DTC brands review request email test
Post-purchase review requests are some of the highest-leverage emails a DTC brand sends — they fuel social proof, repeat purchases, and ad creative. This guide shows how to run a review request email test with marginal, the hosted email MCP server, so an AI agent can generate variants, launch, and measure lift without leaving your editor.
Why review requests need an A/B test for DTC brands
DTC brands live and die by reviews. A few percentage points of extra open and click on your review request flow compounds into more star ratings, more user-generated content, and stronger paid acquisition. But the gap between a request people ignore and one they act on usually comes down to framing — and that's exactly what an email A/B test is built to settle.
Instead of guessing whether to lead with the product name, a thank-you, or an incentive, you let real opens and clicks decide. marginal handles the subject line A/B test, manages the send, and reports within-test lift so you can ship the winner with confidence.
- Reviews drive social proof that lowers acquisition cost.
- Small subject line wins compound across every order.
- Timing and tone vary by category — test, don't assume.
- Within-test lift metrics keep the comparison honest.
Subject line angles worth testing
A review request email test usually starts at the subject line, where the open-rate battle is won. For DTC brands, the strongest angles tend to balance gratitude with a clear, low-friction ask. Generate a handful of distinct variants and let the test prune the weak ones.
- Direct ask: "How are you liking your [product]?"
- Gratitude-first: "Thanks for your order — one quick favor?"
- Social proof: "Help the next shopper decide"
- Incentive-led: "Leave a review, get 10% off your next order"
- Personal: "[First name], 30 seconds for a quick rating?"
Running the test with an AI agent and marginal
marginal exposes four MCP tools — generate_variants, launch_test, get_results, and recommend_next — over a hosted endpoint at https://marginal.sh/mcp. Connect any MCP client and your AI agent can run the whole review request loop conversationally: draft variants, launch the send, read open and click tracking, and get a recommendation on what to send next.
Because marginal is fully hosted, there's no infrastructure to stand up. Authenticate with a Bearer API key, point your client at the endpoint, and start experimenting. AI agent email marketing workflows fit naturally here — the agent does the iteration, you approve the call.
- generate_variants — draft review request subject lines and copy.
- launch_test — run a managed send across your variants.
- get_results — pull opens, clicks, and within-test lift.
- recommend_next — let marginal suggest the follow-up experiment.
Connect your stack and start small
marginal works with Cursor, Claude Desktop, Claude Code, OpenAI Codex, Windsurf, Cline, Continue, and Zed, so you can run a review request email A/B test from whatever tool you already build in. The free tier includes 100 experiments per month — plenty to find a winning subject line for your post-purchase flow.
Begin with one variable. Lock your audience and send time, test two or three subject lines, and let the data name the winner before you layer in body-copy or incentive tests. Full setup details live in the docs at https://marginal.sh/docs/.
- Start with a single subject line A/B test, not five at once.
- Keep send timing consistent so opens are comparable.
- Promote winners into your evergreen review request flow.
- Use recommend_next to plan the follow-up experiment.
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