Email Marketing: Consent, Deliverability and Automation
Email questions test permission, deliverability and lifecycle design rather than creative inspiration. The exam expects you to know why email reaches the inbox or does not, how to segment and automate, and which metrics actually indicate programme health.
Permission and deliverability
Consent must be freely given, specific and recorded - under GDPR and equivalent regimes, with double opt-in as the defensible standard. CAN-SPAM requires accurate headers, a physical address and a working unsubscribe. Authenticate sending domains with SPF, DKIM and DMARC; the DMARC policy record tells receivers what to do with failing mail. Manage list hygiene: suppress hard bounces permanently, retry soft bounces, and watch complaint rates - sustained complaints above roughly 0.1% indicate a list or relevance problem. Purchased lists breach consent rules and destroy sender reputation.
Segmentation and automation
Segment by lifecycle stage, behaviour (opens, clicks, purchase recency), value (RFM: recency, frequency, monetary) and stated preference. Automations run on triggers rather than batches: welcome series, browse or cart abandonment, post-purchase and onboarding, win-back and re-engagement, with suppression rules so customers do not receive contradictory messages.
Metrics
Delivery rate, bounce rate, click rate, click-to-open rate, conversion rate, revenue per email, list growth and churn. Treat open rate cautiously since privacy features can pre-fetch images and inflate opens; prefer click and conversion measures for decisions.
Worked example
A send to 400,000 delivered emails achieves a 1.5% click rate (6,000 clicks) and converts 8% of clickers (480 orders) at an average order value of £60, giving £28,800 revenue against £2,400 send cost. Now split the list by past purchase category and send two tailored versions: version A lifts click rate to 2.1% on its 200,000 segment. Reason carefully - a lift on a segment is not automatically a whole-list win, so the test must be compared like-for-like on a held-back control before rollout.
Traps
- Buying or scraping lists.
- Using open rate as the primary success metric.
- Sending one message to the entire list with no segmentation.
- Testing several variables at once, or calling a test before sufficient volume.
- Ignoring suppression and frequency rules across automations.