Google Quebec hosted Think Quebec and this year they explored digital marketing as a path to quicker, deeper connections between a brand and its consumer. Inspired by Parkour, the popular urban sport of finding the most direct route to your goal, they presented campaigns and strategies as beautiful as they are successful–another discipline at the junction of art and science. James Prudhomme, CEO Datacratic spoke at Google's Think Quebec. His Talk is entitled "Why is programmatic taking off? What is this revolution all about?"
4. Datacratic développe des technologies d'apprentissage machine ("machine
learning") et d'intelligence artificiel permettant l'automatisation de prise de
décision en temp réel pouvant être déployé dans un large éventailde
d'applications.
Datacratic develops machine learning and artificial intelligence
technology, which enables real-time machine based decisioning to
be deployed into a wide range of applications.
5. What is Machine Learning?
• Machine learning, a branch of artificial intelligence, concerns the
construction and study of systems that can learn from data.
• Another way to think about machine learning is that it is the act
of teaching a software program to recognize patterns.
• After learning the pattern, the program can make predictions on those patterns.
• Those predictions can be integrated into products and applications. .
6. Marketer Challenges and Pain Points
• Media Optimization:
• Publishers are unable to serve impressions to users with a high probability of
clicking.
• Conversion Optimization:
• Marketers are unable to target users who have a high probability of converting.
• Segment Reach Extension:
• Can’t scale from a precision segment to a larger qualified audience.
• Precision Targeting:
• Inability to target a high performing subset of an audience with any precision.
8. Audience Modeling Use Cases
Media Optimization:
• Identifies users with a high probability of performing targeted campaign behaviors
Conversion Optimization:
• Uses machine learning models to identify users with a high probability of converting
Segment Reach Extension:
• Identifies a larger pool of users who behave similarly to an underlying segment.
Precision Targeting:
• Focuses on a high scoring subset of a segment target highest quality converters.
9. 9
Lotame Optimizer Case Study – 172% Increased ROAS
Cost per acquisition for the Lotame-powered
campaigns were reduced by
64%.
Users targeted as part of the Lotame
Optimizer campaign demonstrated an
372% increase in conversion rate.
Users targeted as part of the Lotame
Optimizer campaign were 129% more
likely to click on an advertisement.
10. 5 Ways to Use Audience Data to Drive Performance
• Use media optimization to reach your target audience.
• Use 1st party attributes to create a unique audience-segment.
• Use 3rd party attributes to scale to a larger audience.
• Combine 1st & 3rd party for uniqueness & scale.
Use machine learning & modeling to
predict & target your best prospects.
11. Train the model
using every
attribute of the bid
request.
Score bid
requests based
on the campaign
objectives.
Determine
the value of the
bid request and
manage budget.
Calculate the
optimal bid price
and send bid
decision to bidder.
Using Data in RTB Optimization
12. Using Data and Machine Learning to Drive RTB Performance
“Datacratic has enabled AdGear
to provide unprecedented value
to customers using the AdGear Trader
RTB Platform. After integrating RTB
Optimizer from Datacratic, our clients
experienced significant CPA cost
reduction.”
Vlad Stesin, CEO Adgear
Over 65% Reduction in CPA
• CPA Optimization - Manufacturing
• Duration - 2 Weeks
• Averaging 15-71 conversions per day
• Starting CPA level of $8.6 reduced to $3.00
60% Reduction in CPM
• CPM Optimization – Food & Beverages
• Duration - 1.5 Months
• Averaging 200,000 impressions daily
• CPM reduced from $0.8 to $0.32
50% Reduction in CPA
• CPA Optimization – Retail
• Duration - 1.5 Months during holiday season
• Averaging 150 – 200 conversions daily
• 34.7 million impressions, 23.5k clicks,
total spend $48.5k