Testing Whether to Mention the Discount in Your Subject Line
"Save 20% today" versus "Something new just for you" — naming the discount up front is one of the oldest debates in email marketing. This guide shows how to settle it with a discount mention subject line A/B test driven by an AI agent through marginal.
Why the discount-mention question matters
Putting the discount in the subject line tends to lift opens from deal-seekers, but it can also train your list to wait for promotions and can hurt perceived brand value. The only honest way to know how your audience reacts is to run a controlled subject line A/B test rather than guess.
A discount mention subject line A/B compares two framings of the same send: one that states the offer explicitly ("30% off ends tonight") and one that withholds it or uses curiosity. Same audience, same body, same send window — the subject line is the only variable.
- Explicit: surfaces the value immediately, good for price-driven segments
- Implicit: protects margin and curiosity, can over-promise less
- Hybrid: hints at savings without a number ("a treat inside")
Running the test with marginal
marginal is a hosted email MCP server that exposes a small set of tools to your AI agent. Once connected to the endpoint at https://marginal.sh/mcp with a Bearer API key, your agent can draft, launch, and read the experiment without you leaving the editor.
- generate_variants — produce explicit vs. implicit discount subject lines
- launch_test — split the audience and send both arms
- get_results — pull opens, clicks, and within-test lift
- recommend_next — get a data-backed suggestion for the follow-up send
Reading the results without fooling yourself
A discount-mention subject line will often win on open rate simply because it screams a deal — but opens are not the goal. marginal reports open and click tracking plus within-test lift so you can check whether the explicit variant actually drove more clicks, not just more curious peeks.
Watch for the trap where the explicit discount wins opens but the implicit version wins clicks per open, signaling higher-intent engagement. Let the click metric and lift decide which framing your agent recommends for the next campaign.
A workflow for AI agent email marketing
marginal connects to Cursor, Claude Desktop, Claude Code, Codex, Windsurf, Cline, Continue, and Zed, so the entire loop lives where you already work. Ask your agent to test the discount framing, ship it, and report back — no separate dashboard, no manual list splitting.
The free tier covers 100 experiments per month, enough to test discount mentions across several campaigns before committing to a house style. See the docs at https://marginal.sh/docs/ for tool schemas and setup.
- Hosted server — nothing to deploy or self-host
- Authenticate with a marg_live_ Bearer key
- Iterate weekly until your discount-framing strategy is settled
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