Vpon has been a trustworthy partner when it comes to boost installation and first transaction rate through mobile ads for the ridesharing mobile app. Vpon constantly optimized the campaign performance through multiple dimensions and digged out customer insights for our client.
This case study has bagged the Top Mobile Awards in Big Data Marketing Category, proving the outstanding capability of Vpon’s big data analytics. Thanks to the artificial intelligence algorithm optimized by the combination of machine learning and expert experience, both the installation rate and first transaction rate have been improved phase by phase. Meanwhile, the cost per installation and cost per first transaction have been lowered by 81% and 63% respectively.
2. Objective
• Boost mobile app installation
• Lower cost per first transaction (CPFT)
Campaign
Period
June, 2015 – June, 2016
Geographic
Targeting
More than 20 cities across China including
first-, and second-tier cities, such as Beijing,
Shanghai, Guangzhou, and Shenzhen
Ad Format Mobile Banner
Operating
System
• iOS
• Android
Results
• CPFT decreases 63%
• CVR reaches 6.23%
3. From download to transaction
Download and install
ridesharing mobile app
Register an account
Make the first-riding
transaction
APP
5. Phase 1 :
Multi-Dimensional Manual Optimization
Location
Optimization
Creative
Optimization
Time
Optimization
Operating System
Optimization
6. • The first-tier cities in China had better conversion rates
than second-tier cities did.
• It is more likely for mobile users to install a
ridesharing mobile app during commute 32%.
7. Ran A/B testing to maximize the usage of the most effective creative to encourage call-to-actions.
CVR:5.33% CVR:6.82% CVR:4.96%
8. The ROI of advertising investments on iOS system was greater than that
on Android, as the former one performed a better conversion rate.
0%
20%
40%
60%
80%
100%
Clicks Installations First transaction
Comparison of Performance by OS
iOS Android
38%
62%
27%
73%
83%
17%
iOS Android
Installation Rate:
Registration Rate:
First-ride Rate:
Installation Rate:
Registration Rate:
First-ride Rate:
7.3%
64.2%
61.8%
4.5%
55.2%
54.4%
10. Audience Optimization
Vpon’s Data Management Platform (DMP)
Demographics
Interest
Location
Device
App
5 dimensions with 1000+ tags
collected from mobile devices
Audience A Audience B Audience C
Profile 1
Analyzed the data and
developed audience
profiling, segmenting and
look-alike audience targeting
Profile 2 Profile 3
Built Target Audience Prediction Model
11. ** Image for reference only
Transportation
News
Travel
Health & Fitness
Lifestyle
Finance
Business
Sports
Utilities
Shopping
Weather
Tools
Productivity
Entertainment
Personalized
Education
Photography
Music
Communication
Social
医疗
The actual number of apps (by
category) that installed on the
ridesharing mobile app users’
devices
The higher the ratio, the
stronger intensity the
ridesharing mobile app users
installed the apps (by
category).
It is more likely for users who have
downloaded transportation, news, and
travel-related apps to download the
ridesharing mobile app. The analytics
shows that these groups of users show a
higher interest.
Analyzed app preference of the ridesharing mobile app users
12. Data scientists developed
an unique algorithm based
on the massive data
collecting during the data
accumulation stage.
The algorithm model
constantly optimized by
machine learning to predict
real-time the individual
mobile user’s possibility to
download, register, take the
first ride, etc.
The algorithm identified
and prioritized the high
potential users
programmatically to bid ad
impressions via RTB (Real-
time Bidding Engine).
Developed a tailor-made algorithm to achieve a better ROI
14. Increase ROI for Ad Investments
• 1st – 8th Month:CPI and CPFT were fluctuated and higher during data accumulation stage.
• 9th – 12th Month:With the continuous performance optimization, the algorithm worked
in full force and resulted in a significant drop for both CPI and CPFT.
4th Month 8th Month 12th Month
Decrease 81% Decrease 63%
4th Month 8th Month 12th Month
Cost per Installation (CPI) Cost per First Transaction (CPFT)
USD USD
15. Overall Campaign Performance
CVR: 6.23%
Number of mobile app installation: 1,030,000
Number of first-ride after installation: 298,000
0.0%
2.0%
4.0%
6.0%
8.0%
10.0%
Phase 1 Phase 2 Phase 3
Comparison of installation rate
through 3 phases
4.4%
6.7%
8.0%
$0
$10
$20
Phase 1 Phase 2 Phase 3
Comparison of cost per first-riding
transaction by 3 phases
USD
Lowered by 63%