How do I A/B test sending frequency?

You already know that sending more often isn't automatically better. But how do you find the actual sweet spot for your list? That's exactly what a frequency test is for.

Unlike subject line tests, frequency tests take time. You're watching behavior compound over weeks, not hours. A good setup looks like this.

Set up your segments

Split your list into at least three equally-sized groups. A common starting point is weekly, twice weekly, and three times weekly. Each group should have at least 500 to 1,000 subscribers to get readable results (more is better, but that's the floor). Make sure you randomize the split so engagement history is roughly equal across segments.

Here's the part most people skip: keep a control group. One segment stays at your current frequency unchanged. That's your baseline. Without it, you can't tell whether any changes in engagement are caused by frequency or just normal seasonal fluctuation.

Run it long enough

Frequency effects don't show up overnight. Run your test for at least 60 days, and 90 days is better. You're watching for patterns like gradual open rate decline, unsubscribe accumulation, and complaint trends. Short tests almost always make higher frequency look better than it actually is, because the fatigue hasn't set in yet.

Track the right metrics

Don't just look at per-email open rates. Those can actually go up at lower frequencies even when total engagement goes down. What you want to measure is the full picture across the test window:

• Total unique opens per subscriber (not just per send)
• Total clicks per subscriber
• Cumulative unsubscribe rate per segment
• Spam complaint rate per segment
• Revenue or conversions per subscriber (if trackable)

The winning frequency isn't the one with the best individual email stats. It's the one that generates the most cumulative engagement and the fewest exits over the full test period.

Watch for fatigue signals

Now if open rates in a segment start declining week over week while your control holds steady, that's fatigue. If unsubscribes spike in week three of your higher-frequency group, that's your answer. These signals matter more than any single metric snapshot.

One more thing

Optimal frequency isn't the same for your whole list. Your most engaged subscribers can usually handle more email. Your less active ones are already on the edge. Once your test wraps up, it's worth looking at whether frequency changes affect different engagement tiers differently, because they almost always do.

If you're not sure how to read the results once you have them, or if your list is too small to split cleanly, feel free to reach out via our SOS hotline. Free, no pitch, just help.