How do you test placement with your real subscriber list safely?

You've got a new campaign ready and you want to know it'll land in the inbox before you hit send to everyone. Testing on your real list is the right instinct. The trick is doing it in an order that protects your reputation if something goes wrong.

Who counts as "engaged enough" to test first

Your safest test group is subscribers who have opened or clicked at least once in the last 30 days. That's your active core. If you have fewer than 500 of those, widen the window to 60 days and pull everyone who clicked (not just opened) in that window. Clicks are a stronger signal than opens, especially after Apple Mail's open rate inflation blurred the numbers.

Avoid mixing in subscribers who haven't engaged in 90+ days for your test group. Those contacts are exactly where placement problems surface first.

How big should the test send be

If your list is under 10,000, start with 10% of your most engaged contacts. If your list is 10,000 to 100,000, a flat 500 to 1,000 contacts is enough to get a signal. Above 100,000, you can work with 2 to 5% of your engaged segment before opening to the full list.

Don't test on a number so small that the data is meaningless. 50 people is not a useful signal. You need at least a few hundred sends to spot a delivery rate drop or a complaint spike that matters.

What to watch and when to stop

Give your test send about 2 to 4 hours before drawing conclusions. Here's what you're looking for in that window:

• Delivery rate should be above 98%. Anything under 95% is worth stopping and investigating before the full send.
• Spam complaint rate should stay below 0.08%. If it crosses 0.1%, stop the rollout. That threshold is what Gmail uses as a warning signal, and exceeding it repeatedly puts your sender reputation at risk.
• Open rate for your most engaged segment should roughly match your recent baseline. A sharp drop (more than 10 percentage points below your average) can mean spam folder placement.
• Bounces: a hard bounce rate above 2% on a list you've mailed before is unusual. Pause and check for data issues before continuing.

Most ESPs show these metrics inside their dashboard. Mailchimp, Klaviyo, and Brevo all update delivery and complaint data within an hour of sending. If you're on a platform that delays reporting by 24 hours, that's worth factoring into your timing.

The staged rollout

Once your test group looks clean, expand in waves rather than all at once. A simple pattern that works:

1. Send to engaged segment (opened or clicked in last 30 days). Wait 4 hours.
2. If metrics hold, expand to anyone who opened or clicked in the last 90 days. Wait 2 to 4 hours.
3. If metrics still hold, send to the rest of your active list.
4. Leave anyone who hasn't opened or clicked in 6 months for a separate re-engagement campaign. Don't roll them into this send.

So the whole point of staging is that each wave generates positive signals that build a small buffer before the riskier segments see the email.

Isolating a risky test

If you're specifically testing something that might perform differently, like a new content format or an unusual offer, consider sending from a subdomain rather than your root sending domain. Something like sends.yourbrand.com keeps any complaints from directly dinging your main domain reputation while you gather data. You'd need to set up authentication records for that subdomain first, but it's a clean way to test without betting your whole sender history on one unknown.

If you're unsure whether a sudden placement dip is from your test or something else already in motion, our SOS hotline is free. We'll help you read the signals without the pitch.