11. We import a lot of Atlas Data
24 servers
Cloud Storage
Upload 200 + GB
of data per day
( ½ Trillion ICA records )
12. We filter out the relevant cookies
Cloud Storage Elastic Mapreduce
100 Machine Cluster Created on demand. We filter for only the
transactions that we need to process (more than 3.5 billion)
( about 71 million unique cookies a day)
13. Filter by behavior
Filtered Transactions
SKU Table
Generate list of products that have been seen
( Match these cookies to 100,000’s of skus )
14. Match to their affinity
Join transactions
to site genre
information Sport
Enthusiast
70 million
Filtered Transactions
placements
Determinee profile information by the
types of sites the user has visited
( Cookies are matched to 3.5 billion ICA records )
15. …and run custom business rules
Join site
behavior to SKU Table
product info In market
Gamer
Filtered Transactions
Determine the types of products the
user is interested from what they have done on the site
( super–computing power determines some key categories )
16. We bring it all together
category affinity generation
In market
Gamer + Sport Enthusiast
+ Purchaser Home
Theater
( 1 of N “Personalization” segments )
17. Drive a personalized message
User recently purchased
a home theater system
and is now looking for Target Ad
sports games
( 1.7 million per day )
18. Each and every day
This all happens in about 8 hours every day
( not bad )