Travel Cross-Sell Email Test
Cross-selling is where travel brands earn their margin: a booked flight is just the start of seat upgrades, hotels, insurance, and tours. This guide shows how to run a travel cross-sell email test with marginal, the hosted email MCP server, so your AI agent can generate, launch, and measure variants without leaving your editor.
Why cross-sell deserves a dedicated test
In travel, the post-booking window is short and intent is high. A traveler who just confirmed a flight is far more likely to add a hotel, car, or airport transfer in the next few days than weeks later. That makes the cross-sell email worth treating as its own optimization problem rather than a footnote to your confirmation flow.
Small wording changes move real revenue here. Whether you lead with savings, convenience, or destination inspiration changes how many travelers open and click. A structured email A/B test removes the guesswork.
- Test offer framing: bundle discount vs. limited-time upgrade
- Compare urgency language against destination imagery cues in the subject
- Segment by trip type — leisure, business, family — for sharper variants
Setting up the test with marginal
marginal exposes four MCP tools your agent uses end to end: generate_variants drafts subject line and copy options, launch_test ships a managed send, get_results returns open and click metrics, and recommend_next suggests the follow-up experiment. You connect once at https://marginal.sh/mcp with a Bearer API key and drive everything in natural language.
For a travel cross-sell campaign, point your agent at the segment of recently booked travelers, describe the add-on you want to promote, and let it produce several angles. A subject line A/B test runs first because the subject gates every downstream metric.
- generate_variants — produce 3–5 cross-sell subject and body options
- launch_test — managed send with open/click tracking built in
- get_results — within-test lift metrics per variant
- recommend_next — propose the next iteration based on results
Driving it from your AI agent
Because marginal is an email MCP server, AI agent email marketing happens inside the tools you already use. Connect from Cursor, Claude Code, Claude Desktop, Windsurf, Cline, Continue, Zed, or OpenAI Codex and prompt your agent to spin up the cross-sell test, then check back for results.
A typical loop: ask the agent to generate variants for a hotel add-on, launch the email A/B test, wait for engagement to accumulate, then have it read get_results and call recommend_next. The free tier covers 100 experiments per month, enough to iterate on subject lines across multiple travel segments.
- No self-hosting — marginal is fully hosted
- Works with your existing agent client out of the box
- Docs and tool reference at https://marginal.sh/docs/
Reading lift and iterating
Once a variant pulls ahead, use within-test lift metrics to confirm the winner is meaningful, not noise. For cross-sell specifically, watch click-through closely since that signals intent to add the upsell, not just curiosity about the subject line.
Then keep the cycle going. Feed the winning angle into recommend_next and let your agent propose the following subject line A/B test — perhaps swapping the add-on category or testing send timing relative to the original booking.
- Confirm winners with lift before rolling out broadly
- Prioritize click metrics for revenue-driving cross-sells
- Chain experiments so each test informs the next angle
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