How to select and integrate the data all these billions from, which would fill the most reliable user profile? What interests the user has, what age, where he is coming from? What pages does he visit? Big Data tools and AI algorithms come with help. They continuously monitor Internet traffic and learn people's behaviour to create precise profiles of Internet users.
Big Data Warsaw v 4 I "Precise Data Integration Thanks to Big Data Analysis & AI Algorithms" - Wojciech Dwojacki
1. Precise data integration
thanks to Big Data analysis
and Artificial Intelligence
algorithms
Wojciech Dwojacki, OnAudience.com / Cloud Technologies Group
5. By 2021, Internet traffic will be
generating 280,000 petabytes
monthly
Internet traffic (monthly, in petabytes)
Global data volume is growing rapidly
Source: Statista.com
6. Value of global data market in 2019 Value of data market in Poland in 2019
became a currency of XXI century
Source: Global Data Market Size 2016-2018, OnAudience.com
Data
9. How domains are classified?
New website
Classified domains
Classification model PredictionsText transformation
Interests
Demography
Intentions
10. using DMP technology
Data Management Platform (DMP)
The platform makes it possible to control all
users’ data that comes from various sources.
It enables to manage and analyze data about
online activity and users’ behavior.
The DMP helps to profile users and group them
into larger segments, such as Music Lovers or
Sports Fans.
Integrating and managing data
12. DMP makes it possible to collect
data from various channels, for
example:
● web pages
● landing pages
● mobile applications
● offline, e.g., data from CRM
● online campaigns
Data Gathering
13. DMP allows the integration of all data types in
purpose to create a 360-degrees client picture.
1st party data - client’s own data,
e.g., from web pages or CRM
2nd party data - partner’s data,
e.g., from online campaign
3rd party data - purchased from data
providers, e.g., DMP/DSP
Integration
14. DMP makes it possible not only to gather
but also organize all data types, e.g.:
● creating segments tailored to specific
brands
● division of user profiles based on
markets
● creating custom segments, e.g.,
users who bought specific product,
visited chosen website, or
participated in selected campaigns.
Management
15. Activation
DMP gives access to DSP platforms, where
online campaigns are maintaining on. After
segments are created, online campaign is set
in a few clicks, targeted at precisely chosen
group. For example, via:
● display ads
● video ads
● social network ads
● search ads
16. Gathered data is precisely analyzed. The
information serves for optimization of the
campaign and increasing client engagement.
Analysis takes into account:
● demographic data
● online activity
● online campaign efficiency
● best converting profiles
Analysis
18. with using of machine learning algorithms
Profiling of online clients
For creating a user profile, machines “read” and
analyze the content of visiting pages, and identify its
subject.
Based on the analysis, machines add an attribute
into the user profile which indicates a category,
e.g., automotive, art or technology interests.
19. Standard method
Methods of look-alike modeling
● Analysis of data points in target group
● Building groups with the highest probability
level of conversion
● Focused od quantity (large group of profiles)
20. ● Based on customer path
● Focused on quality (smaller
target groups)
Graph method
Methods of look-alike modeling
23. One of the biggest data warehouses worldwide.
Processes more than 27 B user profiles from over 200
global markets worldwide.
A part of Cloud Technologies group - one of the fastest
growing technology companies in Europe, according to
rankings of Financial Times and Deloitte.
Who we are | OnAudience.com