3. The perfect storm called ubiquitous computing
“ Computing is everywhere
and anywhere. ”
4. Clarifying the ambient intelligence paradigm
a pervasive computing principle that is sensitive and responsive
Artificial Intelligence Ÿ
Affective Computing
Embedded &
Networked
Internet of Things
Context aware
Sensor fusion
Personalized
Behavioral profiling
Adaptive
A.I. & learning
Anticipatory
Predictive analysis
Privacy
Trust & Privacy
UX
Human Centric UX
5. Putting some numbers in perspective
Wearables
30B units by 2018
Up to 10 sensors
Connected Devices
US$ 290B by 2018
Up to 6 sensors
Smart-phones
2,8B units by 2016
Up to 12 sensors
6. Context-awareness explained
what I tell my grandma what I do
“ We help your phone know you so
well, it helps you to interface with the
world in a much more natural,
proactive, and personalized
manner, so it behaves like a sixth
sense ”
7. Engage in a context-aware manner
Engaging at the right m o m e n t , under the right c i r c u m s t a n c e s .
Time
Feeling
…
Activities
Habit
Mood
Weather
Road type
…
Behavior
8. How context upgrades your content
Know what you are
listening to
Understand the!
circumstances you
are listening to
9. “ What’s up with mobile advertising? Seriously. ”
10. 2014. And my toaster situation.
• Search or buy activity trigger ad-serving.
• Conversion moments on mobile are a
mere few seconds.
• Retargeting hell.
11. Re Pre-targeting is fundamental to mobile conversion
Re-targeting!Pre-targeting!
Bid on users that have left the session
High conversion at low cost
Can’t get the real-time aspect right
Trying to convert users after the facts
doesn’t work on mobile
Big data spray and pray approach
Experimental
Bid on users before they enter a session
Heavily rely on behavioral algorithmic
processing of contextual data, to better
estimate a conversion moment
Emerging out of recommender systems
and contextual computing
Small data precision targeting
Accepted
12. Traditional versus full context-awareness
Online advertising: based on explicit
signals, expressed by humans.
Mobile advertising: based on implicit
signals, learned by machines.
CLICKS
SENSORS
Richer, real-time and accurate
personas
Clicks, page views, and
audience segments
13. Some fundamental challenges
Engaging for conversion on a mobile device?
Using stale data to understand fleeting moments?
Using explicitly entered data from users?
14. Some groundbreaking opportunities
Internet of Things Internet of Channels
Sensors. Sensors. Sensors and behavioral profiling
The offline channel has become connected
Context-awareness leaps beyond location
15. Mobile users live in the now, man!
Mobile terrain
Past
PresentFuture
A smart-phone is an impulse device that captures the present
where real-time and context-awareness are essential
16. How mobile users want to be served
Sensed & Learned:!
- Just woke up!
- Walking!
- Sunny weather forecast!
- Had a good night’s rest !
Sensed & Learned:!
- At the gym!
- On the treadmill!
- Busy day a head!
- Music playing + 100BPM!
Sensed & Learned:!
- In traffic jam!
- Daily commute home!
- Had a short night of sleep!
- End of a busy day!
- Frown eyes, unhappy face!
- Negative text message detected!
- Aggressive typing text message!
- Stress in her voice!
17. “ Argus Labs has created a pre-targeting platform
capable of near real-time contextualization
and profiling using deep learning ”
Context w Mood w Behavior
18. Sensors as building blocks for attribution data
Sensors are a device’s equivalent to possessing human-like sensing
capabilities and allow to sense and understand contextual cues.
Automatic gender detection, through walking patterns
Mood detection, through mobile handling
Sound source detection (gyroscope sampling)
Activity detection (walking, running, sitting, …)
…
Automatic in- versus outdoor detection
Mood detection, through heart rate and oxygen level detection
…
@arguslabs
19. It’s a service layer
Access through SDK and API
Handling > 1 million requests per second
Ambient sensing technology (no input needed)
Android, IOS*, Microsoft Mobile** & Tizen**
Built for mobile and embedded devices
Proprietary IP and patented
Provide real-time push/profile updates to DMP
*
Being
ported/tested
**
Por0ng
started
20. Personas & Moments
Understand, profile and predict habits
Uncover habits, profile and predict human behavior
Sense and understand context
Sense and interpret the contextual cues of a mobile user
Understand personalities and emotions
Learn personalities and emotional states of a mobile user
LAYER 3
LAYER 2
DEPRESSION SOCCER MOM
SUSCEPTIBILITY SCORE
LAYER 1
BIKING SLEEPING TIMES
RUNNING
LOUD ACTIVE
MUSIC
LOCATION
HAPPY
LAZY SAD
EXTROVERT NERVOUS …
ENGAGED …
IN MEETING IN TRAFFIC STRESSED
INDOOR …
21. Privacy as a trading function
Gyroscope
Accelerometer
Microphone
Camera
GPS
Light
In-App Usage
Expected
value
User
acceptance
Wi-Fi
Browsing History
Calendar
SMS
GSR
The more permissions are required, more added value is expected from the mobile app!
The type and quality (“value”) of a mobile app, not quantity, determines the ad estate value!
22. Start to build mood- and contextually
aware real-time mobile campaigns
23. Context-aware mobile
advertising: turning ads into!
real-time engagement
Filip Maertens
Founder
http://www.arguslabs.com
filip.maertens@arguslabs.com
@fmaertens