big data & personalized content in your cms
TRANSCRIPT
Big data & Personalized content in your CMS
Peter Lem, MSc., VP Engineering O2MC I/O
7th October 2016
© 2015 - Insite Innovations and Properties B.V. 1
How I got here
© 2015 - Insite Innovations and Properties B.V. 2
Jeroen Wijdeven De Ideeënfabriek
Big data
Facebook, the largest online social-network, collected 300 petabytes of personal data since its inception – a hundred times the amount the Library of Congress has collected in over 200 years
© 2015 - Insite Innovations and Properties B.V. 3
Big data
• Volume - The immense scale of data coming towards us.
• Variety - All the different forms of data, from website to sensor data.
• Velocity - The speed at which data are generated.
• Veracity - The uncertainty of how accurate data are.
© 2015 - Insite Innovations and Properties B.V. 4
Big data
© 2015 - Insite Innovations and Properties B.V. 5
The world of connected devices, appliances, machines and what not is growing at the speed of light.
Disruptive technologies - Netflix
© 2015 - Insite Innovations and Properties B.V. 6
Big data companies
© 2015 - Insite Innovations and Properties B.V. 7
The disruptive nature of these technology companies is caused by
• On demand, user friendly user interface• Personalized value• Connected to / owning data nobody else is
Big data companies
© 2015 - Insite Innovations and Properties B.V. 8
Startups hypes
• Uber: valued at 17 billion, without making any profit• Facebook: 8 billion stock value, 1 bilion revenue,
190 bilion market value
Power of big companies
© 2015 - Insite Innovations and Properties B.V. 9
1966• US vs Russia• Costs: 25 billion
The new world: power of big companies
© 2015 - Insite Innovations and Properties B.V. 10
2016• Tusk (Telsa) vs Bezos (Amazon)• Valued at: 12+12 billion
The new world: power of big companies
© 2015 - Insite Innovations and Properties B.V. 11
Unpurposed data
Behavioral data
Primarydata
• Data can be reused without losing value• Different types of data with different
values and volumes
Big data
Machine learning
– Algorithms and not business rules determine what is best based on data
• Provide data and the algorithm will figure out what to do
– Predicting what is going to happen
• Autonomous cars
© 2015 - Insite Innovations and Properties B.V. 12
Big data
• Risks
– Connecting all the data and processing it requires massive amounts of computing power
• And flexibility of each system
– Data governance is a challenge
• Are we allowed to combine all data sources to gather information about an individual?
• E.g. Facebook friends suggestions based on Whatsapp contacts
© 2015 - Insite Innovations and Properties B.V. 13
Big data
• Ethics
– How to handle predictions that go beyond human capacity
• E.g. Minority Report
© 2015 - Insite Innovations and Properties B.V. 14
Big data
• Consider data as an asset, not as costs
• Transparency is key, user should always be in control
• Social science important to deliver personalized value
© 2015 - Insite Innovations and Properties B.V. 15
Simplify Data Life Cycle Management
© 2015 - Insite Innovations and Properties B.V. 16
Traditional data processinginefficient and costly data management process
Our solutioningest any data source for downstream
business implementationssimple, sustainable and secure
Philips HUE: Creating more engagement through better recommendations
© 2015 - Insite Innovations and Properties B.V. 17
Philips HUE are Wi-Fi-powered
light bulbs that can take any color
you like to create exactly the
right mood. The longer HUE lights
are used, the more they become
a part of people’s lives. To further
extend this bond, Philips
approached O2MC I/O to develop
a new app that would create
spot-on recommendations for
personal light scenes
Can we apply this online?
© 2016 - Insite Innovations and Properties B.V. (an Adversitement Group company). 19
Traditional personalisation
Persona matching plugin
Umbraco
3rd Party
1) request www.website.com
2) return HTML
4) return content
3) execute HTML and get content
Our shared principles
© 2015 - Insite Innovations and Properties B.V. 20
• Personalization and A/B testing• Integrated solution in stead of content replacement
• Single point of design• Optimal performance
• Real time• Multiple touch points
• Not just 1 way of measuring success, entire customer journey is important
© 2016 - Insite Innovations and Properties B.V. (an Adversitement Group company). 21
Architecture – personalization with O2MC I/O
send c
onte
nt
with personalised content
Persona matching plugin
Umbraco
1) request www.website.com
3) return HTML
5) return content
4) execute HTML and get content
2)
Ask for
conte
nt
7) collect other data
Getting personalised content
© 2015 - Insite Innovations and Properties B.V. 22
• Umbraco requests the best content for visitor given set of content items to consider• Real time content status taken into consideration• No fixed setup of tests• Multiple tests per website
• O2MC Platform provides content item (id) back to Umbraco• Content management and design in Umbraco• Include other data points as well such as weather• Use business rules or machine learning
Measuring content item effect
© 2015 - Insite Innovations and Properties B.V. 23
• Each page contains 1 line of Javascript code to each page as well as the visitor id
• O2MC Platform can use the SLoC to capture all interactions from this point forward• Not just direct ´click´ on content but also engagement
scoring in orders• Change data model without changing content• Real time extension with data
Environment
© 2015 - Insite Innovations and Properties B.V. 24
• O2MC Platform can run on any platform, installed on Azure, AWS and several others
• DocDB, MongoDB or MySQL for storing the data• Configuration in files, configured in studio environment
O2MC Platform
© 2015 - Insite Innovations and Properties B.V. 25
Next steps
© 2015 - Insite Innovations and Properties B.V. 26
• Proven integration, looking towards scalable extensions
• Integrated reporting (in Umbraco)• Configuration through Umbraco• (Deployment) packages
Your opportunity
© 2015 - Insite Innovations and Properties B.V. 27
Unique websites (e.o.) with real time, personalized content using detailed, omnichannel data
• You can bring customer specific data together in one single environment
• Learnings across several of your customers and personalization algorithms