Two mail estates, two very different inboxes

    Who this was forA services business with a small sales team, sending cold email from a mixed estate of inboxes it had inherited.

    Inbox placement tested before the copy

    All case studies

    The drawing

    Seed inboxesPlacement test acrossevery sending inboxGoogle WorkspaceinboxesInboxes on theinherited relay100 percent inbox85 to 100 percent spamClient facing healthtool

    The problem

    Reply rates were low, and the team was reading that as a copy problem. So the fix on the table was new copy.

    Copy is a fair first suspect. But nobody had checked where the emails were landing. Placement was assumed.

    What we built

    We ran a seed placement test across the whole sending estate before touching a word of copy. Seed inboxes receive a test send from every sending inbox, and each result gets logged by where it lands.

    We published the results to a health tool the client can open any time.

    What it produced

    The estate split in two.

    • The Google Workspace inboxes landed in the inbox 100 percent of the time.
    • Several dozen inboxes on the inherited relay were landing 85 to 100 percent in spam.

    That's the awkward bit. Mail sent through the relay mostly never reached an inbox, and the blended numbers still looked survivable. A good inbox group averaged against a bad one gives you a mediocre number, and a mediocre number gets blamed on the writing.

    New copy wouldn't have fixed any of this. It would've gone to spam too.

    What runs now

    The health tool stays live, so the client can check placement per inbox without asking us. The estate's size is tracked in the client's own database.

    Your scenario

    If your reply rate dropped and your first move is rewriting the sequence, test placement first. Tell us how your sending estate is set up and we'll tell you what we'd test first.

    Related tools: SMTP2GO, BounceBan, Instantly, Domain Deliverability Checker, Inbox Placement Triage

    Some of the links on this page are affiliate links.

    Client names are always kept confidential at MMG, so they have been removed from this case study.

    Published February 24, 2026

    Last updated April 22, 2026