabs
  • Security gateway or hosted-domain environments

Do not assume that a handful of personal accounts you created years ago gives a reliable test. A do-it-yourself check can catch obvious rendering or routing problems, but professional seed lists are designed to provide broader coverage and managed observations.

2. Send the real message, not a simplified substitute

A seed test should reflect the actual sending conditions as closely as possible. Use the real From domain, sending domain, DKIM signing setup, link-tracking domain, subject-line strategy, HTML, text version, attachments if any, and sending infrastructure.

A simplified test from a personal mailbox may not reveal issues caused by production routing, link rewriting, IP reputation, or a different envelope sender. Likewise, changing the subject line after the test may reduce the relevance of the result if the original subject was an important filter signal.

For a campaign with personalization, test a representative rendered version. Make sure merge fields resolve correctly, the unsubscribe link works, and dynamic content does not create malformed HTML or suspicious-looking URLs.

3. Match the sent message to the test

Seed-testing systems need to identify which incoming message belongs to the campaign being measured. Header-based matching is generally more reliable than subject-line matching, particularly when messages have personalized or dynamic subjects.

Mailgun's Inbox Placement documentation describes the use of an X-MG-IBP-ID header for header-based matching. In its Send API workflow, the o:inbox-placement-test-key parameter can trigger a placement test alongside a normal message send. The parameter value serves as an idempotency key, so duplicate requests using the same value do not continually create duplicate tests within the key's retention period.

The broader lesson applies regardless of provider: use a stable, unique campaign identifier where the testing system supports it. That reduces ambiguity when several campaigns have similar subjects or when a campaign contains recipient-level personalization.

4. Wait for results, then investigate the pattern

Placement results are often available within minutes, but some can take longer because mailbox providers may defer, delay, or process messages asynchronously. Avoid reacting to a partial result as though it were final.

When the results settle, look for patterns rather than obsessing over one individual seed. A single spam result could be noise. A consistent spam result across a provider, a region, or an entire campaign variation is a stronger signal.

Common causes of poor seed-test results

A poor seed test usually does not have one magical explanation. Inbox placement is influenced by a combination of authentication, sender identity, infrastructure, audience behavior, message content, and sending patterns.

Authentication failures or misalignment

Authentication is a baseline requirement. SPF identifies servers authorized to send mail for a domain. DKIM adds a cryptographic signature that receiving systems can validate. DMARC applies domain-alignment rules and tells receiving servers how to handle mail that fails the relevant checks.

Google's sender guidelines require all senders to use SPF or DKIM, while bulk senders to Gmail must use SPF, DKIM, and DMARC. Google also advises keeping reported spam rates below 0.3%. Yahoo similarly emphasizes authentication with SPF, DKIM, and DMARC for senders, with requirements that vary by sender category.

Common authentication problems include:

  • DKIM signatures missing on one sending stream
  • A DKIM selector record that no longer resolves
  • SPF records that omit a legitimate sending provider
  • SPF records exceeding lookup limits
  • A visible From domain that does not align with authenticated domains for DMARC
  • A third-party system sending on behalf of a domain without proper setup
  • An email modification after signing that breaks DKIM

A seed test may show these issues directly in authentication results, but always confirm them by inspecting the raw message headers and DNS configuration.

Weak domain or IP reputation

Reputation is an accumulated assessment, not a switch that flips after one message. A new domain, a newly assigned IP, a sender with recent complaint spikes, or a sender with inconsistent traffic can face stricter filtering.

Reputation can decline when recipients complain, ignore messages, receive unwanted campaigns, or encounter mail that resembles phishing or spam. It can improve over time when a sender uses clear consent, sends predictable and wanted mail, removes unhealthy addresses, and focuses on engaged audiences.

A seed test cannot fully measure reputation because seeds do not behave like real recipients. But a broad spam-placement pattern, especially after a volume increase or infrastructure change, can be a sign that reputation deserves attention.

Sudden changes in sending behavior

Mailbox providers notice changes. A domain that normally sends a few thousand messages weekly and suddenly sends hundreds of thousands may appear riskier than one that grows steadily with a clear sending history.

Other meaningful changes include switching email service providers, moving to a dedicated IP, adding a new subdomain, changing the envelope sender, rotating DKIM keys incorrectly, or introducing a new tracking domain. These changes may be reasonable, but they should be planned, monitored, and tested.

Content, links, and formatting issues

Content alone rarely explains every spam placement result, but it can contribute. Risky patterns include misleading subject lines, excessive urgency, deceptive display text, broken links, image-only email, overly aggressive wording, unexpected attachments, or HTML that is malformed enough to resemble low-quality bulk mail.

Links deserve particular attention. A message can be authenticated correctly but lose trust if it uses a newly introduced redirect domain, links to a compromised destination, or displays one domain while sending recipients to another. Check every destination, including links inserted by analytics, personalization, URL shorteners, or third-party widgets.

Poor list quality and weak consent

A message sent to people who did not request it creates negative signals quickly. Stale addresses can bounce. Uninterested recipients can complain or disengage. Recipients who do not recognize the sender may report mail as spam even when the content is legitimate.

Use explicit permission, set accurate frequency expectations, and make the sender recognizable. Before a risky campaign, verify newly acquired or older addresses and suppress known hard bounces, unsubscribes, and complaints. Teams can use an email address verification tool as one part of list hygiene, but verification does not prove consent or interest.

How to improve a poor seed test

The best response to a poor seed test is a structured investigation. Do not make five unrelated changes and send again; doing so makes it impossible to know what caused the improvement or regression.

Start with the basics: inspect headers and authentication

Download or view the original message source from affected seed inboxes. Confirm:

  1. The visible From domain is the intended domain.
  2. SPF passes for the envelope sender where applicable.
  3. DKIM passes and uses the intended signing domain.
  4. DMARC passes with appropriate alignment.
  5. The return path, message ID, reply-to address, and tracking links are expected.
  6. TLS, DNS, and message formatting are correctly configured.

If these checks fail, fix them before revising design or copy. Attractive content cannot compensate for broken authentication.

Isolate one variable at a time

Create a disciplined test plan. If you suspect content, compare two versions with the same domain, audience conditions, and infrastructure. If you suspect a new tracking domain, test the campaign with and without the change where operationally appropriate.

Useful controlled comparisons include:

  • Existing sending domain versus new sending subdomain
  • Established template versus redesigned template
  • Existing link domain versus new link domain
  • A plain-text or minimal HTML version versus a complex marketing version
  • A normal audience segment versus a lower-engagement segment

Document the date, campaign identifier, source domain, volume, audience segment, template version, and result. Deliverability work becomes much more effective when it is treated as an experiment rather than a sequence of guesses.

Improve audience selection before increasing volume

If a campaign is intended for a broad audience but seed placement is weak, do not immediately compensate by sending more frequently or to more people. Start with recipients most likely to recognize and value the message: recent customers, active users, confirmed subscribers, or subscribers who recently interacted with similar campaigns.

This approach can reduce complaint risk and create healthier engagement signals. It also produces a more honest measure of campaign interest. If the message performs poorly among highly engaged recipients, the issue may be relevance or execution rather than deliverability alone.

Make subscription expectations clear

People are less likely to complain when they understand who is emailing them and why. Use an identifiable From name and address, state the brand clearly, honor the promised frequency, and provide a straightforward unsubscribe mechanism.

For bulk mail, one-click unsubscribe support and visible preference controls are operationally important. Google and Yahoo have sender requirements and recommendations around subscription management for qualifying senders. The implementation should be validated in the actual production message, not assumed from a template setting.

Warm up new infrastructure carefully

New domains and IPs need time to establish a sending history. The right process depends on the provider, volume, and audience, but the principle is consistent: begin with mail that recipients expect, prioritize engaged segments, grow volume deliberately, and monitor placement, complaints, bounces, and deferrals.

Do not treat warm-up as a trick to manufacture reputation. It is risk management for a real change in sending infrastructure. A sender that jumps immediately to a large cold or low-engagement audience can create poor signals faster than a seed test can prevent them.

Retest the production-like message

After fixing a likely cause, run another seed test with the same production-like conditions. Compare results by mailbox provider and by placement category. If the first test was poor and the retest improves, continue to monitor live-campaign outcomes because the real audience remains the ultimate signal.

For teams integrating sending into applications, the email API setup documentation should be part of the deployment checklist alongside domain verification, authentication, event monitoring, and template review.

Seed test best practices for developers and marketers

Seed testing works best when it is part of a repeatable release process rather than an emergency action taken only after a campaign goes to spam.

Use a pre-send checklist

Before a major send, verify the following:

  • The sending domain is verified and uses the expected From address.
  • SPF, DKIM, and DMARC are configured and passing.
  • The campaign uses production link-tracking and unsubscribe settings.
  • All personalization renders correctly for representative recipients.
  • Links resolve to secure, expected destinations.
  • HTML and plain-text versions are present and readable.
  • The intended audience has valid consent and appropriate recency or engagement.
  • The seed test includes important mailbox providers.
  • The team knows what result would pause a send and who can make that decision.

The final point matters. A seed test has little value if no one has defined an escalation path. For example, a team might proceed when results are normal, investigate when one major provider shows a material spam pattern, and pause when authentication fails or broad spam placement appears.

Test at meaningful moments

Not every small copy edit needs a full seed test. Testing is most valuable when risk changes materially. Good triggers include a new sender domain, a new ESP, a major template revision, a significant audience expansion, new link tracking, a reputation incident, or a high-stakes campaign.

For recurring newsletters, periodic tests can establish a baseline. If placement suddenly changes while the content and infrastructure remain stable, the team has a reason to examine reputation, audience composition, complaints, or provider-specific conditions.

Keep seed tests separate from recipient data

Seed addresses are testing infrastructure, not subscribers. Do not accidentally add them to customer segments, trigger lifecycle automations against them, or include them in conversion reporting. Label and isolate them in campaign systems.

Also avoid using seed tests to create misleading engagement. The objective is observation, not artificially generating opens or clicks. Real engagement should come from real recipients who chose to receive the mail.

Seed test alternatives and complementary tools

Seed testing covers an important gap, but it should not be the only pre-send or deliverability control.

Email previews and rendering checks

Rendering tests show how an email appears in different clients and devices. They can catch clipped content, broken layouts, inaccessible contrast, missing images, or CSS problems. However, a perfect rendering preview does not tell you whether the email will land in spam.

Spam and content analysis

Spam-analysis tools can inspect headers, HTML, text, and links for common technical or content-related issues. They are useful for debugging, especially during template development. Their scores should be treated as hints rather than a direct model of a mailbox provider's live filtering decisions.

Authentication monitoring and DMARC reports

DMARC reports help identify which systems send mail using a domain and whether authentication aligns correctly. This is valuable for finding unknown senders, broken third-party setups, or inconsistent signing across transactional and marketing streams.

Production telemetry

Event logs, bounce classifications, complaint notifications, unsubscribe events, and engagement reporting provide direct evidence from real mail recipients. These tools show what happened after the send and help validate whether seed-test signals matched production behavior.

The strongest program uses each tool for its proper job: previews for presentation, authentication monitoring for identity, seed tests for likely placement, and production telemetry for real audience outcomes.

When should you trust seed-test results?

Trust seed tests as a directional operational signal. A clean result supports proceeding when other fundamentals are healthy. A poor result is a reason to investigate, especially when the issue repeats across many seeds or a strategically important mailbox provider.

The confidence level rises when the test matches production conditions closely, the seed list is representative of the audience, the outcome is consistent across multiple seeds, and the result agrees with other evidence such as authentication failures, complaint trends, or provider-specific deferrals.

The confidence level falls when the sample is tiny, the message differs from production, results are still pending, the campaign uses highly individualized content, or the interpretation depends on one seed account. Avoid declaring a campaign safe or unsafe based on one isolated mailbox result.

Conclusion: use seed tests as an early-warning system

A seed test is a practical way to check likely inbox placement before a campaign reaches real people. It can reveal spam placement, provider-specific differences, authentication failures, and infrastructure regressions that delivery logs alone may not make obvious.

Its limits are just as important as its benefits. Seed inboxes do not act like subscribers, so a seed test cannot replace permission-based list growth, audience segmentation, authentication, reputation management, or production monitoring. Used alongside those practices, it becomes a high-value early-warning system: test the real message, identify meaningful patterns, fix the underlying cause, retest, and keep measuring live results.

FAQ

Is a seed test the same as a spam test?

Not exactly. A spam test may analyze content, headers, and known filter rules. A seed test sends the message to monitored inboxes to observe likely placement, such as inbox, spam, Gmail Promotions, or missing. Many deliverability platforms combine both capabilities.

Does a seed test guarantee inbox placement?

No. Seed results are directional because mailbox providers also use recipient-specific engagement, complaint, reputation, and behavioral signals. A positive seed test is useful evidence, not a guarantee for every recipient.

What is a good seed-test inbox placement rate?

There is no universal safe threshold because provider mix, audience type, message category, and sending history vary. The key is to establish a baseline for your program, investigate meaningful declines, and treat broad spam placement or authentication failures as urgent issues.

Should transactional emails get seed tested?

Yes, especially when you change sending infrastructure, templates, authentication, link domains, or a critical user flow. Transactional mail has different content and engagement patterns from marketing mail, so test it separately rather than assuming marketing results apply.

How often should you run a seed test?

Run one before high-stakes campaigns and whenever a material sending change occurs. For recurring campaigns, periodic testing helps establish a baseline and can expose regressions before they become widespread deliverability problems.