Running an Upgrade Prompt Email Subject Test
Upgrade prompts are some of the highest-leverage emails you send — a few extra opens can move real revenue. This guide shows how to run an upgrade prompt email subject test with marginal, the hosted email MCP server that lets your AI agent generate, launch, and measure subject line experiments without leaving your editor.
Why upgrade prompts deserve a subject line A/B test
Upgrade prompt emails sit at a delicate moment: the recipient already uses your product but hasn't paid, or hasn't moved to the next plan. The subject line decides whether they even open the message that explains the value of upgrading. Small wording shifts — urgency, benefit framing, social proof — produce measurably different open rates.
Rather than guessing, a structured subject line A/B test isolates which framing actually earns the open. marginal handles the experiment plumbing so you can focus on the messaging hypotheses you want to validate.
- Benefit-led: "Unlock unlimited projects today"
- Urgency-led: "Your free tier resets in 3 days"
- Curiosity-led: "You're close to a limit you'll want to know about"
- Personalized: "{{first_name}}, here's what Pro adds"
How marginal runs the test through MCP
marginal exposes four MCP tools your AI agent can call directly: generate_variants drafts subject line candidates, launch_test sends them as a managed A/B test, get_results returns open and click metrics, and recommend_next suggests the follow-up experiment. Because marginal is a hosted email MCP server, there's nothing to deploy — connect a client and authenticate with a Bearer API key (marg_live_...).
A typical upgrade prompt email subject test flow looks like this from an AI agent session:
- Ask the agent to generate_variants from your upgrade prompt copy
- Pick the strongest two to four subject lines
- Call launch_test to send the managed A/B test to your segment
- Read open/click and within-test lift metrics via get_results
Reading lift and choosing a winner
marginal reports open and click tracking alongside within-test lift, so you can see which subject line outperformed the others on the same audience under the same conditions. For upgrade prompts, watch click-through as well as opens — a clever subject that gets opens but no upgrade-page clicks isn't the winner you want.
Once a variant wins, recommend_next helps you iterate. Maybe the urgency angle beat the benefit angle; the next test refines the urgency wording or layers in personalization. This is where AI agent email marketing compounds: each experiment feeds the prompt for the next.
Getting started with the endpoint
marginal works with Cursor, Claude Desktop, Claude Code, Windsurf, Cline, Continue, Zed, and OpenAI Codex. Point your client at https://marginal.sh/mcp, add your Bearer key, and the subject-testing tools become available to your agent. The free tier includes 100 experiments per month — plenty to validate an upgrade prompt subject test sequence.
Full setup and tool reference live in the docs at https://marginal.sh/docs/.
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