Shopify Back in Stock Email A/B Tests
Back in stock alerts are some of the highest-intent emails a Shopify store sends. A small lift in open rate can recover real revenue from shoppers who already wanted the product. This guide shows how to run a Shopify back in stock email test with marginal, the hosted email MCP server your AI agent can drive directly.
Why back in stock emails deserve their own test
When an item returns to inventory, the window of intent is short. Shoppers who set up a restock alert are deciding fast whether to buy now or move on. The subject line is what gets that email opened before the moment passes — and that makes it a perfect candidate for a focused subject line A/B test.
Generic 'It's back!' copy leaves performance on the table. Testing urgency, product specificity, and scarcity framing against each other tells you what actually moves your Shopify audience to open and click through to the product page.
- High purchase intent: subscribers explicitly asked to be notified.
- Time-sensitive: restocks often sell out again, so opens matter immediately.
- Clear conversion path: one product, one call to action.
- Repeatable: every restock is another chance to apply what you learned.
Running the test with marginal
marginal exposes four MCP tools your agent calls in sequence: generate_variants to draft subject line options, launch_test to start a managed send, get_results to read open and click metrics, and recommend_next to surface the winner and suggest a follow-up. Because marginal is hosted, there's no infrastructure to stand up — connect over the MCP endpoint and start testing.
For a back in stock campaign, ask your agent to generate variants around the specific product and the restock event, then launch the email A/B test against your notify-list segment. marginal handles the send, splits the audience, and tracks engagement so you get within-test lift metrics rather than guesses.
- generate_variants — restock subject lines emphasizing product name, scarcity, or savings.
- launch_test — managed send with automatic audience split.
- get_results — open and click tracking with per-variant lift.
- recommend_next — the winning variant plus a suggested next experiment.
Subject line angles worth testing
Treat each restock as a fresh subject line A/B test. Vary one dimension at a time so you can attribute the lift cleanly, and feed the winners back into your next campaign.
- Product-specific: "The [product] you wanted is back in stock".
- Scarcity: "Restocked — but only a few left".
- Personalized intent: "You asked, it's back".
- Benefit-led: "Back in stock and ready to ship".
- Emoji vs. plain text to gauge inbox tone with your Shopify list.
Let an AI agent own the loop
AI agent email marketing works best when the agent can act end to end. With marginal connected through clients like Cursor, Claude Desktop, Claude Code, Windsurf, Cline, Continue, or Zed, your agent drafts variants, launches the back in stock test, reads results, and recommends the next iteration without leaving the editor.
The free tier includes 100 experiments per month, which is plenty to test restock subject lines across your catalog. Point your client at https://marginal.sh/mcp with a Bearer API key, and see the setup details at https://marginal.sh/docs/.
- Hosted endpoint: https://marginal.sh/mcp — no self-hosting required.
- Bearer auth with a marg_live_ API key.
- Works across all major MCP-capable coding agents.
- Free tier: 100 experiments/month to iterate on restock campaigns.
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