How to Run an Outlook Inbox Placement Test for Cold Outreach
Most placement testing guidance focuses on Gmail. But a large portion of your B2B audience is on Outlook. Here's how to test specifically for Microsoft recipients.
You want to test specifically how your cold emails perform on Outlook and Microsoft 365 accounts. Most testing guidance focuses on Gmail, but a large portion of your target B2B audience is on Microsoft-hosted email — if you only test against Gmail, you might have excellent inbox placement on Google while your emails are going straight to Junk for the exact audience you need to reach.
Why This Matters
Microsoft's filtering differs from Gmail's in ways that matter for cold email. Content, IP reputation, and sender patterns are weighted differently. An email that passes Gmail's filters can fail Microsoft's. You need Outlook-specific testing.
Step-by-Step Testing Process
Create test accounts on Outlook.com — set up 3–5 personal Outlook.com accounts (free accounts). Do not add any contacts or set up any rules. These accounts should reflect a default Outlook configuration.
If you have access to a Microsoft 365 business account, create test addresses there too. M365 business accounts may have additional Exchange Online Protection policies that personal Outlook.com accounts do not — testing both gives you a clearer picture.
Use the inbox placement test to send your actual campaign content from each cold email inbox to all test accounts. Use the exact subject line, body content, links, tracking, and signature you plan to use in production.
Check each test account: Inbox, Junk Email folder, and Other tab (if Focused Inbox is enabled). Note where each email landed.
Record results in a spreadsheet tracking sending inbox, receiving account, placement (Inbox/Junk/Other), and date/time. Run at least 3 rounds over a week to get a reliable baseline.
How to Interpret Results
All test emails in Inbox: You're in good shape for Outlook recipients. Proceed with your campaign.
Mixed results (some inbox, some Junk): Inconsistency usually points to IP rotation. Your outbound emails may route through different Microsoft IPs on different sends. This is common on shared tenants.
All test emails in Junk: Your sending reputation with Microsoft is poor. Diagnose using SNDS, authentication headers, and content analysis. Run the blacklist checker on your domain and sending IP.
Emails in "Other" tab (Focused Inbox): This is not spam placement. Focused Inbox is Microsoft's equivalent of Gmail tabs. Better than Junk but worse than Focused inbox for engagement. Usually means your emails look like low-priority or bulk mail — simplify content and increase personalization.
The Fix Path Based on Results
For Junk placement: Check IP reputation via SNDS, run auth checks with the SPF checker, DKIM checker, and DMARC lookup, and simplify content.
For "Other" tab placement: Simplify content, personalize more, and send fewer emails to build engagement signals.
For inconsistent results: The issue is likely IP-related. Consider whether a dedicated IP or different sending infrastructure would provide more consistent results. Run the burn score calculator to assess overall infrastructure health.
When to Replace Instead of Repair
Testing itself doesn't require repair or replacement — it informs when repair or replacement is needed. If testing reveals that your M365 inboxes consistently underperform for Outlook recipients, WarmInboxes provides prewarmed M365 accounts that can serve as benchmarks for what healthy Outlook placement looks like, and can replace underperforming inboxes immediately.
Mistakes That Make This Worse
- Only testing against Gmail seed accounts and assuming Outlook will be similar
- Testing with plain text but sending campaigns with HTML and tracking
- Testing once and assuming results are permanent
- Not testing against both personal Outlook.com and business M365 accounts
- Ignoring the "Other" tab in Focused Inbox and treating it as equivalent to Primary inbox
Run the checks first
Before replacing anything, run a free inbox placement test. You might find the issue is DNS, not the domain — and save yourself a week of unnecessary work.