Testing Loyalty Reward Subject Lines in Cursor
Loyalty reward emails only earn their open if the subject line makes the perk feel real. This guide shows how to draft, launch, and measure loyalty reward subject lines in Cursor with marginal, the hosted email MCP server for AI agents.
Why loyalty rewards need their own subject line test
A loyalty reward campaign speaks to people who already trust your brand, so the subject line has a different job than acquisition mail. It has to confirm value, signal exclusivity, and avoid sounding like every other promo. Small wording choices — "Your 500 points are ready" versus "A reward just unlocked" — move open rates in ways guesswork can't predict.
Running a subject line A/B test removes the argument. Instead of debating tone in a thread, you launch competing variants to a sample and let opens and clicks decide. With marginal wired into Cursor, that whole loop lives next to the code that triggers the send.
- Reward specificity vs. curiosity: name the perk or tease it
- Urgency framing for expiring points or tier deadlines
- Personalization tokens like member name or tier level
- Emoji usage and how it lands with a loyal segment
Connecting marginal to Cursor
marginal is a hosted email MCP server, so there's nothing to deploy. Add the endpoint to Cursor's MCP configuration with your Bearer API key, and the agent can call marginal's tools directly while you work. No Mailchimp or Klaviyo credentials are required — sends and tracking run through marginal.
Once connected, Cursor's AI agent can reach four tools: generate_variants to draft subject lines, launch_test to start the experiment, get_results to pull live metrics, and recommend_next to surface the winning direction.
- Endpoint: https://marginal.sh/mcp
- Auth: Bearer API key in the form marg_live_...
- Registry id: sh.marginal/mcp
- Free tier covers 100 experiments/month for early testing
Running a loyalty reward subject line test
Describe the campaign to Cursor in plain language — the reward, the audience tier, and the tone you want — and let the agent call generate_variants for a spread of subject lines. Review them inline, trim the weak ones, and ask it to launch_test against your loyalty segment.
marginal handles the managed send and tracks opens and clicks per variant, reporting within-test lift so you can see which subject line outperforms and by how much. When results stabilize, recommend_next tells you which direction to scale for the full reward blast.
- Prompt: "Generate 5 loyalty reward subject lines for our gold-tier members"
- Launch a subject line A/B test to a holdout sample
- Read within-test lift on open and click rates via get_results
- Use recommend_next to pick the variant for the broad send
Making AI agent email marketing repeatable
Because the workflow runs through MCP, you can turn it into a routine your agent repeats every reward cycle. Each new loyalty drop becomes another quick test rather than a fresh round of guessing, and the lift data compounds into a clearer picture of what your loyal audience responds to.
See the tool reference and config details in the marginal docs at https://marginal.sh/docs/, and point Cursor at https://marginal.sh/mcp to start testing loyalty reward subject lines today.
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