For a company to determine who its ideal customer is, how much he would spend, how to turn him from a lukewarm lead into a hot one, and what would keep him coming back might sound like it would require a lot of formulas and late nights at the office. Fortunately, thanks to predictive analytics, only the former is true. Even better, you don't have to be a math whiz to make these determinations, as a growing number of solutions are available that will do the number crunching for you.
Predictive analytics use algorithms that can be applied to large volumes of data to extract patterns and predict future outcomes. This is particularly appealing to companies overwhelmed by an increasing amount of unstructured data, also known as big data.
The creation of the social Web has led to a new wave of unstructured data and, subsequently, to technologies that have sprouted up to decipher it. According to Gartner, analytical systems are expanding their focus on capturing and correlating structured data to include unstructured data as well. Additionally, while machine-learning is not new, what is developing is what Gartner calls "man-machine partnerships" that "learn and deliver prescriptive advice" through analytics to make employees more efficient.
While companies have been investing in tools to help them uncover valuable business insights for years, predictive analytics is not about the simple business intelligence tools of yesteryear. "In predictive analytics, you don't know what data matters," says Mike Gualtieri, Forrester Research principal analyst and author of "The Forrester Wave: Big Data Predictive Analytics Solutions." "In business intelligence, you sit in a meeting and try to decide, 'What do we want to see in the dashboard?' 'What are the KPIs?' 'What reports should we use?' That's all important, but in predictive analytics, you're saying, 'Give me all the data you've got and then the predictive algorithms will find what's relevant.'"
Predictive analytics can help companies in a variety of ways. E-commerce companies, for example, need insights into customers' Web behaviors, which go deeper than simply measuring standard bounce and click-through rates. Thus, companies such as Quiterian, acquired by Actuate last fall, help e-commerce sites gain a better understanding of conversion rates by identifying cross-selling opportunities and the likelihood that a customer will buy in the future.
Any business that sells subscriptions, warranties, and contracts knows that running a business rooted in recurring revenue is no easy task. Based on Gartner statistics, it's estimated that $30 billion gets left on the table annually. This is because a majority of the sales models are tailored around new customer acquisition. A cloud-based solution such as ServiceSource, through its Renew OnDemand analytics, tracks key performance drivers specifically tailored to renewals. ServiceSource examines a recurring business' data streams, determines how often customers are using a service and to what extent, and then builds models that identify the propensity of customer churn.
Regardless of the environment, for predictive analytics to work most effectively, there needs to be some variety to the data coming from multiple sources, Gualtieri maintains. Clearly, there could be other market indicators, such as a competitor's launch of a new product, that could weigh on sales in one quarter. By bringing historical data together with information from external sources, such as credit ratings or Internet data, a company has a better shot at success.