Test Cross-Sell Subject Lines in Continue
Continue lets you drive email experiments straight from your editor. Connect it to marginal's email MCP server and you can draft, launch, and measure cross-sell subject lines without leaving your workflow.
Why cross-sell campaigns need subject line testing
Cross-sell emails live or die in the inbox. A customer who already bought one product won't open a recommendation unless the subject line frames the next purchase as obvious value. Small wording changes — 'Complete your setup' versus 'You might also like' — can swing open rates meaningfully.
A subject line A/B test removes the guesswork. Instead of debating phrasing in a doc, you ship two or more variants to a sample and let opens and clicks decide which framing earns the cross-sell conversation.
- Frame the second product as completing or upgrading the first purchase
- Compare urgency-led vs. benefit-led subject lines
- Test personalization tokens against generic phrasing
- Measure open and click lift before a full send
Connecting Continue to marginal
marginal is a hosted email MCP server, so there's no infrastructure to run. Add it to Continue as a remote MCP endpoint at https://marginal.sh/mcp and authenticate with your Bearer API key (marg_live_...). Once registered, the tools appear to Continue's agent like any other capability.
The four tools cover the full loop: generate_variants drafts cross-sell subject lines, launch_test sends the experiment, get_results reports open and click metrics, and recommend_next suggests where to take the campaign.
- Registry: sh.marginal/mcp
- Endpoint: https://marginal.sh/mcp
- Auth: Bearer API key (marg_live_...)
- Docs and setup: https://marginal.sh/docs/
Running a cross-sell subject line test from your editor
Prompt Continue with the cross-sell context — the product the customer already owns and the accessory or upgrade you want to promote. The agent calls generate_variants to produce candidate subject lines, then launch_test to send them as a managed A/B test through marginal.
After the send, ask Continue to pull get_results. You'll see within-test lift metrics on opens and clicks, so you can declare a winner with evidence rather than instinct. recommend_next then proposes the follow-up — broaden the winning angle or test a new hook.
- Describe the owned product and the cross-sell offer in plain language
- Let the AI agent generate and launch competing subject lines
- Review open/click lift per variant in get_results
- Use recommend_next to plan the follow-up campaign
Scaling AI agent email marketing
Because Continue keeps the experiment loop inside your coding environment, cross-sell testing becomes a repeatable workflow rather than a one-off task. You can wire it into release notes, post-purchase flows, or seasonal pushes and let the agent iterate on subject lines as you ship.
The free tier covers 100 experiments per month, which is plenty to validate cross-sell messaging across several product pairings before committing to a broader rollout.
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