After analyzing 670 million mobile app push notifications, we saw several interesting trends:

Push notifications are the cornerstone of every mobile app’s user engagement and retention strategy, yet we know so little about them. Andrew Chen, head of user growth at Uber, has previously written about why 60% of users choose to turn off push notifications and why some push notifications have a click-through rate of up to 40%. This time, he analyzed 671 million push notifications based on push notification data from the mobile marketing automation tool Leanplum and found some interesting trends, especially in terms of the timing of push notifications.

Push notification activity during workdays in North America

Let’s first look at the hourly push notifications marketers send and how users interact with those push notifications. The following chart shows push notification send and open metrics for an average weekday in North America, based on a sample of millions of notifications. Data is normalized to the local hour and represents raw sends and opens for each scheduled hour. The blue line shows the raw number of sends, using the left axis. The red line shows the raw number of opens, using the right axis.

You can see an interesting trend in how push notifications are sent and opened throughout the day, with a small peak around noon, a slightly larger peak around 3pm, and the largest peak in the evening. The trend in the evening is interesting; after 6pm, push open rates start to rise relative to the previous hour, while push counts become lower. This suggests that while mobile apps send a large number of push notifications in the evening, push notifications during the evening hours may be more effective as user engagement appears to be high.

No matter how you look at it, this curve is very interesting and we should observe a typical day to understand the behavioral patterns of how people spend their day.

Research on daily activities reinforces this trend

The figure below simulates the average minute-by-minute activities of 1,000 people through a visual way, about what they are doing, such as making phone calls, exercising, shopping, and working. This indirectly proves that the above trend is reasonable. You can watch the full visualization by clicking this link.

Between 7am and 9am, people wake up and start their morning activities: eating, personal care, commuting, and starting work. These consecutive morning tasks and productivity activities explain why user push notification engagement is low before 12 noon. Around 3pm, we can see a correlation between a shift in daily activity and higher push and open rates. This could be because people take a break to have coffee or go out, and these are the times when you often check your phone. By about six in the evening, most people have left work and transitioned to leisure activities.

This explains why push notification open rates are so high starting between 6pm and 9pm. It’s already off work, and people are likely on their phones, browsing apps and social media . Leisure activities last for a few hours, and after ten o'clock in the evening, most people move on to personal care and sleep. The peak period driving notification engagement is similar in duration to the leisure time after work.

Media consumption

So we see that leisure activities are correlated with higher push notification open rates, but what exactly are people doing during these leisure times? Channel Shows’ daily media consumption charts reveal some interesting trends:

A few points to note:

There are two main peaks in Internet usage: one is 8 am and the other is 7 pm.

Usage of iOS and Android apps starts to rise around 7 a.m. and peaks at 9 p.m.

Television clearly belongs in after-get off work time, with a big peak between 7 and 11 p.m.

As you can see from these two icons, TV seems to be an important part of leisure time, especially after get off work. With these two graphs, you can see that TV seems to be an important part of leisure time, especially after get off work. Voice communications (such as phone calls) tend to occur during the day, while online communications (SMS/email) occur throughout the day.

Push Notification Engagement vs. Media Consumption

By combining this daily pattern with Leanplum push notification engagement data, we can speculate why people choose to be pushed in the evening. I bet we're seeing the shitty phone-as-second-screen effect, with people getting push notifications while casually watching TV. They probably have their phone in their pocket or close at hand and can easily open the app.