Email Frequency & Send-Time Optimization
Segmentation tells you who gets which message. It never tells you how often to hit send, or what hour to hit it.
Get cadence wrong and even a perfectly segmented, perfectly written email lands in a fatigued inbox. Get timing wrong and a great email sits unread under forty others. Both are fixable, and both have real 2026 data behind them.
Quick Summary
- MailerLite's 2025 analysis of 12 billion emails found unsubscribe rate bottoms out at twice-weekly sends (0.33%) and roughly doubles for both daily blasts and sub-monthly trickles (0.87%).
- Salesforce's 19-billion-send frequency study found subscribers getting 5+ emails a week unsubscribed at 0.58%, versus 0.07% for 1 to 2 emails a week, an 8x gap.
- There is no single "best time to send", best time varies by list, industry, and individual subscriber.
- AI-driven per-subscriber send-time optimization (Klaviyo's Personalized Send Time) beat one-size-fits-all scheduling by up to 35% on click rate in beta.
- Your frequency ceiling is not a number you look up, it is a threshold you find by testing upward until unsubscribes or complaints tick up, then backing off one notch.
The Frequency-Fatigue Tradeoff
More email is not automatically worse, and less email is not automatically safer. The data shows a curve, not a straight line.
MailerLite's 2025 study of 1.4 million campaigns found sub-monthly senders had the highest unsubscribe rate of any group, 0.87%, nearly double the twice-weekly rate of 0.33%. Subscribers who forget why they signed up leave just as fast as subscribers who feel spammed.
| Frequency | Open Rate | Click Rate | Unsubscribe Rate |
|---|---|---|---|
| Less than monthly | 35.11% | 3.74% | 0.87% |
| 1-3 times/month | 33.48% | 4.32% | 0.54% |
| Weekly | 33.22% | 4.87% | 0.38% |
| Twice a week | 32.98% | 5.31% | 0.33% |
| Daily | 30.04% | 4.97% | 0.36% |
The sweet spot in this dataset sits between weekly and twice-weekly, the highest click rate, the lowest unsubscribe rate, before daily sending starts eroding both.
Salesforce's separate analysis, run across 19 billion sends, found the same pattern from a different angle: brands sending 1 to 2 emails a week saw a 0.07% unsubscribe rate. Brands sending more than 5 a week saw 0.58%, an 8x jump. Newsletter creators publishing more than three issues a week saw opt-outs spike to 0.39%, an 85% increase over once-weekly senders.
Unsubscribes are the visible cost. Spam complaints are the expensive one. A healthy unsubscribe rate sits below roughly 0.2 to 0.3%, with anything past 1% signaling a targeting or frequency problem. But spam complaints do more damage per incident: they feed directly into the sender-reputation score that Gmail, Outlook, and Apple Mail use to decide whether your next email, to everyone, lands in the inbox or the spam folder. Frequency fatigue rarely shows up as a wave of unsubscribes first. It shows up as a slow decline in open rates as inbox providers quietly start filtering you, followed by an unsubscribe spike weeks later once subscribers notice you in spam and go looking for the option to leave. Watch open-rate trend lines, not just the unsubscribe count.
Momentum check: frequency has a ceiling, and the ceiling is measurable. The next question is when, inside your chosen frequency, to actually pull the trigger.
Fixed Best Times vs. Per-Subscriber AI Optimization
For a decade, "best time to send" meant a generic answer like Tuesday at 10am, based on aggregate open-rate studies across thousands of accounts.
That answer still isn't wrong, it's just averaged away the thing that matters: your list is not the average list. A B2B SaaS audience checks email between meetings; a consumer fashion list checks it on the couch after dinner. A single fixed time optimizes for the middle of your list and underserves both ends.
Two levels of send-time tooling exist today:
- List-level "smart send time." Tools like Klaviyo's Smart Send Time and Mailchimp's Send Time Optimization analyze your account's own historical opens and clicks, then pick one best time for the whole send. Better than a generic Tuesday-10am rule of thumb, still one timestamp for everyone.
- Per-subscriber AI send-time optimization. Klaviyo's Personalized Send Time goes further: using reinforcement learning, it predicts a different optimal moment for each individual profile within your sending window, based on that person's own engagement history and patterns from similar subscribers. During Klaviyo's beta, top-performing campaigns saw a 35% increase in click rate, and premium sunglasses brand Shady Rays measured a 10%+ lift in placed-order rate across 30+ campaigns using it.
Per-subscriber optimization needs a cold-start runway. Without roughly 2 to 4 weeks of engagement history per subscriber, the model is guessing rather than learning, and a well-chosen manual send time will beat it. It also works best for non-urgent sends like newsletters and product announcements. Anything time-sensitive, a flash sale ending at midnight, a webinar starting in an hour, should still go out at a fixed time to everyone, since a coordinated deadline beats individually-optimized opens.
Momentum check: the tooling question has a clean answer, use AI send-time optimization once you have the data and non-urgent content to justify it. The frequency question is harder, because your ceiling is specific to your list.
Finding Your Own Frequency Ceiling
Benchmarks tell you where other lists land. They do not tell you where yours does. Use this four-step framework to find your ceiling directly.
Step 1, Start at your current frequency and log the baseline. Record open rate, click rate, unsubscribe rate, and spam-complaint rate for your last 4 sends at whatever cadence you're on now.
Step 2, Increase by one send per week, hold for a full cycle. If you send weekly, try twice-weekly for 3 to 4 sends. Do not change content strategy at the same time, isolate frequency as the only variable.
Step 3, Watch two numbers, not one. Unsubscribe rate rising slightly is normal and often fine. Spam-complaint rate rising, or open rate trending down across the cycle, means you have crossed the ceiling.
Step 4, Back off one notch and lock it in. If twice-weekly triggered a complaint spike, weekly is your ceiling for now. Revisit the test in 3 to 6 months, ceilings shift as your list grows and your content improves.
Segment before you scale frequency. The engagement-tier segments from the segmentation lesson (active, cold, lapsed) are the natural place to apply different ceilings. Your active segment can often absorb twice-weekly sends happily. Sending that same frequency to your cold segment is exactly how you manufacture the spam complaints this framework is trying to catch.
Key Takeaways
- The frequency-unsubscribe relationship is a curve: too rare and too frequent both spike unsubscribes, the low point sits between weekly and twice-weekly for most lists.
- 5+ emails a week is the danger zone in the Salesforce data, unsubscribe rate jumps 8x versus 1 to 2 a week.
- Fixed best-time sending still works, and remains the right choice for time-sensitive campaigns and lists without enough engagement history.
- Per-subscriber AI send-time optimization outperforms fixed timing once you have 2 to 4 weeks of data and non-urgent content, up to 35% higher click rates in Klaviyo's reported beta results.
- Find your frequency ceiling by testing one notch up at a time and watching spam-complaint and open-rate trends, not just the unsubscribe count.