After enough production incidents, every SFMC engineer internalizes a pre-send checklist. The ones we've hit personally:
- Subject line still had [DRAFT] when 300k got the email
- Personalization field defaulted to blank, making every subject line Chào ,
- Wrong Send Classification caused the email to miss the primary audience
- Wrong DE selected - the campaign went to the welcome-series list instead of promo
- Unsubscribe link 404'd because the landing page wasn't published
Five minutes of checklist prevents most of these. Here's ours.
The checklist
Why each item
Preview with data from real segments
Dynamic Content Blocks only surface bugs when rendered against actual data. Preview with a Gold tier subscriber, a Silver subscriber, a brand-new subscriber. If any breaks, the real send will break for those groups.
Preview with missing-data subscriber
The most common production break: %%FirstName%% field empty for a subset. Without Default Value, subscribers see Hi ,. Always preview against a subscriber with minimal data.
Outlook test send
Outlook doesn't support many CSS properties. Background colors, some gradients, many layout tricks. If 20%+ of the audience uses Outlook, testing there is mandatory.
Subject line placeholder check
[DRAFT], [TEST], [APPROVAL] are marketer habits that slip into production. Configure Subject/Preheader Validation in Email Studio admin to block send if these strings appear - covered in a later post.
Link testing
Every link - hero, every product card, footer social, unsubscribe. Click them. 404s happen. Broken unsubscribe is a compliance violation in some jurisdictions.
Send Classification correctness
Wrong classification sends to the wrong audience. Commercial classification sends to marketing-opt-in subscribers; Transactional sends to everyone including unsubscribed (for order confirmations only).
Sending a promotional email with Transactional classification is an opt-out violation and damages sender reputation.
Audience double-check
"Marketing Gold Tier" filter sounds right. Click into the DE - 12 rows? Your filter condition is wrong. 2.4M rows? You're about to send to everyone including unsubscribed. Visual check every time.
Volume sanity check
Expected 50k, showing 500k? Something else got included. Expected 50k, showing 1.2k? Filter stripped too much. Both warrant investigation before Send.
Automating what's automatable
Some of these can be system-enforced:
- Subject/Preheader Validation - blocks send on flagged strings
- Verification Activity - halts automation if audience count is out of range
- Content Builder Approvals - forces a review step before send
But don't rely on automation alone. The checklist is the human layer that catches what the system doesn't.
The five-minute rule
If all eleven items take more than five minutes, the process is the problem:
- Test Send destinations saved in a distribution list (not typed each time)
- Approval step embedded in the tool (not a Slack ping)
- Volume expectations documented in the brief (not recalculated each send)
Optimize the checklist, don't skip it.
Team onboarding
New team members run the checklist supervised for their first five sends. After that, solo but audited weekly for the first month.
Pin the checklist in the team wiki. Print it and tape it over the Send button if needed.
Takeaway
Production email incidents are almost always preventable. The specific causes differ - typo, wrong audience, bad classification - but the cure is the same: a systematic pre-send checklist that a human actually runs. Five minutes per send. Zero marketers have regretted it.
Auditing SFMC send processes on client accounts? Our Salesforce team designs checklists and governance frameworks tailored to the team size and send volume on production engagements. Get in touch ->
See our full platform services for the stack we cover.