Pre-send testing: a workflow that catches real failures

Testing is risk management, not perfectionism

Pre-send testing exists because the cost of a broken campaign exceeds the cost of thirty minutes of QA. A single inverted logo in Outlook or a clipped Gmail message can waste a product launch list. Testing does not promise perfection in every client; it promises no catastrophic surprises in the clients your audience actually uses.

Small teams skip testing because enterprise tools feel expensive and manual seed inboxes feel tedious. This guide sequences free and paid steps so you can start with zero budget and add automation when send volume justifies it.

Separate rendering tests from deliverability tests. Rendering asks whether the HTML looks correct. Deliverability asks whether authentication passes and content triggers spam filters. Both belong in the workflow, but they measure different failures.

The minimum viable checklist

Send a proof to at least three live accounts you control: Gmail web, Apple Mail on iPhone, and Outlook desktop if you sell B2B. Click every link, including unsubscribe and view in browser. Confirm merge tags resolved.

Validate SPF, DKIM, and DMARC on the proof using Gmail message headers or a header parser. If any leg fails, stop and fix DNS before scheduling.

Run a spam content scan with Mail-Tester or your ESP's built-in checker. Scores are imperfect, but sudden drops after a rewrite flag risky copy or link shorteners.

Scaling with preview tools

Litmus and Email on Acid capture screenshots across dozens of clients simultaneously. They pay for themselves when a single template ships weekly to heterogeneous audiences. Export shares to Slack so stakeholders approve visually, not from their personal Gmail only.

HTML validators such as HTML Email Check catch unclosed tags, oversized files approaching Gmail clip thresholds, and missing alt attributes before you upload to an ESP.

Accessibility checks matter too: logical reading order, sufficient contrast, and meaningful alt text. Some enterprise buyers now ask for basic ADA diligence on customer communications.

Link, tracking, and personalization QA

UTM parameters should be consistent with your analytics taxonomy. Broken utm_campaign values silently ruin reporting. Personalization fallbacks must render when a field is empty: never ship Dear {{ first_name }} to a live list.

Tracking domains should match authenticated sending domains where possible to avoid mismatched reputation signals. Test with ad blockers and Apple Mail Privacy Protection in mind: opens will undercount, so do not treat open rate alone as proof of rendering failure.

If you use AI-generated HTML from platforms like Brew, run the same QA pass you would on agency HTML. Generation speed does not remove responsibility for links and compliance footers.

Document and regress

Save screenshots and checklist sign-offs per send in a shared folder. When something breaks in production, compare against the last known good proof to see what changed: template, audience, or DNS.

Re-run full previews when you change ESPs, add a new module type, or alter footer compliance language. Incremental edits feel safe but accumulate risk.

After major mailbox provider announcements, such as Gmail bulk sender updates, re-test authentication and unsubscribe headers even if templates did not change. Policy moves faster than design.

Sources

• Litmus pre-send checklist
• Mail-Tester
• Gmail email sender guidelines
• Brew documentation: sending and QA

Lectures associées

• Why the same email looks different in every inbox
• Email client CSS: what works, what fails, and what to avoid
• Email deliverability for founders: the first 90 days