Key takeaways
- A seed test sends your real email to test inboxes at different providers and records where it lands.
- Build seeds at Gmail, Outlook.com, Yahoo and the business providers your audience uses.
- Send the test exactly like a real campaign: same domain, platform, content and links.
- Seed results are a snapshot, not your real inbox rate, because seeds have no engagement history.
- Combine seed tests with Postmaster Tools and engagement by provider for a reliable picture.
In this guide
- Delivered is not the same as inbox
- How a seed test works
- Build your seed list
- Run the test correctly
- Read the results
- The limits of seed testing
- Estimate inbox placement with your own data
- When to run seed tests
- Free vs paid seed testing
- Keep your seed accounts healthy
- Interpreting mixed results
- Seed testing and privacy
- From test to action
- A simple seed testing log
- Frequently asked questions
Delivered is not the same as inbox
Your email platform reports a message as "delivered" when the receiving server accepts it. What happens next is invisible to the platform: the message can land in the inbox, in a tab such as Gmail Promotions, or in the spam folder. The delivery rate can be 99% while a large share of your mail sits in spam.
Inbox placement testing, or seed testing, is the most direct way to see where a campaign actually lands at each mailbox provider.
How a seed test works
- You create or rent a list of test mailboxes, called seeds, spread across the providers your audience uses.
- You add these seeds to your real campaign, or send them an identical copy through the same platform.
- You (or a tool) check each seed mailbox and record where the message landed: inbox, tab, spam, or missing.
- You compare results by provider and look for patterns.
Build your seed list
Start from your audience. Export your list and count contacts by domain. A typical consumer list is dominated by gmail.com, then Microsoft domains (outlook.com, hotmail.com, live.com), Yahoo and AOL, then local providers. A B2B list contains many business domains hosted on Google Workspace or Microsoft 365.
| Provider | Seeds to create | Why |
|---|---|---|
| Gmail | 3 or more personal accounts | Largest provider, tabs and spam folder |
| Outlook.com / Hotmail | 2 or more | Different filtering, frequent junk issues |
| Yahoo / AOL | 2 or more | Own filters and complaint sensitivity |
| Google Workspace business domain | 1 or 2 | Business filtering differs from consumer Gmail |
| Microsoft 365 business domain | 1 or 2 | Uses Microsoft business filtering |
| Local providers in your markets | 1 each | Important if your audience is regional |
Paid seed testing tools offer dozens of seeds and automatic reports. A home-made list of ten to fifteen accounts is enough to spot major problems.
Run the test correctly
- Same setup as the real send: same platform, same sending domain, same From address, same content and links. Testing with a different setup tests something else.
- Same time as the real send, or just before it. Reputation changes over time.
- No engagement on seeds before reading the result. Opening, moving or clicking changes future placement for that seed.
- Check every folder: inbox, all Gmail tabs, Outlook "Other", spam or junk.
- Record results in a simple table with date, campaign, provider and placement.
Read the results
| Result | What it usually means | Next step |
|---|---|---|
| Inbox or tabs everywhere | No major placement problem | Keep monitoring engagement |
| Spam at one provider only | Reputation issue with that provider | Check that provider's tools and your engagement there |
| Spam everywhere | Authentication, blocklist or content/link problem | Check SPF/DKIM/DMARC, blocklists and links |
| Missing at some seeds | Rejected or heavily delayed | Read bounce logs and delivery errors |
| Gmail Promotions | Normal for marketing mail | Not a failure; see the Promotions tab guide |
| Outlook "Other" tab | Normal for many newsletters | Not junk; Focused vs Other is user-specific |
When results point to spam at Microsoft, see Outlook-specific fixes. When they point to authentication, start with SPF, DKIM and DMARC.
The limits of seed testing
Seed tests are useful, but you should know their weaknesses to avoid wrong conclusions:
- Seeds have no engagement history. Real subscribers who open your emails every week are treated better than an inactive seed. Seeds can underestimate your real inbox rate.
- Filtering is personalized. Gmail and Microsoft adapt placement to each user. One seed is one user, not your whole audience.
- Small samples. With two seeds per provider, one result can be noise. Look at repeated tests and trends.
- Seeds can be recognized. Widely used seed networks may be treated differently by some providers.
Use seed tests together with real engagement data and Postmaster Tools. When all three tell the same story, you can trust the diagnosis.
Estimate inbox placement with your own data
Your own campaign data gives a second view, based on real subscribers:
- Segment results by mailbox provider: clicks per delivered email for Gmail, Microsoft, Yahoo and business domains.
- Compare with your historical average for each provider.
- Look for a sudden drop at one provider. If clicks at Microsoft fall by half while Gmail stays stable, you probably have a Microsoft placement problem, even before any seed test.
Use clicks rather than opens: privacy features can register automatic opens and make placement look better than it is.
When to run seed tests
- Before important campaigns (launches, seasonal promotions)
- After any change: new platform, new domain, DNS changes, new template
- During a warm-up, weekly
- When engagement drops at a specific provider
- After a blocklist removal or reputation incident
Free vs paid seed testing
| Home-made seed list | Paid seed testing tool | |
|---|---|---|
| Cost | Free | Subscription or per test |
| Number of seeds | 10 to 15 | Dozens, many providers and countries |
| Time per test | 15 to 30 minutes of manual checking | Automatic report |
| Extra checks | None | Often authentication, blocklists, content analysis |
| Best for | Small senders, occasional checks | Agencies, high volume, frequent testing |
Start with a home-made list. If you find yourself testing every week or managing several domains, a paid tool saves time.
Keep your seed accounts healthy
- Log in to each seed account regularly so providers do not close it for inactivity.
- Do not use seeds for anything else, such as signups or personal mail.
- Never mark test emails as spam or not spam: it changes future results.
- Use the same seeds over time so results stay comparable.
- Keep the seed addresses in a separate segment that you include in each campaign.
Interpreting mixed results
Results are rarely all green or all red. Typical mixed situations and what they usually mean:
- Gmail fine, Microsoft in Junk: Microsoft reputation or authentication for high-volume rules; see Outlook 550 5.7.515 if you also see bounces.
- Consumer inboxes fine, business domains in spam: security gateways reacting to new domains, link reputation or attachments.
- One seed in spam, others in inbox at the same provider: often noise or a personalized result; retest before acting.
- Everything in spam after a template change: check links, image-only content and any new third-party domain in the template.
Act when the same result repeats over several tests, or when seed results and engagement data agree.
Seed testing and privacy
Seeds are your own test accounts, so no subscriber data is involved. If you use a paid tool, you send it a copy of your campaign; avoid including personal data of real customers in test sends.
From test to action
A seed test is only useful if it leads to a decision. Before each test, write down the question you want answered, such as "Does the new template land in spam at Microsoft?" or "Is our new subdomain ready for full volume?". After the test, write the decision: keep, change, retest next week. This habit prevents the common trap of collecting results without changing anything.
A simple seed testing log
| Date | Campaign | Gmail | Outlook.com | Yahoo | Business | Notes |
|---|---|---|---|---|---|---|
| Week 1 | Newsletter | Promotions 3/3 | Inbox 1/2, Junk 1/2 | Inbox 2/2 | Inbox 2/2 | Watch Microsoft |
| Week 2 | Promotion | Promotions 3/3 | Junk 2/2 | Inbox 2/2 | Inbox 1/2 | Microsoft issue confirmed |
A log like this turns isolated tests into a trend, which is what you need to act with confidence. Keep it next to your reputation monitoring notes.
Frequently asked questions
How many seed addresses do I need?
At least two or three per major provider your audience uses. More seeds give more stable results, which is why paid tools use dozens.
Are seed tests accurate?
They show trends and problems, but they are not perfect. Seed inboxes do not engage like real subscribers, and filters personalize placement for each user.
Is a Promotions tab result a failure?
No. The Promotions tab is part of the inbox. Marketing emails often land there, and subscribers still see and read them.
Should I test every campaign?
Not necessarily. Test before important campaigns, after changes to your setup, and when engagement drops at a specific provider.
Can I build a free seed list?
Yes. Create a few free accounts at major providers. Paid tools add more seeds and automated reports, which is useful at scale.
