Everyone wants a glimpse into the future; the insights and information found there would be inherently lucrative for the individual (or business) that possessed them. And, while soothsayers haven’t yet proven themselves to be a reliable business asset, today’s predictive analytics software has.
“Predictive analytics,” in whole, is data analysis that provides enterprises with predictions about future events, as based on current and historical data. This longstanding science has been around in various forms for centuries, but has only recently become reliable—and affordable—enough for most companies to use in their day-to-day processes.
With success, however, a number of myths have grown around big data and predictive analytics. The five below, especially, need to be dispelled so that ventures of all sizes and stages can begin to enjoy smarter, more efficient decision making:
Myth 1: “Big data is a silver bullet.”
Predictive analytics has the promise and potential to be a pervasive trend that shapes the wider economy, especially if more business decision makers become data driven (as opposed to depending on pure “gut instinct”). It is also true that increasing the amount of, and access to, information— especially about your customers’ behavior—will provide a competitive advantage to certain businesses, just like the Internet. However, in many cases, relatively undifferentiated Internet and Web capabilities did not result in great business growth, and the same is true of instituting predictive analytics.
Big data is not a silver bullet for the enterprise. Instead, better data management and analytics are tools to help organizations make better decisions. Even “small data” can be very useful in providing small and medium businesses with a roadmap of where to invest, build and diversify without having to make a big IT investment.
Myth 2: “Only big companies need predictive analytics.”
While it is clear that Amazon, Target, Walmart, Zipcar and other big businesses have been the early adopters (and significant beneficiaries) of predictive analytics—especially customer analytics—companies of all shapes and sizes stand to benefit from the intelligence analytics bring. For example, analyzing customer retention patterns based on clusters of customer profiles and behavior—and then using said data to design targeted promotional offers—can have an immediate impact on any venture.
Myth 3: “The best way to introduce predictive analytics to a company is going ‘bottom-up’ or ‘top-down.’”
For some, the bottom-up approach involves IT personnel and data analysts to implement an enduring solution. For others, a top-down approach involves solving this business challenge with significant resources, strategy, and culture, and therefore should involve the CEO, CMO, or other executives.
In reality, a predictive analytics implementation does not have to conform to either edict. A bottom-up process may establish a good foundation for companies. Beginning with a given department in other companies—especially in the marketing group—can be very fruitful. Similarly, a top-down process can be either productive or short-lived. Executive involvement does not always guarantee success.
Myth 4: “To implement predictive analytics, you’ll need your own Ph.D.”
Predictive analytics is going retail (or being “democratized,” as we like to say). By this, we mean that it is becoming widely available and does not require a multimillion-dollar IT infrastructure.
And, while unique big data problems may require a Ph.D. (or even a group of Ph.D.s), many new areas of predictive and business analytics are available for easy access through software as a service (SaaS) solutions. What’s more, time-to-value does not equal months or years anymore; you can now get useful results in a shorter timeframe, and without the need for your own Ph.D.
Myth 5: “All we need to do is hire [insert your favorite consulting or technology firm here] and we’ll have predictive analytics.”
There are a group of companies who see predictive analytics as a technology or a software problem. And, they have a list of “go-to” companies—sometimes technology vendors with management or technology consultants—which go about solving predictive analytics in a traditional way by selling large amounts of infrastructure, data storage, software, and hardware to companies. In reality, predictive analytics is more of a business and cultural problem that needs more than just technology (or episodic visits by management consultants) to establish and institute an ongoing solution.
Conclusions
There is a lot to think about when considering the addition of predictive analytics to your business decision mix. And, in fairness, there is an array of different approaches—and ultimately different outcomes—from which organizations may choose. During the review of products and options, however, it is good to separate myths from the practical and systemic realities that accompany this science.