A rising buzzword in CRM and marketing circles, Predictive Analytics can sound a bit like magic. Our computers are going to simply “know” when someone is ready to shop for our company’s products and will automatically feed those perfect leads to our sales team.
Naturally, it’s not quite that simple, but that’s the aspiration. Marketers today are looking to use big data and machine learning to bridge the gap between the anonymous masses of the Internet and those customers they know are likely to (and would welcome the chance to) engage.
Predictive Analytics is the practice of looking for patterns within systematically compiled data in order to anticipate behaviors and trends.
An auto insurance company, for example, might use predictors (variables) like age, gender, location, and driving history to assess the risks a specific policyholder presents.
Predictive Analytics works when multiple data points can comprise a forecasting model. These variables need to be germane to the behaviors in question: for example, age might be a determinate factor in whether or not someone could potentially shop for the latest LEGO set. Which movies that person recently attended could help predict which LEGO set they might like, and so on. A variable that perhaps isn’t that useful, in this example, might be someone’s height.
That said – we don’t actually know. There may well be a correlation between someone’s height and LEGO purchases. That’s where we turn to computer modeling to tell us.
Using the power of machine learning, one can predict future probabilities with an acceptable level of reliability. Rather than banking on human guesses, or presumed if-this-then-that logical suppositions, computer analytics look for unbiased, actual patterns within large data sets and work to identify leading identifiers for behavior.
Scientists have used such modeling for many years in weather forecasting, medical prognoses, and geological research. As computing power and our access to data increases, more and more can be applied to predictive analytics – including customer behaviors.
Salesforce.com is uniquely positioned to lead the industry. Salesforce.com itself has millions of customers, and their customers have data on billions of their own customers. It’s all stored in the Salesforce cloud. In the past, a company could see (to some degree) the behaviors of its established customers. Salesforce now has the potential to pool behavioral information across multiple industries and thousands of companies. Likewise, browser and operating systems track individual behavior with increasing detail. The Internet of Things will no doubt yield yet more predictors on which to base models. We have access to more data than ever before.
Salesforce.com has made a number of acquisitions in this emerging industry. The platform will no doubt bring their sizable R&D muscle to bear in the years ahead. While initially available only to enterprise-level companies with the resources to devote to predictive analytics, Salesforce.com may popularize the capability as well.
This has obvious privacy implications, and we should suspect the self-serving agendas of commercial companies. However, predictive analytics could help deliver better ads to consumers as well. It could, for example, recognize when an individual isn’t ready to, say, buy a new pair of designer shoes.
Imagine a world in which the only ads you see are those that genuinely interest you. The goal of predictive analytics in marketing is to identify ideal customers ready to speak with us.