Send Time vs Subject Line Experiment: Which to Test First
When you only have so many experiments to run, choosing between a send time vs subject line experiment matters. This guide breaks down where each variable moves the needle and how to test both cleanly using marginal's email MCP server.
Why the two variables behave differently
Subject lines and send times influence different stages of the funnel. A subject line A/B test mainly affects whether someone opens at all — it's a content lever. Send time affects when a message lands at the top of an inbox and how likely it is to be seen before newer mail buries it.
Because they act on different mechanisms, the right starting point depends on your current metrics. If opens are low relative to your list quality, subject lines are usually the bigger win. If opens are decent but inconsistent across campaigns, send time is worth isolating.
- Low open rate, healthy list → start with subject lines.
- Opens swing wildly by day or hour → test send time.
- New audience with no baseline → subject lines first, since they compound across every future send.
- Time-sensitive content (events, sales) → send time often dominates.
Running each experiment with marginal
marginal is a hosted email MCP server, so your AI agent calls a small set of tools to design, launch, and read results. For a subject line A/B test, you generate competing variants, launch a split, and measure within-test lift on opens and clicks.
Send time experiments work the same way structurally: hold the subject line constant and vary the launch window across segments. The key is changing only one variable per test so the lift you measure is attributable.
- generate_variants — produce subject line candidates for an A/B test.
- launch_test — start a managed send with the split you define.
- get_results — pull open/click tracking and within-test lift metrics.
- recommend_next — get a suggested follow-up experiment based on results.
A clean sequence to test both
Rather than confounding both variables at once, sequence them. Lock down your best subject line first, then use that winner as the fixed control while you probe send time. This keeps each result interpretable and lets an AI agent email marketing workflow iterate without muddying attribution.
- Round 1: subject line A/B test, send time fixed. Record the winner.
- Round 2: hold the winning subject line, vary send time across segments.
- Round 3: re-test subject lines with the better send window confirmed.
- Use recommend_next between rounds to prioritize the next variable.
Getting started
Point your MCP client at https://marginal.sh/mcp and authenticate with your Bearer API key. The free tier includes 100 experiments per month, which is enough to run a full send time vs subject line experiment sequence and validate your results. Full tool reference lives at https://marginal.sh/docs/.
- Endpoint: https://marginal.sh/mcp
- Works with Claude Code, Cursor, Windsurf, Cline, Continue, Zed, and more.
- Hosted — no self-hosted server to maintain.
- Free tier: 100 experiments/month to start testing both variables.
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