How to Run a Click Through Rate Email Test
Open rates tell you whether a subject line earns attention, but click-through rate tells you whether the email earns action. This guide walks through running a click through rate email test with marginal, the hosted email MCP server, so your AI agent can test, send, and read results without leaving your editor.
What a click through rate email test measures
Click-through rate (CTR) is the share of recipients who click a link inside your email. A click through rate email test compares two or more variants to see which version drives more of those clicks — useful when you're optimizing calls to action, link copy, or the subject line that sets reader expectations before they open.
Subject lines influence clicks indirectly: a misleading subject can inflate opens but tank CTR, while an honest, compelling one tends to attract readers who actually want what's inside. marginal tracks both opens and clicks per variant, so you can see the full funnel rather than guessing.
- Open rate: did the subject line earn the open?
- Click rate: did the content earn the click?
- Within-test lift: how much better is the winning variant?
Running the test with marginal's email MCP server
marginal exposes four MCP tools your AI agent calls directly: generate_variants drafts candidate subject lines, launch_test sends them as a managed split, get_results returns open and click metrics, and recommend_next suggests the follow-up variant to try. Because marginal is hosted, there's nothing to deploy — point your client at https://marginal.sh/mcp and authenticate with a Bearer API key.
A typical click-through-focused flow asks the agent to generate variants emphasizing different value propositions, launch a subject line A/B test, then poll get_results once enough recipients have engaged. The within-test lift number tells you whether the click difference is meaningful or noise.
- generate_variants — draft subject line options
- launch_test — managed split send with click tracking
- get_results — open/click metrics per variant
- recommend_next — data-driven next iteration
Why an AI agent makes CTR testing faster
AI agent email marketing turns a multi-step process into a conversation. Instead of switching tools to draft copy, configure a split, and export metrics, you describe the goal and the agent orchestrates the MCP tools. marginal works inside Cursor, Claude Desktop, Claude Code, Windsurf, Cline, Continue, Zed, and OpenAI Codex.
This tight loop encourages real iteration. When a variant underperforms on clicks, recommend_next proposes the next experiment so you keep improving rather than shipping once and hoping.
- No context switching between editor and email platform
- Generate, launch, and read results in one workflow
- Free tier covers 100 experiments per month
Getting accurate click-through results
Reliable CTR numbers depend on sample size and clean tracking. marginal handles click tracking on managed sends automatically, so every variant is measured the same way. Give each test enough recipients before declaring a winner, and treat early lift figures as directional until the numbers stabilize.
See the docs at https://marginal.sh/docs/ for tool parameters and result fields, then wire up your client to start your first click through rate email test.
- Wait for adequate sample size before acting on lift
- Compare click rate alongside open rate, not in isolation
- Use recommend_next to chain experiments toward higher CTR
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