Breaking through the hype: the realities of data science, artificial intelligence and machine learning

Artificial intelligence (AI), machine learning (ML), and data science were three of the biggest buzzwords of 2018 and it looks like that trend is set to continue. Fospha’s Head Of Research Alexandra Darmon helps cut through the jargon to explore what makes all three so important, and how they differ, in the context of modern, digital marketing.

“First of all, data science isn’t just changing the business landscape, it’s influencing the way people interact and make decisions, infiltrating every facet of our lives. Where there is data, there is data science. Any situation and any problem involving data observations can largely be solved using data science – this is what makes it such a crucial business tool.”

Fospha’s head of research – Alexandra Darmon

While there are examples of misuse, where data science is used for fraud and cybercrime, the technology is still overwhelmingly underutilized.

So, what is data science?

Data science is a multi-disciplinary blend of data analysis, algorithm developments, and technology, designed to solve complex analytical data problems. It can allow powerful predictive and analytical capabilities such as multi-touch attribution (MTA) and marketing mix modelling (MMM).

With masses of consumer data at the fingertips of all businesses, data science helps to turn the numbers into key, actionable insights.

With consumers sharing a huge amount of data every day, research from IDC shows that the volume of digital data is increasing by 40% to 50% annually.

By unifying statistics and data analysis, data science can extract knowledge from both structured and unstructured data, diving in at a granular level, to understand complex behaviors and trends, and therefore inform future business decisions.

Forrester predict that 80% of all data is currently unstructured, with this number set to rise to 93% by 2020.

With the volume of unstructured data growing rapidly, AI and ML can help marketers make better sense of large amounts of data, coming from a variety of sources and formats across the digital media marketplace. The insights from this process can lead to more unified measurement of marketing performance as well as delivering accurate channel-mix predictions. However, marketers have been slow to adopt and properly apply these disciplines within their wider marketing strategies.

Forrester’s 2017 Consumer Wave identified that not a single marketer considered AI/ML a priority when buying AdTech.

As marketers face an ever-increasing degree of complexity in measurement and strategy, they need a solution that’s designed to evolve and innovate in line with the industry. With this abundance of choice, comes a great deal of responsibility, especially when it comes to machines making the decisions for us.

Understanding artificial intelligence and machine learning

As the consumer digital journey becomes increasingly complex, a range of AI-fuelled ad tech solutions have emerged promising to provide more accurate measurement and more effective personalization.

With the ability to acquire human-like intelligence skills, AI is the power of a machine to make decisions and provide solutions quickly and automatically. However, it’s important to note that not all businesses operate from the same definition of AI and while the foundation is generally the same, the focus of artificial intelligence shifts depending on the entity that provides the definition.” What is certainly true is that machine learning (ML) lies at the heart of it – a method of developing intelligence which feeds the ‘machine’ data so that it can ‘learn’, by adjusting the coefficients and parameters of the model for a specific case or problem. For example, by stitching together fragmented digital journeys (with touchpoints spread across multiple devices), ML can help to create a more unified view of consumer behavior and buying patterns.

We’re still at the very beginning of what these disciplines can ultimately deliver, but it is clear that a total lack of human-led decision making can be risky, leading into many questions around ethics. Effective marketing requires a powerful combination of AI/ML tools with real people with bias and human consciousness.

This human approach to optimizing the power of data science is central to Fospha’s ethos and approach. Excited about the future of artificial intelligence, we’re determined to be at the center of new developments. We recruit the best data science talent through our partnership with Imperial College London, and our annual hackathons (which bring together over 200 of the best computer science graduates in any given year).