7. B3
@jongreenhalgh
Good news (or bad). Our jobs will be different in five years’ time
Jobs capable of being
automated
47%
New jobs created
every decade
16%
19. B3
@jongreenhalgh
In digital media, we should also express caution
Reason 2
Algorithms
can be
biased
Reason 1
The ad tech
in use is still
limited
Third-party tools – Do they deliver improvement?
AI bidding models – What’s behind them?
Agency tech – What’s the value?
CPA bidding – Purpose?
Look-a-like modelling – Brand impact?
Dynamic Search Ads – Control?
26. B3
@jongreenhalgh
We do this with three core skill sets:
Data and
technology
60+ coders,
developers, machine
learning experts
Specialists
60+ search
specialists who
challenge
Creativity
Psychological and
behavioral experts
Ai has been around since 1950. It is integrated into our daily lives whether we know it or not and it's here to stay.
The purpose of artificial intelligence is fundamentally based on 3 things; solving problems, learning and most crucially, adapting. A machine has to be conscious of its environment.
The purpose of artificial intelligence is fundamentally based on 3 things; solving problems, learning and most crucially, adapting. A machine has to be conscious of its environment.
Ai is changing the world around us:
47% of all current jobs are estimated to be capable of being automated.
We currently create 16% of jobs every decade.
To stay ahead and protect ourselves, we should become experts in identifying the proper use of AI on our own industries and being part of the conversation.
The skill is being able to evaluate the big picture and provide a path to achieve and add value from AI.
Google Marketing Keynote - fundamental part of the business for the past year and will continue to be for the foreseeable future:
Critical technology helps marketers analyse countless real time signals. As humans, this level of personalisation and automation is impossible. It's also impossible based on pure logic coding - Google learns about you and delivers you results that matter.
In one motion, you can search and Google will analyse multiple signals, including information from the physical world, such as location.
Google is heavily invested in focusing on better assisting experiences. The hope here, no doubt, is that the better they make users experiences, the more advertising revenue opportunities there are across multiple touchpoints, not just search.
The Google focus is on:
Customer Intent
Data
Machine Learning.
All 3 of these things support them in better assisting experiences (encouraging ad engagement and revenue opportunities).
Example of different outcomes to a search, for different individuals.
Intent, context and identity.
Going to the cinema:
User 1: cinemas near me - Intent: It's an option to visit the cinema at the weekend. Context: has friends visiting. Identity: Hasn't bought with Vue before.
User 2: cinema tickets - Intent: considering a cinema trip this week. Context: Not sure what film the want to see, needs some inspiration. Identity: Has bought tickets with Vue before to action movies.
User 3: baywatch cinema tickets tonight - Intent: will be going to the cinema tonight. Context: We know they want to watch Baywatch. Identity: Hasn't bought with Vue before.
Each of these people are interested in going to the cinema, but an improved understanding of intent, context and identity can change ads displayed for the better and improve response time.
To do this successfully, we need to understand all of the data points available to us, which has to be done through machine learning and we have to utilise technology like Google's to support us.
We also need clear strategies to handle our data and have a clear picture of who our customers are, what they need from us and what they are likely to want in the future. Customer segmentation is a good start point, but we need to use technology to support the real time flow of information - we can't model this information retrospectively - we need real time solutions to deliver the right combination to customers.
Google wants brands to centralise their data with them to be able to deliver the best experiences for customers and keep people in their revenue loop. They need data and AI to stay competitive.
IBM Watson is a different solution for brands, but one that is flexible to business needs. The idea of IBM Watson is to provide an algorithm, flexible to all uses.
IBM describe Watson as 'cognitive computing'. It's a really key point - the computer isn't based on logic anymore as it has been since the dawn of computing in our era. It is no longer based on 100's of human written logic and variations.
Watson is designed to observe, interpret, evaluate right / wrong and ultimately decide on the best course of action.
Watson has been developed to understand grammar, context, culture and intent.
Watson needs and relies on human interaction.
Any use case of Watson needs human teaching and consistent intervention. The key point here is that as humans we are investing valuable time in teaching an algorithm that learns from what you tell it to compute thousands of iterations that we could never touch as humans.
Consumer Tool
Amazon Machine Learning is based on the same proven, highly scalable, ML technology used for years by Amazon’s internal data scientist community.
The service uses powerful algorithms to create ML models by finding patterns in your existing data.
Then, Amazon Machine Learning uses these models to process new data and generate predictions for your application.
Amazon Machine Learning is highly scalable and can generate billions of predictions daily, and serve those predictions in real-time and at high throughput. With Amazon Machine Learning, there is no upfront hardware or software investment, and you pay as you go, so you can start small and scale as your application grows.
Ai is changing the world around us:
47% of all current jobs are estimated to be capable of being automated.
We currently create 16% of jobs every decade.
To stay ahead and protect ourselves, we should become experts in identifying the proper use of AI on our own industries and being part of the conversation.
The skill is being able to evaluate the big picture and provide a path to achieve and add value from AI.
Ai is changing the world around us:
47% of all current jobs are estimated to be capable of being automated.
We currently create 16% of jobs every decade.
To stay ahead and protect ourselves, we should become experts in identifying the proper use of AI on our own industries and being part of the conversation.
The skill is being able to evaluate the big picture and provide a path to achieve and add value from AI.
Ai is changing the world around us:
47% of all current jobs are estimated to be capable of being automated.
We currently create 16% of jobs every decade.
To stay ahead and protect ourselves, we should become experts in identifying the proper use of AI on our own industries and being part of the conversation.
The skill is being able to evaluate the big picture and provide a path to achieve and add value from AI.
Google Photos
Example of how image recognition powers organisation of photos.
It's free to upload thousands of images - I have 10,000+ images stored with Google, for free.
They are assisting my everyday life, my experiences. And they know a huge amount about me in the value exchange. I will share my data, my life, in return for great assistance and a solution for my horrendously low iphone storage capacity.
It's not perfect though - dancing photos, great. Mountain photos are not.
Imagine a world where we rely completely on these algorithms. Would Google target me with a mountineering trip because it thinks I've been to some mountains in the past and I have in fact been to Cornwall? It's potentially a dangerous situation for advertisers.
Ai is changing the world around us:
47% of all current jobs are estimated to be capable of being automated.
We currently create 16% of jobs every decade.
To stay ahead and protect ourselves, we should become experts in identifying the proper use of AI on our own industries and being part of the conversation.
The skill is being able to evaluate the big picture and provide a path to achieve and add value from AI.
Artificial Intelligence doesn’t always deliver results. The filter bubble is artificial intelligence creating a situation for us which is making society worse and people more blinkered.
Our interactions with artificial intelligence and technology, generally, are based on doing things better, brighter and faster.
BETTER:
Test the technology
Where are we failing?
How could we be delivering better results?
Is there better technology we could be using?
BRIGHTER:
Engage the best minds. Collate thoughts of the best way to address the problem.
Engage the creativity of the industry. We build our own technology that addresses challenges – technology the genuinely ads value.
FASTER:
Use and build technology that makes us FASTER but also more ACCURATE
Challenge how easy it is to solve a problem