Predictive Send Time
Using a subscriber's engagement history to predict the specific moment they're most likely to open an email.
Definition
Predictive send time is the technique of analyzing a subscriber's historical email engagement, such as when they've opened and clicked in the past, to predict the specific time they're statistically most likely to open a future email. It's closely related to send-time optimization: send-time optimization is the name most platforms give to the feature, while predictive send time describes the underlying technique that powers it.
Why It Matters
A single fixed send time can never be right for everyone on a list. Predictive send time personalizes delivery timing per subscriber instead of per campaign, which typically improves open rates because more emails arrive while a subscriber is actually checking their inbox rather than sitting unread for hours.
How It Works
A model looks at each subscriber's history of opens and clicks, identifying patterns like which hours and days they tend to engage. Combined with timezone data, this produces a predicted optimal send window for that specific person. New subscribers without enough history are typically assigned a reasonable default based on their timezone or aggregate patterns from similar subscribers, with predictions improving as more of their own engagement data accumulates.
Best Practices
- 1Give new subscribers a sensible default rather than leaving them unscheduled while data accumulates
- 2Set a maximum delivery window so time-sensitive campaigns don't get delayed too long
- 3Re-evaluate predictions periodically, since a subscriber's habits can change over time
- 4Use predictive send time for recurring sends like newsletters, where individual timing has the most room to help
- 5Disable predictive timing for genuinely time-critical sends, like a flash sale ending in hours
How Predictive Send Time Is Calculated
| Input | What it contributes | When it matters most |
|---|---|---|
| Historical opens and clicks | Engagement timing signals; opens can include automated privacy prefetch | Once a subscriber has meaningful history |
| Timezone | A local-time baseline | Always, especially for new subscribers |
| Aggregate patterns from similar subscribers | A cold-start estimate | New subscribers with little or no history |
| Recency weighting | Prioritizes recent behavior over old patterns | Subscribers whose habits have changed |
Predictive Send Time vs Timezone Sending
| Approach | What it optimizes for | Limitation |
|---|---|---|
| Timezone sending | Same local clock time for everyone | Ignores individual habits within a timezone |
| Predictive send time | Individual engagement patterns per subscriber | Needs engagement history to be accurate |
| Combined approach | Timezone as a baseline, refined by individual history | Requires a platform that supports both natively |