1. Big Data and Marketing:
What’s in it for me?
Davy Cielen
2. WHAT IS BIG DATA?
A vast quantity of UNSTRUCTURED data, which we
now have the ability to process in REAL TIME.
3. WHY IS IT BECOMING IMPORTANT NOW?
Rise of smartphones
Rise of sensor networks
Social network adaptation
4. HOW BIG IS BIG DATA?
Every 2 days we create as
much information as we did
from the beginning of time
up to 2003.
By 2020 analysts predict the
amount of data will be 50x what it
is today.
5. WHAT MAKES BIG DATA SO DIFFERENT?
Veracity
Velocity Variety
Volume
8. AN INTRODUCTION TO BIG DATA TECHNOLOGY
Map reduce
framework
Specialized databases
& tools
Tools to perform analysis
in real time or on a huge
scale
Hadoop:
A distributed file
system
Everybody is talking about Hadoop , Hbase and other technologies. In fact I can spent an hour with nothing than name dropping
but the most important thing in big data, is data then privacy issues and not technology.
The availability of new data is a much bigger driver for the big data revolution than is any other break through.
The fact that we are able to collect and combine all these types of information are far more important than any single technology.
Business models and new products filled with sensors are far more important than any technology.
It is up to you as a business to build a model in which you give a customer so much value for his data, that he doesn’t mind at all that
You can see his most intimate details as long as he is served by you. At that moment, privacy is much less of an issue than you think.
Big data really means big data. At least 40% of the internet traffic today comes from video. If today
Introduction
Right message, right person
What is it?
One single truth of the customer, its DNA. Detailed insights of who the customer really is.
Combining internal and external datasources.
Benefit?
Better customer selection
Relevant message to customer
Price based or should it emphasize the security, social or coolness factor of your product?
Why now?
Integration of different data sources with very different structures
Examples?
NGDATA = customer integration
Introduction
Right message, right person, right time = event driven marketing = trigger marketing
Why now?
The need for real time analysis + the sheer volume + the unstructured nature of the data
Integration of multiple data sources
Types of triggers
Transactional triggers: based on an action
Recurring triggers: :based on an individual profile and details
Behavioural triggers: based on an activity
Treshold triggers
Most marketers only react to big events such as new year, holidays, ….
Examples
Using social data & by integration of data sources we can discover more fine grained events , examples:
End of contract or heritage
Customer visited a certain page on your websitej
Big life events
Customer did a simulation
Soomething that happened on a partner website
Be ready to analyze the results to see what content works best
Introduction
Right message, right person, right time & actionable place
http://www.slideshare.net/havas-media/locationbased-marketing-lbm-global-media-trends?qid=c8e9ce50-2e57-48ff-a9f5-ec84cacc8ea3&v=default&b=&from_search=3
Why now?
Enabled by gps, rfid, wifi, checkin on social media:
The amount of sensor data + the need to act in real time
Geospatial data is best handled by geospatial databases
Examples
Different adds based upon weather profile in a region
Find friends in the neigbourhood
Informing instead of marketing
Coca cola allows you to make a personalized drink when you approach a vending machine
Coca cola (Gefen Team) shows your name when you approach a billboard
Almost 20% of the people are already using LBM
Expected to bring $10 billion by 2016
Out of home
Events
Search and display
Point of sale
Mobile apps
Location based consumer usage throughout the world:
Navigation 46%
Find restaurant and entertainment venues 26%
Find friends 22%
Public transport schedule 19%
Deals and offers 12.5%
Klant die op een luchthaven nog een aanmelding krijgt voor een reisverzekering
Klant die de vakantiebeurs of autobeurs bezoekt
Gamification van marketing actie
Personalisation of car insurance based on telematics devices
Use of workout data to adjust rate
Discount targeting: give discounts to a store nearby
Introduction
Do you know that when a friend divorces, that your likelihood on a diverce also rises? In fact, even when a friend of a friend divorces, you also have a higher chance of getting diverced.
What?
Network based marketing exploits information about your friends and relations to give you a better targeted message.
Why now?
Network data is streaming data, the speed of this data just to fast for traditional techniques
Analyzing networks requires special types of database structures
Examples?
- Simple count models have a good predictive power for market share but also capture share of attention
- Events in the first and second degree of separation are good predictors for events such as babies, mariages and divorces
- Events in the first and second degree of separation can build awareness, an accident with a friend will increase your awereness for the need of insurances.
Pitfalls?
Privacy
Introduction
What?
Why?
There is a tremendous amount of information in non structured data such as video, images, audio
Examples?
Lots of customers don’t use a loyalty card for smaller transactions: Recognize a customer based on a video to get a completer view and give targeted marketing
How about recognizing missing children from security cams?
Recognize tumor from images
Use video cameras to detect interest in a product. You know what people buy, but you don’t know which products get the attention of customer
Detect searching behavior from customers => optimize for store layout
Detect objects and give relevant information about this object
Analysing speech data from a call center data for names of competition of upcoming events, products of interest
How about recognizing friends and social relations from pictures (either in the public domain or based on competitions and or marketing actions)
Summarize previous + introduce next topic
Introduction?
Do you know that facebook, google and Amozon all have bank licences?
What?
When you have an installed customer database and you have lots of data about a customer, you are ready to enter new markets and defeat incumbents that lack the data.
Why?
Data is not available in every sector, so it may be an unfair battle. Take wholesalers for instance, it is very difficult to have customer data
Examples?
Google has done this multiple times: they started as a search engine, went into advertising, mobile phones, operating systems and google walled ….
Amazon started as a online bookstore, but now also delivers computing services
Apple started as a computer manufacturer, but now sells watches, mobile phones, tablets, …
Introduction:
Imagine that you dig a hole in your garden and find the oil pouring out of this hole.
What is it?
People realize that data is as valuable as oil but not everybody has access to customer data.
While there have been data collectors and integrators for a long time, also your data may be valuable too for another company.
For instance:
Who would not want to have a peek inside the customer data from Colruyt? Their data is a test lab where you can try products and see the effects.
If you have detailed information of X,Y coordinates for Belgium, this too can be a data product.
Twitter gives you the ability to buy tweets so you can monitor their data while integrators like datasift help you to get normalized data from multiple social media sites.
It is not only interesting to give data on an individual level, even selling aggregated data can be very lucrative certainly when you have build models
on top of that are beyond the reach of your customers, this is something that we see more often in Europe.
Telecom operators can sell you data about traffic
How about bypassing the middle man?
What?
The amount of data many companies have give the customer a superior customer experience and allow you to either bypass the middleman
or give other companies to work with your customers based upon your data.
Build new services on top of this data
Why now?
Expertise,
access to data that other companies want but that you can’t share
Examples
Apple sits on top of an amount of customer data and allows other companies to access its customers
Google uses your search history to allow other companies to have better targeted adds
Introduction:
What?
Why now?
Examples?
Introduction:
As a teenager a dreamt about having a X-ray glasses and I could see all the beautiful girls naked. Well it might not be so far yet
and probably due to some privacy issues never will be, we now are able to add a lot of information to the things we see and do.
What?
Even the simplest camaras now can recognize humans and focus on their faces. We are able to process more types as data
such as videos and speech and than recognize people from it. Although not perfect but combined with prior knowledge about you
Facebook is able to make pretty good recognitions on your pictures.
Not only video and pictures can be used but also information from your GPS and or hearth rate monitor.
Why now?
Imagine the amount of data that a multiplayer game sends out?
The availability of sensors and the power of integrating different data sources in real time,
gives us now new possibilities into augmenting the reality and use for instance gamification
Examples?
Now combine this with other information sources, and you are able to give customers a new experience such as gamifying the running experience.
Race yourself is an application on Google glass that transforms your running experience into a game.
Introduction:
Do you realize how smart products have become?
What:
Products are now smarter then ever. A cell phone now combines functions from many separated devices such as cameras, computers and GPS.
But this is not only limited to cell phones. Think about cars, glasses, heating systems, refrigerators, …. Almost everything has become smarter.
Why now:
Sensors are cheaper then ever and every product is connected to the internet. There are millions of RFID devices around us that enable us to track devices all around.
Combine this with an unprecedented data processing capabilities and smarter products are born.
Examples:
Google now allows you to find where you have parked your car and will alert you that you will face heavy traffic on the way to your next appointment appointement.
Smart meters like google nest help you to optimize your energy expenditure.
New cars can parc themselves.
New experiences:
While inventing a driving car themselves is not something every company can do, it is certainly a new experience that is build on top of
A combination of engineering and data processing tools.
To build such a car you need to be able to analyse so much information in such a low latency that you need extreme analytical and data processing power.
Why now?
Big data is not only about processing vast amounts of data. In fact the quantity can be much smaller that you would expect but also about being able to
Perform calculations on a small device such as a watch or even much smaller.
Examples?
Google car
Google glasses
Microsoft Wii