The challenge this session’s speaker and his colleagues faced in trying to learn more about customer experiences was that insights are fragmented across different systems such as Oracle Eloqua, CRM, and web analytics. To better understand their contacts, they started with the corporate data warehouse, which was missing a lot of this lower-value and detailed data. When they considered expanding the data warehouse, it was difficult to define what questions they wanted to answer in advance, because it varies for each campaign they run. Thus they embarked on building a Hadoop-based data lake, for the flexibility to ask any questions with an ad hoc schema on read approach, against any customer data sets in varying levels of detail, to better understand what their visitors want to consume.
Breakout Session
Wednesday, Apr 26, 5:45 p.m. | Mandalay Bay D
https://oracle.rainfocus.com/scripts/catalog/oracleCx17.jsp?search=BRK1098
1. Journey to
Marketing Data Lake
BRK1098
Oracle Modern Marketing
Experience | Las Vegas | Apr 26-28
Sumit Sarkar
Product Marketing
@SAsInSumit
linkedin.com/in/meetsumit
The challenge this session’s speaker and his colleagues faced in trying to learn more about customer experiences was that insights are fragmented across different systems such as Oracle Eloqua, CRM, and web analytics. To better understand their contacts, they started with the corporate data warehouse, which was missing a lot of this lower-value and detailed data. When they considered expanding the data warehouse, it was difficult to define what questions they wanted to answer in advance, because it varies for each campaign they run. Thus they embarked on building a Hadoop-based data lake, for the flexibility to ask any questions with an ad hoc schema on read approach, against any customer data sets in varying levels of detail, to better understand what their visitors want to consume.
Breakout Session
Wednesday, Apr 26, 5:45 p.m. | Mandalay Bay D
Pillar: Marketing
Marketing Track: Data-Driven Marketing
Product: Oracle Data Management Platform (Oracle BlueKai)
Level: Intermediate
Session Type: Breakout Session
https://go.oracle.com/moderncx-speaker-information
https://oracle.rainfocus.com/scripts/catalog/oracleCx17.jsp?search=BRK1098
Data Lakes are loaded with raw data (no “T”) and create the “Schema on Read” on business demand
To really get big data value, you need to store all types of structured and semi-structured data in a data lake, from CRM data, to social media posts.
You don’t have to have all the answers upfront, or even the questions. Lakes store raw data that can be transformed as questions arise.
Use a variety of tools based on what you’re asking.
Everyone talks about a single, unified view of data
http://info.zaloni.com/hubfs/Architecting_Data_Lakes_Zaloni.pdf
By Ben Sharma and Alice LaPlante
The project is still run by marketing engineers, so we don’t get the perks of an IT driven project. We expect to continue expanding use and value to then elevate this to production and start looking to do more buying than building.
Statistical analysis of detailed data
Log data to study application characteristics
Performance lab data
The project is still run by marketing engineers, so we don’t get the perks of an IT driven project. We expect to continue expanding use and value to then elevate this to production and start looking to do more buying than building.
Statistical analysis of detailed data
Log data to study application characteristics
Performance lab data