Best Practices

Field-tested guidance for deploying and operating RME at enterprise scale.

Deliverability

1. Authenticate every sending domain with SPF, DKIM, and DMARC before the first production send.
2. Separate transactional and marketing traffic onto different subdomains and IP pools.
3. Warm new IPs and domains over two to four weeks; never jump to full volume.
4. Keep hard bounce rate under 2% and complaint rate under 0.1%.
5. Validate addresses at capture, not at send time.
6. Re-engage or remove recipients with 12 months of no engagement.

Architecture

1. Send asynchronously from triggers — never call out in the same transaction as DML.
2. Use templates rather than inline bodies so content is versioned and reviewable.
3. Set Idempotency-Key on every external send request.
4. React to outcomes with Platform Events instead of polling status.
5. Keep one provider config per environment and per business unit.
6. Configure a failover provider for every production provider.

Content

1. Write a meaningful subject under 60 characters; front-load the value.
2. Always generate a plain-text alternative.
3. Inline CSS, absolute image URLs, and descriptive alt text.
4. One clear call to action per message.
5. Keep SMS inside a single segment when possible and always include an opt-out on marketing traffic.

Governance

1. Restrict the Marketing category to a small set of users.
2. Require template review before activation.
3. Review the audit log monthly for suppression removals and consent overrides.
4. Set per-user and per-org send ceilings before go-live.
5. Keep sandboxMode enabled in every non-production org.

Testing

1. Refresh sandboxes and re-point provider configs to test credentials immediately after.
2. Use RME_Test.mockProvider in every Apex test; never call a live provider from tests.
3. Preview every template against real record ids before activation.
4. Run a seed-list send before any large campaign.
5. Load test bulk paths in a Full sandbox with production-scale data.

Operations

1. Dashboard delivery rate, bounce rate, complaint rate, and provider health daily.
2. Alert on circuit breaker events and on failure-rate anomalies.
3. Rehearse a provider outage: disable the primary and confirm failover.
4. Document the runbook for a reputation incident before you need it.

Continue to Troubleshooting.

Last updated 1 month ago