This session demonstrates how Amazon Mobile Analytics can help you make the most of your mobile analytics, keeping tabs on key trends such as active users, revenue, retention and behavioral insights. View webinar on demand here: https://www.brighttalk.com/webcast/9019/195163

1. © 2015, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Hosted by AWS Technical Evangelists, Danilo Poccia & Ian Massingham
AWS Mobile Web Day
24th March 2016
Follow AWS on Twitter at @AWScloud & AWS Mobile Services at @awsmobile

2. © 2015, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
Sandeep Atluri, Data Scientist, AWS Mobile
Analyze Mobile App Data and
Build Predictive Applications

3. “If you can’t measure it ,you can’t improve it”
-Lord Kelvin

4. Retrospective
reporting to analyze
trends and to know
what's happening in
the business
Predictive
applications to
anticipate user
behavior and to
enhance experience
Inquisitive
pattern finding to
discover latent user
behavior and to frame
strategies accordingly
Three Types of Data Driven Development

5. How many users use the app and how often?
What are key user behaviors in the app?
Your
Mobile
App
How to predict user behavior and use those
predictions to enhance their experience ?
In the Context of a Mobile App

6. THREE TYPES OF DATA DRIVEN DEVELOPMENT
Retrospective
reporting to analyze
trends and to know
what's happening in
the business
Predictive
applications to
anticipate user
behavior and to
enhance experience
Inquisitive
pattern finding to
discover latent user
behavior and to frame
strategies accordingly

7. Let’s just say you have built a music appMusic App

8. Let’s just say you have built a music app
What are some of the questions that would help you analyze trends in the app?
Music App

9. Engagement
How many users use
the app daily to listen
music ?
How many times
users open the app to
listen music in a day?
How many new users
have been acquired
to the app ?
Monetization
How many paying
users does the app
have ?
How much does a
average paying user
pay ?
Retention
How many people
returned to listen
music in the first 7
days after they
have installed the
app ?
Behavioral
How many users
shared or liked a
particular artist ?
Few Key Questions to Analyze Trends in the App

10. 65
66
66
66
66
66
Millions
Time spent in the app (total session length)
3,000
3,050
3,100
3,150
3,200
3,250
3,300
3,350
Thousands
Number of app opens (Sessions)
Push notification
Marketing campaign
Push notification Marketing campaign
Music
App
Engagement

11. Marketing campaign did successfully improve app opens but did not
result in users spending more time in the app
65
66
66
66
66
66
Millions
Time spent in the app (total session length)
3,000
3,050
3,100
3,150
3,200
3,250
3,300
3,350
Thousands
Number of app opens (Sessions)
Push notification
Marketing campaign
Push notification Marketing campaign
Engagement
Music
App

12. $1,000
$2,000
$3,000
$4,000
$5,000
$6,000
Revenue
Promotion in app store
0
20
40
60
80
100
Number of paying users
Promotion in app store
Monetization
Music
App

13. $1,000
$2,000
$3,000
$4,000
$5,000
$6,000
Revenue
Promotion in app store
0
20
40
60
80
100
Number of paying users
Promotion in the app store increased the over all revenue and more
importantly the number of paying users as well
Promotion in app store
Music
App
Monetization

14. 50%
55%
60%
65%
70%
75%
80%
36%
40%
44%
48%
52%
Introduction of demo video
Week 1 Retention
Introduction of demo video
Day 1 Retention
Music
App
Retention

15. 50%
55%
60%
65%
70%
75%
80%
36%
40%
44%
48%
52%
Introduction of demo video
Week 1 Retention
Introduction of demo video
Day 1 Retention
Changing the first time experience in the app has significantly improved
retention in the app
Music
App
Retention

16. 0
5
10
15
20
25
30
35
Thousands
Number of purchases of a music track
Promoting the track in the gallery
Number of likes/shares for an Artist
0
1
2
3
4
5
6
7
Thousands
Promoting the track in the gallery
Music
App
Behavioral

17. 0
5
10
15
20
25
30
35
Thousands
Number of purchases of a music track
Promoting the track in the gallery
Number of likes/shares for an Artist
0
1
2
3
4
5
6
7
Thousands
Promoting the track in the gallery
Promoting the track has not only increased purchases for the track but
has also increased the number of shares for the artist
Music
App
Behavioral

18. Is there a easy way to track all these
metrics automatically as soon as
users start to use your app ?

19. Amazon Mobile Analytics
“Collect, visualize and export your app usage data at scale”
Scalable and
Generous Free Tier
Scale to billions of
events per day from
millions of users.
Own Your Data
Data collected are not
shared, aggregated, or
reused
Focus on metrics that
matter. Usage reports
available within 60
minutes of receiving data
from an app
Fast

20. Key Business Metrics
1. Monthly Active Users (MAU)
2. Daily Active Users (DAU)
3. New Users,
4. Daily Sessions
5. Sticky Factor
6. 1-Day Retention
7. Avg. Revenue per DAU
8. Daily Paying Users
9. Avg. Paying DAU
Get Metrics Important for Your App in a Single View

21. Get a Detailed View of Each Business Metric

22. Track Unique Behavior to Your App Using
Custom Events

23. Retrospective
reporting to analyze
trends and to know
what's happening in
the business
Predictive
applications to
anticipate user
behavior and to
enhance experience
Inquisitive
pattern finding to
discover latent user
behavior and to frame
strategies accordingly
Three Types of Data Driven Development

24. Going beyond standard metrics will give
you more insight in to user behavior

25. How does usage pattern vary for users with different demographic profiles?
Who are the most engaged users and what are their usage patterns ?
How does user population distribute across countries and platform ?
How much time does it takes for a user to convert to a paying user ?
Music App
Few Questions that Will Help you Understand
your Users Better

26. 23%
8% 7% 6% 6%
4% 3% 3% 2% 2% 2% 2% 2% 1%
29%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15+
Highly engaged users opening
the app more than 15 times
Users with low engagement
Number of times app was opened in last 7 days
Who are the Most Engaged Users and What
are their Usage Patterns?

27. 23%
8% 7% 6% 6%
4% 3% 3% 2% 2% 2% 2% 2% 1%
29%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15+
Highly engaged users opening
the app more than 15 times
Users with low engagement
Number of times app was opened in last 7 days
Design strategies to influence users with low engagement and convert
them to highly engaged users
Who are the Most Engaged Users and What
are their Usage Patterns?

28. 30 20 10 0 10 20 30
16-20
20-25
25-30
30-35
35-40
40-45
45-50
50-55
55-60
60-70
70+
Female Male
Time spent in the app(minutes)
Age
How Does Usage Pattern Vary for Users with
Different Demographic Profiles?

29. Understand your core user demographic profile and deliver relevant
content to them
30 20 10 0 10 20 30
16-20
20-25
25-30
30-35
35-40
40-45
45-50
50-55
55-60
60-70
70+
Female Male
Time spent in the app(minutes)
Age
How Does Usage Pattern Vary for Users with
Different Demographic Profiles?

30. How Does User Population Distribute Across
Countries and Platform?

31. Formulate new user acquisition plans in countries that the app has low
penetration
How Does User Population Distribute Across
Countries and Platform?

32. 6%
8%
10%
12%
13% 13%
12%
10%
8%
5%
4%
1 2 3 4 5 6 7 8 9 10 11+
Number of days before first in app purchase
How Many Days Does it Take for a First Time
User to Convert to a Paying User?

33. 6%
8%
10%
12%
13% 13%
12%
10%
8%
5%
4%
1 2 3 4 5 6 7 8 9 10 11+
Number of days before first in app purchase
Target users who have spent more than 8 days in the app and are yet to
purchase
How Many Days Does it Take for a First Time
User to Convert to a Paying User?

34. Auto Export to Amazon Redshift

35. Simple &
intuitive
Integrate with
existing data
models
Automatically
collect common
attributes
Schema for Your App’s Event Data

36. Now Easy to Query and Visualize
Your
Mobile
App

37. SQL Query
select “app opens”,
count(users) as “frequency”
from (
select
client_cognito_id as “users”
,count(*) as “app opens”
From
AWSMA.v_event
Where
event_type=‘_session.start’
And event_typestamp between
getdate()-7 and getdate()+1
Group by client_cognito_id
)
Group by “app opens”
23%
8%
7%
6% 6%
4%
3% 3%
2% 2% 2% 2% 2%
1%
29%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15+
Number of times app was opened in last 7 days
Who are the Most Engaged Users and What
are their Usage Patterns?

38. SQL Query
select
a_age as “age”
,a_gender as “gender”
,avg(m_session_length) as “time spent”
From
AWSMA.v_event
Where
event_type=‘a_session.duration’
And event_typestamp between
getdate()-90 and getdate()+1
Group by
m_age
,m_gender
30 20 10 0 10 20 30
16-20
20-25
25-30
30-35
35-40
40-45
45-50
50-55
55-60
60-70
70+
Female Male
Age
Time spent in the app(minutes)
How Does Usage Pattern Vary for Users with
Different Demographic Profiles?

39. How Does User Population Distribute Across
Countries and Platform?

40. Retrospective
reporting to analyze
trends and to know
what's happening in
the business
Predictive
applications to
anticipate user
behavior and to
enhance experience
Inquisitive
pattern finding to
discover latent user
behavior and to frame
strategies accordingly
Three Types of Data Driven Development

41. Predicting user behavior will help
you deliver personalized
experience for users

42. Susan has been using the app for more than 6
months now but she hasn’t opened the music app
in the last ten days
Predictive Application by Example
Music
App

43. Susan has been using the app for more than 6
months now but she hasn’t opened the music app
in the last ten days
What would you do to bring her back to the app again ?
Music
App
Predictive Application by Example

44. “Susan, you haven’t listened to your favorite
artists in a while. Want to check them out? ”
Push Notification
Predictive Application by Example

45. “Susan, you haven’t listened to your favorite
artists in a while. Want to check them out? ”
But what’s the best time to send her this push notification ?
Push Notification
Predictive Application by Example

46. SELECT e.time_stamp
FROM events e
WHERE customer =‘SUSAN’
AND event_type = ‘_push_notification_open’
HAVING e.date> GETDATE() - 30
You can start by looking at all
the different time slots she has
opened a push notification in the
last 30 days
One Way To Do is…

47. SELECT e.time_stamp
FROM events e
WHERE customer =‘SUSAN’
AND event_type = ‘_push_notification_open’
AND date_part (dow,e.date ) in (6,7)
HAVING e.date> GETDATE() - 30
But her usage pattern changes
on weekends.
You can edit the query to filter
for weekends only
One Way To Do is…

48. SELECT e.time_stamp
FROM events e
WHERE customer =‘SUSAN’
AND event_type = ‘_push_notification_open’
AND date_part (dow,e.date ) in (6,7)
HAVING e.date> GETDATE() - 60
Pattern is not clear as she
opened in multiple time slots on
different days.
You can go back in time to get a
more clear pattern
One Way To Do is…

49. SELECT e.time_stamp
FROM events e
WHERE customer in (‘SUSAN’,’JOE’,’BOB’,…..)
AND event_type = ‘_push_notification_open’
AND date_part (dow,e.date ) in (6,7)
HAVING e.date> GETDATE() - 60
but what about other users ?
tweak the query again
One Way To Do is…

50. SELECT e.time_stamp
FROM events e
WHERE customer in (‘SUSAN’,’JOE’,’BOB’,…..)
AND event_type = ‘_push_notification_open’
AND date_part (dow,e.date ) in (6,7)
HAVING e.date> GETDATE() - 120
….and again
One Way To Do is…

51. SELECT e.time_stamp
FROM events e
WHERE customer in (‘SUSAN’,’JOE’,’BOB’,…..)
AND event_type = ‘_push_notification_open’
AND date_part (dow,e.date ) in (6,7)
HAVING e.date> GETDATE() - 120
Use machine learning technology to
learn business rules from your data

52. Best time to Send
4 PM
9 AM
2 PM
Machine learning automatically
finds patterns in your data and
uses them to make predictions
Better Way To Do it is…

53. Machine learning automatically
finds patterns in your data and
uses them to make predictions
Your data + machine learning
= personalization in the app
Best time to Send
4 PM
9 AM
2 PM
Better Way To Do it is…

54. Building and scaling machine learning technology is hard
Machine learning expertise is rare
Closing the gap between models and applications is
time-consuming and expensive
Why Aren’t there More Machine Learning
Applications Today?

55. What if there were a better way?

56. Easy to use, managed machine learning service built
for developers
Robust, powerful machine learning technology
based on Amazon’s internal systems
Create models using your data already stored in the
AWS cloud
Deploy models to production in seconds
Amazon Machine Learning

57. +
Amazon Mobile Analytics Amazon Machine LearningAmazon Redshift
Leverage Mobile App Data in Amazon Redshift to
Build Predictive Applications Using Amazon ML

58. Train
model
Evaluate and
optimize
Retrieve
predictions
Building Predictive Applications with Amazon ML
1 2 3

59. Evaluate and
optimize
Retrieve
predictions
Train
model
- Create a Datasource object pointing to your mobile app data
- Explore and understand your data
- Transform data and train your model
Building Predictive Applications with Amazon ML
1 2 3

60. Create a Datasource Object

61. Explore and Understand Your Data

62. Train Your Model

63. Train
model
Evaluate and
optimize
Retrieve
predictions
- Understand model quality
- Adjust model interpretation
1 2 3
Building Predictive Applications with Amazon ML

64. Explore Model Quality

65. Train
model
Evaluate and
optimize
Retrieve
predictions
- Batch predictions
- Real-time predictions
1 2 3
Building Predictive Applications with Amazon ML

66. Amazon Mobile
Analytics
Amazon
Redshift
App events
Data SourceStrategies
Predictions
Mobile app
developer
Amazon Machine
Learning
+
Now Build Predictive Applications Using Your
Mobile App Data Easily
Your
Mobile
App

67. Predict users with low probability to purchase in the app and send discount coupon via in-app notification
Predict users with high probability to churn from the app and send push them notification to re-engage
Identify users with high probability to share the app and reach out to them to do the same
Recommend relevant content to users based on similar user’s behavioral patterns
Few Strategies that can be Used Effectively
via Machine Learning

68. Thank you !
For further questions please email us at
amazon-mobile-analytics@amazon.com