When an email underperforms, separate delivery, measurement, audience, and message fit before rewriting the copy. Changing all four at once hides the cause and can make a temporary problem worse.
Diagnose from the outside in
- Was it accepted by the receiving system? Compare attempted sends, accepted messages, bounces, and blocks by provider. A dashboard’s “sent” count may not mean inbox placement.
- Did it reach the intended audience? Check suppression, consent status, list source, and recent changes to collection or import.
- Can the action be measured? Test links, redirects, mobile rendering, and event recording. Use a tagged test address so the team can observe the full path.
- Is the message relevant? Only after the first three pass, review subject, expectation, content, and next action.
Example: If one mailbox provider shows a sudden rise in rejections immediately after a new domain setup, inspect domain authentication and sender configuration before changing the subject line. Confirm current provider-specific guidance; do not copy a DNS record from a random blog.
Isolate changes
Pause questionable imports or an affected stream if consent or source quality is uncertain. Check official documentation for the sending provider and mailbox providers, then involve whoever manages DNS when authentication is involved. Do not repeatedly resend the same campaign as a “test.”
Keep a change log of sending domain, authentication, list source, volume, template, and date. Change one likely cause, run a controlled test, and compare by provider and audience cohort. Opens can be distorted by privacy features; combine them with delivery evidence, clicks, replies, and downstream actions.
Optional AI-assisted work brief — Delivery diagnostic sequence
Inspect the sending domain, authentication status, recipient mix, bounce and complaint patterns, list acquisition, and recent volume changes before touching copy. Separate hard bounce, soft bounce, deferral, spam complaint, and engagement. Recommend one controlled remediation at a time and identify evidence that would disprove it. Do not advise sending to unconsented or suppressed contacts.
Worked example — Do not rewrite copy before checking delivery
If opens and clicks fall together immediately after a sender-domain change, inspect authentication, deferrals, bounce classes, and mailbox-provider responses before changing subject lines. If delivery is stable but clicks fall on one message, copy and destination become more plausible causes. Change one layer at a time and preserve the evidence window; otherwise the team cannot tell whether the fix worked.
AI disclosure: AI tools assisted with drafting. The diagnosis example is illustrative and not a current provider-status claim.