Finding Advantage in Edge Technologies
Felipe Thomaz
Associate Professor of Marketing and Advanced Technologies
University of Oxford
My work covers a diverse set of topics around edge technologies and novel methods in business and beyond. Over the past decade I have been focused on AI and Quantum technologies, their integration into business processes, and novel applications (leading to IP transfers to the University and an Oxford spinout company). Contextually, I’m interested in the function of media, advertising, and brands along with their influence on decision making, as well as the structure of markets. I work closely with several global companies, NGOs, intergovernmental agencies, and governments.
Exploring Applications of Artificial Intelligence and Quantum Technologies in Business
Beyond the hype of sales cycles, using new technology capabilities to change how we solve problems, approach operations, competition, and market structures.
Artificial Intelligence
Currently focusing on Graph Neural Networks, and a new application book. I was one of the faculty that started the Oxford Diploma in AI for Business, and was its Academic Director for years. Currently Program Direction of the Oxford AI Strategy and Leadership Program.
Quantum Technology
I’m currently working on Optimisation algorithms and the development of Quantum Game Theory. A book on preparing the business stack for any Quantum reality is forthcoming. And along the same timeline, I’m building the Oxford Quantum for Business Program.
Mathematical Foundations
I have almost always focused on graphs theory and combinatorics in my work, and it continues to follow me across contexts. I’m now working on higher-order and increasingly general graph-like structures, and exploring how to use them to solve problems that are not classically approached through such representations.
Novel Methods and Insights into Markets, Media, and Advertising
Advertising and Media are a industry worth more than USD1 Trillion / year, and yet, it often relies on technologies and techniques that are inconsistent fundamental aspects of market operation.
No Silver Bullets
Existing advice on media planning, a process akin to portfolio optimisation (with additional complications), often comes down to ‘copying someone else’s homework.’ Or worse, it simplifies the problem to maximising a single KPI like reach, awareness, etc. My work with colleagues on this domain show that media plans in practice are extremely inefficient, and often disconnected from actual outcomes. We show how different plan archetypes benefit different industries and different categories across different goals.
Media Functional Form
Departing from the question of how media (along with creative) combine to influence decision making, a more fundamental question arises on how and where each specific media channel works. Necessarily, for combinations to matter, they must be functionally distinct (beyond audience composition). Examining over 1 million consumer journeys ending in actual purchase decisions, I was able to identify channel specific functions across categories, as well as an individual level coefficient explaining a person’s receptivity to influence.
Broadly, What do we know?
The construction and execution of ‘advertising’ entails an immense amount of separate strategic choices and interdependent activities. It is also a ‘game’, in the sense that outcomes are dependent on competitors’ actions, and not just your own. However, if you were to examine the discussions on the topic, it would be easy to assume that it is simplistic and devoid of guidance, much less rigorous scientific examination (which is undeniably false). I was asked to collate a bibliography with commentary on this body of work, from the most meaningful papers in the field.