Inbox Placement Testing: How to Find Out Where Your Cold Emails Actually Land

Most outbound teams find out they have a deliverability problem months too late — after replies quietly dry up and a quarter of pipeline goes missing. The reason is simple: sending software reports that a message was accepted by the receiving server, not where that server filed it. Accepted and delivered are not the same thing. Inbox placement testing closes that gap by telling you what folder your cold emails actually land in before you burn a list of good prospects.

Delivery rate vs. inbox placement: the difference that costs you meetings

Your sending tool shows a delivery rate of 98%. That number only means the recipient mail server did not bounce the message. It says nothing about whether Gmail routed it to Promotions, whether Outlook dropped it in Junk, or whether Google quietly filed it in a folder nobody opens.

Inbox placement is the percentage of delivered messages that reach the primary inbox. On a healthy cold outreach setup it should sit above 90%. When it falls to 60% you are still seeing a 98% delivery rate in your dashboard while four out of ten prospects never see you. This is also why open rates can mislead you — a low open rate rarely tells you whether the cause is your subject line or a spam folder.

How a seed list test works

A seed list is a set of real mailboxes across the major providers that you control or that a testing service controls. You send your live campaign copy to those addresses, then check where each message landed. Steps:

1. Build or buy a seed set. Cover Gmail (personal and Google Workspace), Outlook.com, Microsoft 365, Yahoo, and at least one corporate mail server if your ICP uses them. Twenty to forty seeds is enough for a directional read.
2. Send the real thing. Use the actual sending inbox, the actual template, the actual links and tracking settings. A stripped-down test email will pass when your real campaign fails.
3. Record placement per provider. Log inbox, promotions/other, spam, or missing for every seed. Provider-level detail matters — Microsoft filtering behaves nothing like Google filtering.
4. Repeat per sending inbox. Placement is a property of the sender, not just the copy. One burned inbox in a rotation of eight can drag down an entire campaign.
5. Test on a schedule. Weekly during a ramp, monthly once stable, and always before scaling volume or launching a new domain.

Reading the results

Placement patterns point to different root causes:

• Spam at Google only: usually engagement-driven. Your reply rate is too low, your list is too loose, or your volume ramped too quickly.
• Spam at Microsoft only: usually reputation- or content-driven. Check whether your IP or domain appears on Spamhaus or SORBS with a blacklist check.
• Promotions tab across the board: your email looks like marketing. Strip images, tracking pixels, HTML signatures with logos, and link-heavy footers.
• Missing entirely: silent filtering at the gateway. Verify authentication with a full SPF, DKIM and DMARC setup.
• One inbox bad, the rest fine: retire or re-warm that mailbox rather than rewriting the campaign.

Fixing placement, in priority order

Work top to bottom. Copy changes are the last thing to touch, not the first.

1. Authentication. SPF, DKIM and a DMARC record that at least reports. Non-negotiable in 2026 for anyone sending at volume.
2. List hygiene. Bounces are the fastest way to lose placement. Run every list through email verification and keep hard bounces under 2%.
3. Warmup and volume discipline. A new mailbox needs weeks, not days. Follow a proper email warmup ramp and cap daily sends per inbox.
4. Engagement quality. Tighter targeting lifts replies, and replies lift placement. A smaller, better-fit list outperforms a bigger one on every placement metric.
5. Content. Plain text, one link at most, no attachments, no tracking pixel if you can live without it. Then review spam trigger words.

Common testing mistakes

• Interacting with your own seeds. Opening and starring every test message trains filters on fake behaviour and inflates future results.
• Testing once and calling it solved. Placement drifts. A domain that was clean in March can be filtered in June after a volume spike.
• Using only free personal accounts. Your buyers are on corporate Microsoft 365 tenants with third-party gateways. Test where your prospects actually are.
• Changing five things at once. Fix one variable per test cycle or you will never know what worked.
• Ignoring the sending domain split. Keep outreach off your primary domain by using dedicated secondary domains.

Build it into your weekly routine

Treat placement as an operating metric, not a fire drill. A workable cadence: run a seed test every Monday across all active inboxes, log placement by provider in a shared sheet, pause any mailbox that drops below 80% inbox placement, and re-warm it for two weeks before returning it to rotation. Pair that with the fundamentals in our email deliverability guide and you will catch problems while they are still cheap to fix.

The teams that consistently book meetings from cold email are not the ones with the cleverest subject lines. They are the ones who know, every week, exactly where their emails are landing.