Finding the Push Notification Sweet Spot: The Data Behind Frequency, Timing, and Fatigue
Send too few notifications and users forget you exist. Send too many and they uninstall your app. Here is what 50 million push notifications taught us about finding the balance.
There is a moment in every product team's life when someone asks: "How many push notifications should we send per day?" The honest answer is: it depends. The useful answer requires data.
We analyzed 50 million push notifications across 200 applications to understand the relationship between notification frequency, engagement, and the dreaded opt-out. Here is what we found.
The Frequency Curve
The relationship between push notification frequency and engagement is not linear. It follows a curve that every product team should understand:
1-2 notifications per week: Engagement rates are high (15-25% open rate), but total reach is limited. Users appreciate the restraint but you are leaving engagement on the table.
3-5 notifications per week: The sweet spot for most applications. Open rates remain strong (12-20%), and the increased frequency drives significantly more total engagement. This is where most successful apps operate.
1-2 notifications per day: A risky zone. Open rates drop to 8-12%, and opt-out rates start climbing. However, apps that deliver genuinely time-sensitive content (news, finance, sports) can sustain this frequency.
3+ notifications per day: Unless you are a messaging app or a real-time trading platform, this frequency actively harms your relationship with users. Open rates drop below 5%, and opt-out rates can reach 5-8% per week at this level.
"The optimal notification frequency is not a fixed number. It is the maximum frequency at which your average notification still delivers genuine value to the user."
The Timing Effect
Frequency matters, but timing can be equally impactful. Our data revealed several patterns:
The Golden Windows
Across all app categories, three time windows consistently produced the highest engagement:
8:00-9:00 AM local time: Users are starting their day, checking their phones, and receptive to information. Open rates in this window are 22% higher than the daily average.
12:00-1:00 PM local time: The lunch break check. Users are browsing, and a well-timed notification catches them in a receptive state. Engagement is 18% above average.
6:00-8:00 PM local time: The evening wind-down. Users are relaxed and have time to act on notifications. This window has the highest click-through rate (as opposed to just opens), suggesting users are more willing to take action in the evening.
The Dead Zones
11:00 PM - 7:00 AM: Obvious, but still violated by 23% of the apps we analyzed. Notifications sent during sleeping hours are not just ignored. They generate active resentment and are the number one cited reason for opt-out in user surveys.
Monday 9:00 AM: Counterintuitively, the very start of the workweek is a poor time for most non-work apps. Users are overwhelmed with work notifications and actively filtering out non-essential messages.
Personalized Send Times
The most sophisticated approach is to send each notification at the time when each individual user is most likely to engage, based on their historical behavior. Apps that implement personalized send times see a 40% improvement in open rates compared to a fixed-time approach.
This requires tracking when each user typically opens your notifications and building a per-user optimal send time model. It is worth the engineering investment.
Understanding Notification Fatigue
Fatigue is not just about frequency. It is about relevance decay. A user who receives 5 highly relevant, personalized notifications per week will not feel fatigued. A user who receives 2 generic, irrelevant notifications per week will.
We identified three stages of notification fatigue:
Stage 1: Selective Ignoring (Mild)
The user starts ignoring some notifications but still opens others. Open rate drops 20-30% from baseline. This is a warning signal, not a crisis.
Intervention: Reduce frequency slightly and increase personalization. Review which notification types are being ignored and consider retiring them.
Stage 2: Habitual Dismissal (Moderate)
The user swipes away notifications without reading them. Open rate drops 50-70% from baseline. The user has developed a muscle memory of dismissing your notifications.
Intervention: Stop all non-essential notifications for this user for 7-14 days. When you resume, lead with your single highest-value notification type. You need to re-earn their attention.
Stage 3: Opt-Out or Uninstall (Severe)
The user disables notifications or uninstalls the app entirely. This is usually irreversible.
Intervention: Prevention. Track the decline from Stage 1 to Stage 2 and intervene before Stage 3. Once a user opts out, your only remaining channel is email, and they are probably not thrilled about that either.
Content Type Performance
Not all notifications are created equal. Here is how different content types performed in our analysis:
| Content Type | Avg Open Rate | Avg Click Rate | Opt-Out Impact |
|---|---|---|---|
| Personalized insights | 24% | 8% | Very low |
| Time-sensitive alerts | 22% | 12% | Low |
| Achievement/milestone | 20% | 6% | Very low |
| Social (new follower, etc.) | 18% | 10% | Low |
| Promotional (sale, offer) | 12% | 7% | Medium |
| Generic reminders | 8% | 3% | High |
| Re-engagement ("Come back!") | 6% | 4% | High |
The pattern is clear: notifications that are about the user (their data, their achievements, their social connections) outperform notifications that are about your product (your sale, your feature, your content).
Building a Frequency Policy
Based on our analysis, here is a framework for setting your notification frequency:
Step 1: Categorize Your Notifications
Divide your notifications into three tiers:
- Tier 1 (High value): Directly requested by the user or delivering personally relevant data. No frequency limit needed.
- Tier 2 (Medium value): Product-related but not user-initiated. Limit to 3-4 per week.
- Tier 3 (Low value): Promotional, re-engagement, or generic content. Limit to 1-2 per week.
Step 2: Implement a Daily Cap
Set a hard cap of 2-3 notifications per day per user across all tiers. If a Tier 1 notification needs to go out and the user has already received 3 today, suppress a Tier 3 notification to make room.
Step 3: Monitor Per-User Fatigue Signals
Track each user's notification engagement trend. If their open rate drops below 50% of their personal baseline, automatically reduce their frequency for 2 weeks.
Step 4: Let Users Choose
Give users granular control over notification types and frequency. Users who feel in control opt out less, even at higher frequencies, than users who feel spammed at lower frequencies.
The Bottom Line
Push notification strategy is not about finding a magic number. It is about building a system that:
- Sends the right notifications (relevant, personalized, valuable)
- At the right frequency (enough to be useful, not enough to annoy)
- At the right time (when the user is receptive)
- With the right fallbacks (fatigue detection, frequency capping, user control)
The companies that do all four consistently achieve 3-5x the engagement of companies that optimize for only one or two. And more importantly, their users want to receive their notifications, which is the only sustainable foundation for long-term engagement.