Re-engagement Email Subject Line A/B Test
Dormant subscribers don't open everything you send — so the subject line is where your win-back campaign lives or dies. This guide shows how to run a re-engagement email subject line A/B test with marginal, the hosted email MCP server your AI agent can drive end to end.
Why subject lines matter most for re-engagement
A re-engagement campaign targets people who've stopped opening. By definition, your baseline engagement is low, which makes the subject line the highest-leverage variable in the whole lifecycle email. A small lift in open rate cascades into more clicks, more reactivations, and a cleaner list.
Because re-engagement audiences are skeptical and inattentive, guesswork is expensive. An email subject line test removes the debate: you send competing lines to comparable slices of the same dormant segment and let opens decide.
- Curiosity vs. urgency vs. plain-spoken "we miss you" framing
- Incentive-forward lines ("Here's 20% to come back") vs. value reminders
- Personalization tokens and short-form mobile-friendly phrasing
- Re-permission asks ("Still want these?") that double as list hygiene
Running the test with marginal
marginal exposes four MCP tools your agent calls in sequence: generate_variants drafts re-engagement subject lines, launch_test sends them across your dormant segment, get_results returns open and click tracking with within-test lift metrics, and recommend_next suggests the line to roll out to the remainder of the audience.
Everything is managed and hosted — there's no self-hosted infrastructure to run. You authenticate with a Bearer API key (marg_live_...) and point your client at https://marginal.sh/mcp.
- generate_variants — produce several re-engagement subject line candidates
- launch_test — managed send of the subject line A/B test
- get_results — open/click tracking and within-test lift
- recommend_next — pick the winner for the rest of the segment
An AI agent workflow for win-back
Because marginal is an email MCP server, you can drive the whole re-engagement workflow from your editor or agent — Cursor, Claude Code, Claude Desktop, Windsurf, Cline, Continue, Zed, or OpenAI Codex. AI agent email marketing means describing the goal once and letting the agent generate, launch, measure, and recommend without manual exports.
A typical lifecycle email loop looks like: define the dormant cohort, ask the agent for five subject line variants, launch the test on a holdout slice, wait for statistically meaningful opens, then promote the winning line to the full segment.
Tips for re-engagement subject line tests
Keep your variants genuinely distinct so the test teaches you something — testing two near-identical lines wastes a send. And measure on the metric that matters: for re-engagement, open rate is the primary signal, but watch downstream clicks to confirm the winning line attracts reactivators, not just curious openers.
- Test 3–5 contrasting angles, not minor word swaps
- Hold out a slice so you can roll the winner to everyone else
- Use the free tier's 100 experiments/month to iterate weekly
- Read the docs at https://marginal.sh/docs/ for tool schemas
Get started with marginal
marginal is a hosted email marketing MCP server at marginal.sh. Sign up free, create an API key, and connect https://marginal.sh/mcp from Cursor, Codex, or Claude Desktop.
- 100 experiments/month on the free tier
- Four MCP tools: generate_variants, launch_test, get_results, recommend_next
- Listed in the MCP Registry as sh.marginal/mcp