Massive datasets are now available to businesses. These datasets contain highly personalized information about their customers. Having tools and techniques to interrogate that information can reveal hidden behaviours, trends and personal propensities – all of which can be used to vastly improve how marketing designs and delivers their messages. Predictive analytics has come of age.

According to Zion Market Research, the global predictive analytics market was valued at approximately $3.49 billion in 2016 and is expected to reach approximately $10.95 billion by 2022, growing at a CAGR of around 21% by 2022.

For businesses managing the omnichannel, being able to predict customer behaviour can have a profound and practical impact on business processes. From supply chain optimization to customer targeting and better segmentation, implementing predictive analytics is now a vital component of your business’s overall development strategy.

How your business engages with its customers has been rapidly evolving. Social media and mobile technologies have changed how consumers interact with the companies they buy from. According to Comscore mobile devices now account for 81% of all adult online minutes.

Being able to offer consumers services and products that support their buying patterns, has taken on a whole new dimension thanks to advances in data collection and analysis.

Says Marketo in their report: “Big Data is all about a firm’s ability to store, process, and access the data it needs to operate effectively, make decisions, reduce risks, and serve customers. In the context of marketing, Big Data refers to all the information created by today’s buyers—from keywords they use and content they share to the webpages they visit and emails they open. Individually, all that data is just noise – but predictive analytics can turn that noise into actionable intelligence.”

Speaking to Silicon, Mark Pybus, Big Data Practice Director, Contino said: “A key challenge for CIOs when it comes to predictive analytics is prescription. Predictive analytics is still insight, and there needs to be an understanding and a prescription to use that data to achieve the required aims. There is also a challenge, especially within machine learning approaches, in the ability to explain how a prediction was obtained. This can lead to a lack of trust in predictive analytics, especially when automation of existing tasks is part of the change, creating a barrier to adoption.”

Balancing the need to use more predictive analytics and ensuring the tools in use are accurate is a challenge for all companies. What is clear as marketing continues to evolve is advanced tools will reveal more about the choices consumers make when buying. The information revealed, is, however, only as good as the data the analytical engine is using. Here, a deep overhaul of how data is collected and stored is a prerequisite for high-level accurate predictive analytics.

Understanding behaviour

In the MicroStrategy, 10 Enterprise Analytics Trends to Watch in 2020 report,
Frank J. Bernhard, Chief Data Officer and Author, “SHAPE—Digital Strategy by Data and Analytics,” states: “In 2020, the spotlight on deep learning will be the nexus between knowing and doing as this technique becomes widely experimented and used to provide autonomous functions within finance, marketing, operations, and supply chains at the speed of computing power. No longer just a buzzword, the pragmatic advent of deep learning to predict and understand human behaviour is a tempest disruptor in how companies will perform with intelligence against their competitors.”

The shift in how businesses can utilize the information they have flowing into their organization will massively expand, as a perfect storm of data collection and data analysis delivers new insights into a range of consumer behaviour.

Steve is CEO, and founder of Black Swan Data also commented to Silicon: “What machine learning can do is analyze data from across multiple sources such as transactional, social media and CRM sources. By being able to analyze more data, it can give greater insight into patterns, which in turn leads to better predictions. This automation also increases the speed that companies can analyze their data, allowing them to process it in real-time rather than data from months ago. Another benefit that is brought by machine learning is its ability to learn from previous models and, adapt itself. This allows it to train itself to identify potential outliers which it can ignore, again making its final prediction more precise.”

It’s effortless to use the latest predictive analytics platforms without first understanding how they are built. What data they are using and, how this information connects to the end customer. Using AI systems, for instance, must have an ethical component to ensure privacy and security are maintained for individual customers.

“As more and more knowledge workers become comfortable working with data, they should also become conversant with data ethnography, or the study of what the data relates to the context in which it was collected, and the understanding that data alone might not give them a complete picture,” says Chandana Gopal, Research Director, IDC.

Indeed, when Silicon spoke with Stefan Hogendoorn, CTO at Cloud Technology Solutions, he explained the cultural and structural shift that is needed within a business to unlock the power predictive analytics offers: “It’s important to understand that any changes an organization makes based on predictive analytics could influence the environment they’re working in.”

Hogendoorn concluded: “So if the data sets don’t reflect what’s truly going on, an organization might end up with a ‘reality gap’ between what is happening and what the business thinks is happening. Bridging this reality gap is crucial for CIOs. While these problems have been identified in the industry for quite some time, the increasing amount of predictive analytics being done means we see it more and more.”

A predictive analytics future

Is marketing and customer service about to be revolutionized? The tools needed to deliver the insight necessary to make close personal connections with each customer are now available. The key is to use these tools effectively with the data being collected.

Says Mark Pybus, Big Data Practice Director, Contino: “Over the next few years, prescriptive analytics will see much more development and use. As machine learning becomes more achievable in more organizations, we’ll see predictive analytics used in new areas. However, we’ll likely see growing pains as businesses struggle to make use of the extra insight and, in some cases, make incorrect predictions that result in harmful actions.”

Cloud Technology Solutions’ Stefan Hogendoorn also advised: “Currently, predictive analytics is predominantly based on numerical data that we’re trying to collect and understand to make predictions on. A lot of information that we have today is not only stored in databases or numerical values but is also captured via text, video, images etc. But, by utilizing machine learning to process the data, our predictions become more accurate.”

Is the future predictable? Indeed, when focused on buying behaviour, businesses can now make accurate predictions about future purchasing patterns.

Peter Ruffley, CEO at Zizo, explained: “Data quality and data trust are two key challenges. As soon as we start to use data to predict what is possible, we have to ensure that we are using the right data, from the right systems, and that we can get that data to those people who need it to make decisions in a timeframe that makes sense. Predictive analytics is not forecasting as we are expecting to use the information to make operational, not strategic decisions. And as soon as the end-user believes that the data output is incorrect, they begin to either doubt or provide their own input into the predictive model – which can prove dangerous if not correctly monitored and assessed.”

What is certain is that the predictive analytical systems in development will gain more intelligence. As collected data also becomes more personalized and granular, the actionable data being extracted will increasingly form the basis of every marketing decision.

Silicon in Focus

Dr Iain Brown, Head of Data Science at SAS UK & Ireland.

Dr. Iain Brown is the Head of Data Science for SAS UK&I and Adjunct Professor of Marketing Analytics at University of Southampton. Over the past decade he has worked across a number of sectors, providing thought leadership on the topics of Risk, AI and Machine Learning.

What is the current state of predictive analytics?

“Though predictive analytics has been around for decades, it’s a technology whose time has come. More and more organizations are turning to predictive analytics to increase their bottom line and competitive advantage. This is due to growing volumes and types of data and more interest in using data to produce valuable insights. Faster, cheaper computers and easier-to-use software have accelerated adoption, plus there is a business need driven by tougher economic conditions and a need for competitive differentiation.

“With interactive and easy-to-use software becoming more prevalent, predictive analytics is no longer just the domain of mathematicians and statisticians. Business analysts and line-of-business experts are using these technologies as well.

“Predictive analytics is accelerating outcomes across multiple industries, whether improving customer experiences, fraud detection, or predicting medical diagnoses. The financial services industry, with huge amounts of data and money at stake, has long embraced predictive analytics to detect and reduce fraud, measure credit risk, maximize cross-sell/up-sell opportunities and retain valuable customers. Commonwealth Bank uses analytics to predict the likelihood of fraud activity for any given transaction before it is authorized – within 40 milliseconds of the transaction initiation.”

How much of an impact is Machine Learning having on the accuracy of predictive analytics?

“Machine learning is a method of data analysis that automates analytical model building. It is a branch of artificial intelligence based on the idea that systems can learn from data, identify patterns and make decisions with minimal human intervention. It’s a science that’s not new – but one that has gained fresh momentum. While many machine learning algorithms have been around for a long time, the ability to automatically apply complex mathematical calculations to big data – over and over, faster and faster – is a recent development due to much greater computing power.

“All of these things mean it’s possible to quickly and automatically produce models that can analyze more critical, more complex data and deliver faster, more accurate results – even on a very large scale. And by building precise models, an organization has a better chance of identifying profitable opportunities – or avoiding unknown risks.

“Examples of its use are in things like self-driving cars, online recommendations for consumers, knowing what customers are saying about a business on social media or fraud detection.”

How do you expect predictive analytics to develop over the next few years?

“Predictive analytics is set to become much more widespread in the health service among other industries, helping to improve patient care.

“The NHS has already announced a £250 million investment to create a national AI lab to help combat disease. Over the next few years, we can expect to see clinicians leverage AI, predictive analytics and machine learning to spot irregularities in blood samples, identify patients with tumours and predict cancer development.

“Investing in a secure organizational foundation must be prioritized if the NHS is to derive maximum value from predictive analytics. Employing a system where every clinician can access all data, administrator and application will facilitate advances in predictive diagnostics and forecasting abilities and ultimately, help save lives.”