Back in Stock Email Tests for Subscription Boxes
When a sold-out subscription box returns, the restock email is one of your highest-intent sends. This guide shows how to run a subscription boxes back in stock email test with marginal, the hosted email MCP server, so your AI agent picks the subject line that actually converts waiting subscribers.
Why back in stock emails matter for subscription boxes
Subscription box shoppers often join waitlists for limited drops, seasonal editions, or sold-out tiers. The back in stock notification reaches them at a moment of peak demand — but a flat subject line can bury that urgency in a crowded inbox.
A structured email A/B test removes the guesswork. Instead of debating phrasing internally, you ship two or more subject lines, measure opens and clicks on the same audience, and let the data decide which restock message wins.
- Urgency framing: "Back in stock" vs. "Almost gone again"
- Scarcity cues: limited quantities or restock windows
- Personalization: box name, tier, or waitlist position
- Clear CTA pull-through from subject to landing page
Running the test with marginal
marginal exposes four MCP tools your AI agent can call directly: generate_variants, launch_test, get_results, and recommend_next. For a restock campaign, your agent drafts subject line variants from the box context, launches the managed send, and reads back within-test lift metrics once data lands.
Because marginal is hosted, there's nothing to deploy. Connect your client to the endpoint at https://marginal.sh/mcp, authenticate with a Bearer API key, and the subject line A/B test runs end to end inside your agent workflow.
- generate_variants — produce competing back in stock subject lines
- launch_test — send variants and split tracking automatically
- get_results — pull open and click rates with lift
- recommend_next — suggest the winning angle for the next restock
A practical AI agent workflow
Treat each restock as a repeatable loop. Your AI agent for email marketing can read which box returned to inventory, generate a handful of subject line variants tuned to that product, and launch the test against your waitlist segment.
Once opens and clicks settle, the agent reviews results and stores the winning pattern. Over several drops, you build a library of restock subject lines proven against your own subscribers rather than generic best-practice templates.
- Start narrow: test subject lines before body copy
- Keep segments consistent so lift is comparable
- Re-run recommend_next to compound learnings per box
- Use the free tier (100 experiments/month) to validate before scaling
Getting connected
marginal works with the MCP clients you already use — Cursor, Claude Desktop, Claude Code, Windsurf, Cline, Continue, Codex, and Zed. Point your client at the marginal endpoint, add your Bearer key, and your agent can immediately call the test tools.
See the docs at https://marginal.sh/docs/ for tool schemas and setup, and the MCP endpoint details at https://marginal.sh/mcp to wire up your subscription box back in stock email tests.
- Endpoint: https://marginal.sh/mcp
- Auth: Bearer API key (marg_live_...)
- Registry: sh.marginal/mcp
- Docs: 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