Ironhack Data Analytics Bootcamp - Final Project - Real problem Statement and dataset from OutPush.
- Introduction to outpush
- Problem statement
- Conclussions
- Presentation
- Push-notifications ads network
- Launched in december 2020
- Middleman between advertisers and publishers
How does it work?
- Clients install a module
- Outpush send personalised ads to the client users
- 1 click = money for the client
Subscribers receive same notifications multiple times: is it usefull?
Challenges of this project:
- Understand how do repeated impressions affect users ‘Click Through Rate’ over time
- Perform statistical analysis
*‘Click Through Rate’ = Number of clicks on Ads / Number of impressions of the Ads
Boost the notification push during the time periods with higher subscribers activity:
- Wednesdays, the higher 'Click Through Rate' of the week
- Saturdays, the lower 'Click Through Rate' of the week
- Between 09:00 AM and 11:00 AM the higher 'Click Through Rate' of the day
Subscribers lose interest on the ads embeded in the noticitions over the time:
-
A subscriber “interest retention” plan could help to boost the Ads 'Opened Rate' for longer time subscribed users.
The optimal number of Ads repetition should be in average 4:
-
CTR decreases after every repetition of the Ads.
- 1st impression: 17.9%
- 4th repetition: 7.1%
-
The probability of opening the Ads decrease after every repetition
- Probability on the 1st impression: 0.8%
- Probability on the 4th repetition: 0.3%
-
The conditional probability of clicking the add, given it was not opened before, drops from 0.79% in the first impression to 0.38% in the 4th repetition.
To see the presentation, click in the below picture.