Dr. Hair is Professor of Marketing at Kennesaw State University, USA and a visiting Professor at Henley Management College, UK.  He formerly was Director, Entrepreneurship Institute, Louisiana State University, USA. 

He has published 44 books, including market leaders Marketing, Thomson Learning, 8th edition, 2006, used at over 500 universities globally, and Multivariate Data Analysis, Prentice-Hall, 6th edition, 2006, which enjoys more than a 70 percent market share worldwide in the field of predictive analytics. This book can be easily bought through the website of Prentice Hall.

In addition to publishing over 100 refereed manuscripts, he has presented executive education and management training programs for numerous companies, has been retained as consultant and expert witness for a wide variety of firms, and is frequently an invited speaker on marketing and business research topics.

Predictive Analytics Insight was fortunate enough to grab this exclusive interview with the guru.

PAI: What is predictive analytics?
JH: Predictive analytics is the effective use of information to help us identify customer relationships as they currently exist or as they are likely to exist in the future. We can use this information to help us solve problems or pursue more sales opportunities. The surge in take up of predictive analytics can be largely attributed to the availability of data. In many ways, the theory behind predictive analytics was developed by scientists over a hundred year ago – they just didn’t have the data to make it a reality.

PAI: Why is it experiencing a surge in uptake?
JH: In the past, organizations threw away data because it was seen as a disposable bi-product of the business. The problem was that data was too expensive to collate and too expensive to store, but breakthroughs in technology have turned this attitude on its head. In the early 1990s, the cost of storing 1 MB of information was $15 – today it costs less than a penny. It’s also a lot cheaper to collect information about customer behavior, as this is often done automatically through electronic payments and the internet. Previously we relied on expensive face-to-face and telephone interviews. Furthermore technology is getting more sophisticated. In the US you can now program the cash register to distribute random surveys to customers.

PAI: Who is using it?
JH: The beauty about predictive analytics is that its application is not limited to any one sector as all companies are able to benefit from it. In the US, 90 per cent of all Fortune 500 companies are using the methodology. Having said that, some sectors have been much quicker than others to see its potential. Large telecommunications companies, airlines, banks and public sector organizations have all used predictive analytics to get closer to their customers as well as boost the bottom line.

PAI: But is it just limited to big firms?
JH: Most national economies are driven by smaller and medium-sized companies. We are certainly seeing growing interest among smaller firms, which are waking up to the value of this approach as they look to compete in a wider marketplace. However, the one sticking point is that this data must be collected over several years if it is to have the most value. Once you have built up a snapshot of where you are today, it is then much easier to predict future trends and consumer requirements. Many smaller firms are starting from scratch.

PAI: Who will make most use of it within an organization?
JH: Anyone who uses information about customers can take advantage of predictive analytics, but once again there are some departments that are more obvious candidates. Finance and marketing for example. Finance managers working for credit card companies and insurance companies can use the technology to identify, minimize and reduce fraud. In the medical sector, firms can use the systems to predict the effectiveness of surgical procedures or medication. It works best in any department where staff rely on clear metrics to take the best decision.

PAI: What are the main objectives?
JH: There’s little doubt that improving customer service is one of the most popular metrics, but this in turn feeds into the much bigger metric of profitability. The more closely you are able to map products to the needs of the customer, the more up selling you are able to carry out and the more profitable the business is likely to become. If a firm can constantly improve its customer service, then it’s likely to outperform the competition. On a macro level, this will eventually lead to a better stock price and a bigger market capitalisation.

PAI: How can firms use predictive analytics to gain competitive advantage?
JH: Firms gain competitive advantage by better identifying their customers’ needs and better matching their capabilities and products to suit. For example, in the case of a large supermarket like Tesco’s, the products they sell will vary by neighborhood. Predictive analytics means that they will be better able to understand how sales vary according to district to boost sales, improve supply chain efficiencies and reduce wastage. And if consumers quickly find the products they want, news of this great service soon spreads by word of mouth and cuts out the competition.

PAI: How else does it add value?
JH: Collating data and analyzing it throws up all sorts of interesting insights. For example, it is commonly accepted that 75 to 80 per cent of purchases in a supermarket are unplanned. One US firm identified through predictive analytics that young fathers buying disposable nappies were highly likely to buy beer on the same shopping trip. This information enabled them to plan the shop floor accordingly to ensure they passed as many products as they could on the way to the alcohol section of the store. Call center operatives perhaps gain the most from Predictive Analytics as it can quickly analyze the most suitable options for customers and communicate that directly to the operator through the computer screen. As a result, they are more likely to offer the customer something they really want.

PAI: What is the most important data to gather?
JH: Firms must evaluate their own, individual situation. But if you think that data might be useful, then collect it, especially as storage costs are so low. The reason is that more often than not, there are surprises hidden in data that you may not anticipate. I personally think it is much better to collect too much data rather than too little. If you do not have data, you can’t solve the problem. However, in deciding which information to gather, it is best to assemble a team of experts from across the business, each with a unique insight into which data is most crucial.

PAI: When can you expect to get a Return on Investment?
JH: Most firms should expect to see a Return on Investment in about two years. After two years, the bulk of firms will not only recover their original investment, but they are also likely to see a rise that is many times greater than the initial outlay.

PAI: Who is using it?
JH: A large telco company like Vodafone will use predictive analytics to develop new tariffs and bundles. Using predictive analytics, each month Vodafone will develop about 15 to 20 models, looking at different packages for different market segments. By analyzing mountains of data about their customers, the firm can identify significant changes in the market place and respond to this by developing the most appropriate products.

PAI: Can it help mitigate risk?
JH: Firms are also able to use predictive analytics to help them take managed risk. In general, a high-risk project also promises a high return if it pays off. With predictive analytics, firms are able to use the data to create a range of what-if scenarios, which drastically reduce the risk of failure. This helps firms weigh up a range of different strategic alternatives before going ahead with a project.

PAI: What are the typical obstacles encountered in a predictive analytics roll out?
JH: Some people worry that they will be polaxed by too much data on their customers, but this is rarely the case. New technology means it has never been easier to store and sort data. Indeed some larger companies store data for the future, even though they do not have a current use for it. The main two obstacles to successful roll outs are a lack of data and poor-quality data. Firms need to devote more effort to making data more useable so that it can easily be converted into key metrics that the system can read and act upon.

PAI: Any more issues?
JH: There is a significant lack of trained people in companies able to carry out predictive analytics. The trouble is the best operatives will need to boast a wide range of skills including marketing skills, accounting skills, and in the ideal scenario IT skills as well. But firms also need to train users how to get the most out of these systems. Many employees in the field will not like the idea of gathering data on their laptops, and try to stick to outdated methods such as pen and paper.

PAI: What is the future for predictive analytics?
JH: At the moment, the bulk of predictive analytics systems focus on numbers, but there are already plans to convert the focus to text as well. More advanced systems are already capable of recognizing phrases in different languages such as French, German and English, which will enable multinational firms to reap the benefits of predictive analytics even quicker. In the meantime there is certainly going to be even more data that needs collecting, and the quality of this data will become increasingly important.
As a counter to this, predictive analytics tools will become increasingly powerful as well as easier to use with user-friendly graphical interfaces and the widespread use of algorithms. And finally applications will become real-time, so that firms will be able to respond even quicker to fluctuations in the market place.