Email marketing is a leaky pipe. Learn where it leaks.
Every campaign is the same five-stage funnel: sent → delivered → opened → clicked → converted. Each stage is a percentage, and the percentages multiply — which is why 10,000 emails can end as 11 sales. Model your own campaign below.
Drag to rotate · disc area = people remaining
2024-era benchmarks to sanity-check against
“Good” depends on industry, but cross-industry medians are remarkably stable. If your numbers are far below these, the leak is usually list quality (stale addresses), subject lines (opens), or offer-to-audience fit (clicks).
| Metric | Typical median | Strong |
|---|---|---|
| Deliverability | 98–99% | >99% with clean lists |
| Open rate | 19–23% | 30%+ (segmented) |
| Click-through of opens | 7–11% | 15%+ |
| Click-to-conversion | 2–5% | 8%+ |
| Unsubscribe per send | 0.1–0.3% | <0.1% |
Where “data mining” fits
The data-mining half of that internship is the part that moves the funnel. Instead of one blast, you cluster subscribers by behavior — recency, frequency, monetary value (RFM analysis) — and by attributes like signup source or past category purchases. Then each segment gets a different subject line, send time, and offer.
Worked example: at the defaults above (10,000 sent), batch-and-blast yields ~10 orders ≈ $480. Toggle segmentation on: opens rise to ~27%, clicks to ~15% of opens, and the same list produces ~20 orders ≈ $960 — double the revenue with zero extra sends. That multiplication effect is why segmentation is the highest-ROI skill in email, and why the classic stat that email returns roughly $36 per $1 spent (Litmus) is driven almost entirely by targeting quality.
Legal floor: CAN-SPAM (US) and CASL (Canada) require a working unsubscribe honored within 10 business days, a real postal address, and no deceptive subject lines. GDPR-covered lists additionally require provable opt-in consent.