How do you know when a segment is too small?
Still have a question, spotted an error, or have a better explanation or a source we should cite?
You've got a segment in mind, but you're wondering if it's worth the effort. Here's how to tell if it's actually actionable.
For A/B testing: You need at least 3,000 to 5,000 people. Why? Random variation. A segment of 500 might show a 15% open rate, but that could genuinely swing to 12% or 18% in another week. The number alone won't tell you if your email changes actually worked.
For custom campaigns: Do the math on effort versus payoff. If building a custom message takes 4 hours but only generates $200 in revenue, your segment's probably too small. Merge it with similar groups instead.
For deliverability: Mailbox providers watch sending patterns. A one-time micro-campaign looks strange to them. They prefer consistent volume. Very small, sporadic sends can actually hurt your reputation.
What's the baseline? If your list has fewer than 10,000 people, segments under 500 are usually too small. If you're over 100,000, go smaller than 2,000 and you're losing signal.
That said, small segments aren't worthless. You can merge them with related groups or use them to exclude people from other campaigns. But if a segment stands alone and needs custom treatment, question whether it's really a segment or just a one-off.
Want to know if your current segments are sized right. Start by auditing your segments and checking which ones are actually pulling their weight.
Contributors
Who worked on this answer
Every name links to their profile. Every company links to their site. Real people, real accountability.
