Few months back I was tasked for one assignment - Optimal time to send push notifications for customers. I was looking at segmenting the shared datasets and figuring out trends/patterns to drive recommendations accordingly. Today after reading some related articles I got a better perspective on the ML behind push notifications.
Some Key learnings are
Personalized push notifications enabled by artificial intelligence
What is the best time to send push notifications?
Insights from Analyzing 1.5 Billion Push Notifications
Happy Learning!!!
Some Key learnings are
- Time Zone based Analysis
- Delivery time vs Notification Open rate duration
- Recommend Optimal time based on user app engagement patterns
- Personalized times based on app usage history
- Windowing of Sent Time
- Windowing of Read Time
- Windowing on Count of Emails sent
- 70% customers received more than 5 emails
- 30% customers received less than 5 mails
- Develop Model with < 5, > 5 to predict Predict Average Read time for different windows
- Combination of rules based, linear regression, classification based approach
- Users Login time
- Users Engagement Window period (Early Morning / Lunch / Evening)
- How many times user checks the App
- Earliest login time in a day
- Login time during weekends
- Message check duration between weekday vs weekend
- Respose for different product categories / age groups
- New categories - News / Travel / Games category based analysis
- Region
- Location
- Career / Handset Type
- Age / Gender
- Android vs Ios Based Analysis
Personalized push notifications enabled by artificial intelligence
What is the best time to send push notifications?
Insights from Analyzing 1.5 Billion Push Notifications
Happy Learning!!!
No comments:
Post a Comment