← All guides marginal.sh

Running an Inbox Placement Subject Line Test

Updated June 2026 · Hosted email MCP server

An inbox placement subject line test helps you understand which subject lines actually reach and engage your audience — not just which ones look good on paper. With marginal, AI agents can generate variants, launch the test, and read back open and click metrics without leaving their editor.

What an inbox placement subject line test measures

True inbox placement is something only mailbox providers can confirm, but the signal you can act on is engagement: opens and clicks across subject line variants sent to comparable segments. If a subject line consistently underperforms, it's often a sign of poor relevance or filtering — and a subject line A/B test surfaces that quickly.

marginal frames this as a within-test comparison. You send competing subject lines, track who opens and clicks, and read lift between variants so you can keep the winner.

How marginal runs the test

marginal is a hosted email MCP server, so your AI agent talks to it through four tools. You don't operate any infrastructure — the sending, tracking, and metrics are managed for you.

A typical flow has the agent draft alternatives, launch a managed send, then return for results once data accumulates.

Where AI agent email marketing fits

Because marginal speaks MCP, AI agent email marketing workflows become conversational. From Cursor, Claude Desktop, Claude Code, Windsurf, Cline, Continue, Zed, or OpenAI Codex, you can ask the agent to spin up a subject line test and report back — all against the same endpoint at https://marginal.sh/mcp.

Authentication is a single Bearer API key (marg_live_...), and the free tier covers 100 experiments per month, which is plenty to validate the approach before scaling.

Reading results and iterating

After a test runs, call get_results to compare variants side by side. Look at lift between subject lines rather than absolute numbers, since list size and timing affect raw counts. When one subject line clearly leads, lock it in and let recommend_next propose your follow-up.

Treating each inbox placement subject line test as one cycle in a loop — generate, launch, measure, repeat — keeps your subject lines improving against real engagement instead of guesswork. See the docs at https://marginal.sh/docs/ for tool details.

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.

Documentation · Sign up · FAQ