A pre-send email QA check answers one practical question: is this exact campaign, for this exact audience, ready to leave the system? It is the final control between a polished draft and a preventable mistake.
The useful version is more than “send a test to yourself”. A design test can look perfect while the real campaign contains an old audience, a broken tracked link, an unmonitored reply address or an unsubscribe path that never reaches the sending list. A reliable preflight checks the live version, records the evidence and gives one named person the authority to stop the send.
This guide is for permission-based business email. It is operational guidance, not legal advice. Rules differ by jurisdiction and circumstance. If your audience evidence is uncertain, complete the email consent-record audit before using this final checklist. For the broader system, see the beginner’s guide to email marketing.
The short answer: check the campaign, not just the template
A sound pre-send process has seven gates: scope, audience, identity, links, replies, opt-out, and rendered delivery. Each gate ends in an observed result. “Looks fine” is not an observed result; “all five distinct destination URLs returned the intended page, and the two tracked variants preserved their parameters” is.
Keep account-level controls separate. SPF, DKIM, DMARC, DNS and sending reputation need ongoing ownership, but rebuilding them before every newsletter is not useful. Google’s current email sender guidelines describe authentication, message formatting, spam-rate monitoring and additional requirements for higher-volume senders. Your campaign preflight should confirm that the approved sending setup has not changed and escalate an unexpected failure; it should not improvise a domain repair minutes before launch.
Start with a frozen send brief
Before clicking through the email, write the version you are approving. Record the campaign name, sender, purpose, audience rule, exclusions, scheduled time and owner. Add the final subject line, preheader and reply address. If any of those details change, the approval no longer describes the same send.
Save an audience count with the selection time. The count is a reconciliation aid, not proof of permission. It helps you notice that a filter expected to select 820 subscribers now selects 8,200, or that an exclusion has silently disappeared.
A compact scope record can look like this:
Campaign:
Purpose and primary action:
Sender name and From address:
Reply route and owner:
Audience rule and snapshot time:
Exclusions applied:
Final subject and preheader:
Scheduled time and timezone:
Approver:
Stop conditions:Do not copy last month’s approval and change only the date. Reuse the structure, then inspect the current evidence.
Gate 1: reconcile the real audience
Open the audience or segment the system will actually send to. Compare its rules with the brief, not with your memory. Check inclusion conditions, exclusions, duplicates, inactive or bounced addresses, and recent preference changes. If the platform shows a final recipient preview, save or record it after the segment refreshes.
Permission and relevance are related but different. A person may have valid evidence for a monthly newsletter but not for an unrelated partner promotion. The UK ICO’s current electronic-mail marketing guidance explains that organisations remain responsible when using a bulk-email platform, must not hide their identity, and must provide a valid opt-out contact route. It also distinguishes consent from limited soft opt-ins.
For the preflight, record which permission or exception the responsible person has already established for this campaign type. Do not relabel an exception as consent, and do not resolve an uncertain record by assuming it is “probably fine”. Hold the affected records or the entire send while the issue is reviewed.
Run one controlled suppression test using addresses you own where the platform permits it. Represent an eligible subscriber and an opted-out or held record. Preview the actual audience. A normal template test may skip recipient rules, so it cannot prove suppression works.
Gate 2: verify identity, subject and preheader together
Read the inbox envelope as a recipient will see it: display name, From address, subject and preheader. They should identify the sender and set an accurate expectation for the message. Google advises that display names should clearly and consistently identify the sender, rather than imitating urgency, message threads or content labels.
Check personalisation fallbacks with a record that has the field and one that does not. Look for doubled spaces, an exposed merge tag or a sentence that makes no sense without the value. Review dates, prices, deadlines and claims against the source used by the writer. An attractive subject does not repair an inaccurate body.
If the message serves US recipients, the Federal Trade Commission’s CAN-SPAM business guide requires accurate routing information and subject lines for commercial email, alongside location and opt-out requirements. Treat the final applicable legal review as a named responsibility rather than a box copied from a generic template.
Gate 3: build and test a link inventory
List every distinct destination in the final email: logo, text links, buttons, images, social links, preference link, unsubscribe link and any “view in browser” route. Include regional or personalised variants. A repeated call-to-action may share one URL, but do not assume visually similar elements are wired identically.
For each destination, check:
- Result: the page loads and matches the promise made by the link text.
- Protocol and domain: the address uses the intended secure domain, without a staging or editor URL.
- Tracking: approved parameters are present once, encoded correctly and not overwriting the destination’s own query string.
- Access: the page does not require an unexpected login, location or device.
- Campaign truth: stock, dates, prices and offer terms still agree with the email.
Test the links from the received seed message as well as from the editor. Some platforms rewrite URLs during sending. If a tracked link fails, pause the send; removing tracking without understanding the failure can hide a broader problem.
Gate 4: send a real reply and follow it
If the email invites questions, replies are part of the product. Send a reply from a controlled recipient account and follow it to the destination. Confirm the address accepts mail, the message is not silently discarded, and the named owner knows where to find it.
A monitored mailbox can still fail operationally if messages arrive in a folder nobody checks, a departed colleague owns the alias or an automation creates a ticket without the campaign context. Record the observed route and owner.
A no-reply address is not automatically wrong, but it must fit the message. Do not write “reply and tell us” above a route that rejects replies. If readers should use another contact method, state it clearly and test that method from the received email.
Gate 5: test unsubscribe and preference handling end to end
Open the unsubscribe route from a controlled seed address. It should be visible, understandable and usable without an account login. Complete the action and confirm that the status reaches the system that controls sending. Then rebuild or refresh the audience and verify that the address is excluded.
The ICO says each subsequent marketing message must offer a simple way to opt out and gives a clear link or direct reply as practical examples. Google requires marketing and subscribed messages sent at higher Gmail volumes to support one-click unsubscribe headers as well as a visible body link. The visible link and the technical header solve related but different parts of the recipient experience.
Do not test by unsubscribing a real customer. Use an address you control, label the event and restore it only through the same documented process a genuine subscriber would use. A manual database change can make the test pass while leaving the public route broken.
Gate 6: inspect the received email, not only the editor
Send the final version to a small seed group that represents the clients and devices your audience actually uses. Check a narrow phone view, a wider desktop view, images-off behaviour and dark mode if it matters to your template. The goal is not pixel-for-pixel sameness. It is readable hierarchy and a usable action when clients make ordinary rendering changes.
Look for readable text size, sensible line length, useful alt text, visible focus or link cues, buttons that are large enough to tap, and a layout that does not force horizontal scrolling. Confirm the main point appears without relying on an image. Image-only instructions, tiny footer controls and low-contrast link text are practical failures even if the desktop screenshot is attractive.
Review the plain-text alternative when your sending platform creates one. Remove raw tracking fragments, repeated navigation and missing destinations. If the platform generates it automatically, inspect the output rather than assuming it is coherent.
Gate 7: use stop, fix and send decisions
Give every check one of three outcomes:
- Send: the check passed on the final version and the evidence is recorded.
- Fix and re-test: the issue is understood, bounded and assigned to an owner.
- Stop: permission, identity, suppression, unsubscribe, essential destination or another material control cannot be established before the deadline.
Do not average the results. Six green checks do not cancel one broken unsubscribe route. After a fix, repeat the failed check and any dependent checks. A subject-line change may require only copy review; an audience-rule change requires a fresh audience reconciliation and suppression test.
This approach is the campaign-level companion to the context filter for email marketing advice: keep safeguards non-negotiable, and test tactics without pretending every choice has the same risk.
A worked hypothetical preflight
Imagine a fictional two-person consultancy preparing a monthly operations note for 1,140 subscribers. The editor sends a design preview to herself and finds no spelling errors. The owner then runs the campaign preflight.
The audience preview shows 1,154 recipients. Fourteen records entered through a new event form whose newsletter wording has not been reviewed. Those records are held, returning the audience to 1,140. This is an invented example to demonstrate the method, not a reported result.
The link inventory contains six distinct destinations. Five work. A button labelled “download the worksheet” points to last month’s file because the editor duplicated the campaign. The owner replaces the URL and tests the received tracked link, not only the editor button.
A reply reaches the shared inbox but is assigned to a former contractor. The team changes the rule, sends a second reply and observes the message in the current owner’s queue. The unsubscribe test records the preference, and the refreshed audience excludes the seed address.
Only then does the owner approve the scheduled version. The result is not “the email is perfect”. It is a reproducible record that the material risks were checked on the version being sent.
Copy this pre-send sign-off
Campaign and final version:
Audience rule, snapshot time and count:
Permission basis reviewed by:
Suppression test result:
Sender, subject and preheader result:
Distinct links tested and failures:
Reply route tested, destination and owner:
Unsubscribe path and refreshed-audience result:
Seed clients and views checked:
Accessibility and plain-text notes:
Account-level sender configuration changed?:
Open issues, owner and deadline:
Decision: SEND / FIX AND RE-TEST / STOP
Final approver and time:Keep the record with the campaign. If a material field changes after approval, reopen the relevant checks. The checklist is useful because it helps someone make and explain a decision, not because every row eventually turns green.
Reader Q&A
Does a test email prove the real campaign is ready?
No. A design test may bypass the live audience, suppression rules, personalisation and automation logic. Use a controlled seed group, preview the actual segment and test the paths the scheduled campaign will use.
How many links should I test before sending?
Test every distinct destination and tracked variant in the final campaign. Repeated buttons may share a URL, but image links, text links, unsubscribe links and personalised or regional variants can behave differently.
Should a marketing email use a no-reply address?
A no-reply address is a business choice, not a universal requirement. If the message invites questions, use and monitor a reply route. If replies are not handled in that inbox, make the alternative contact path clear and test it.
Is an unsubscribe link enough for the pre-send check?
No. Confirm that it is visible, opens without an account login, records the choice in the system that controls sending and prevents the address from re-entering the audience. Higher-volume Gmail senders also need the required one-click unsubscribe headers.
Who should give final approval to an email campaign?
Assign one named owner who can stop the send. Specialists may approve copy, data or legal questions, but the final record should show who reviewed the actual scheduled version, which audience snapshot they checked and what unresolved issues remain.
What should I do when one pre-send check fails?
Pause the affected send, record the failure and fix the narrow cause. Rebuild the final audience or message if the change affects it, then repeat the failed check and any dependent checks. Do not send because the deadline is close.
Make the next send easier
After the campaign is sent, use an email reporting dashboard to connect the checked links and audience with measured outcomes, consistent reporting windows and the next action.
Run the checklist on the next real campaign, then keep the evidence that will shorten the following review: a stable link inventory, named reply owner, approved seed accounts and a documented audience rule. Repeated checks should become faster because the system is clearer, not because the team stops looking.
The final discipline is simple: test the route the campaign will actually use, stop on material failures, and record the version you approved. That is what turns a test email into a controlled send.
