Leveraging the Three Stages of Analytics

William Cao, chief analytics officer at Catalyst, agrees, adding that predictive analytics uses modeling techniques to forecast what might happen in the future. “It generates data that is statistical [or] probabilistic in nature and answers the question ‘What might happen?’” he says.

Predictive analytics can include statistics, machine learning, deep learning, data mining, and simulation, according to Puget. Once again using the example of a retail store, he notes that predictive analytics “could be used to detect trends in sales”—a process that would involve leveraging historical sales figures, past price levels, and external factors such as weather conditions. Using these data sets, a retail manager could build a predictive model that relates sales to pricing, weather conditions, and any other available information, and then extrapolate subsequent sales levels. “Predictive analytics helps understand why things happened, and what is most likely to happen next,” Puget says.

“When applying predictive analytics, one doesn’t know in advance what data is important. Predictive analytics determine what data is predictive of the outcome you wish to predict,” Snow adds. She offers the example of building a model that is able to predict which customers are likely to churn. “In this use case, a predictive algorithm will identify attributes that are present in data sets of customers that have churned and identify those attributes in current clients,” she explains.

DEFINING PRESCRIPTIVE ANALYTICS

While predictive analytics identifies potential future outcomes, prescriptive analytics helps organizations decide which actions to take in response to the latest trends by determining the likely effect of different decisions. In other words, prescriptive analytics aim to “determine the best course of action to achieve an optimal business outcome,” according to Nevala.

“Knowing what might happen next is one thing; knowing what to do next is another, particularly when accounting for often complicated product and service portfolios, multiple engagement and fulfillment channels, and the needs and preferences of a diverse, mobile, and distributed customer base,” she notes.

Cao adds that prescriptive analytics is “based on predictive analytics, but takes it one step further to incorporate optimization or simulation to yield the most possible outcome or solution.” Put simply, he says prescriptive analytics answers the “what should we do?” question.

Prescriptive analytics encompasses multiple products and services and helps users decide which one should be used to engage with a particular audience, according to Cao. As an example, he says that prescriptive analytics could leverage data to produce actionable recommendations for customer-facing employees while they are engaging customers—in a retail store, for example—to deliver a better customer experience and, ultimately, increase revenue.

Olly Downs, CEO and chief scientist/technology officer at Amplero, characterized it similarly, saying that prescriptive analytics uses data mining and modeling to “provide paths to preferred outcomes” and “decide upon the correct message or experience to deliver to each user segment to drive a preferred engagement event.”

He notes that machine learning–powered platforms are particularly suited to this process.

USING PREDICTIVE AND PRESCRIPTIVE TOGETHER

While comprehending each of these three analytics phases individually is important, businesses need to understand how they work together if they hope to gain any sort of competitive advantage. This is especially true of the latter two stages, as predictive and prescriptive analytics are interdependent.

“Predictive and prescriptive analytics are symbiotic.… Predictive analytics identifies potential future outcomes or actions. Prescriptive analytics goes a step further to determine the likely impact or outcome of taking different actions, thus helping determine the best course to achieve an optimal or prescribed business outcome,” Nevala reiterates. “Predictive analytics can identify emerging opportunities, threats, or trends. For those of interest to the business, prescriptive analytics provides the ability to objectively analyze which of a multitude of potential reactions will best deliver a desired result.”

Often, prescriptive analytics are embedded into operational processes to facilitate real-time decision making, which can include presenting the right mobile offer to a customer just in time, providing ranked treatment options to a physician, or executing financial trades at lightning speed, she notes.

“The difference between predictive and prescriptive analytics sits in the key word ‘optimization.’ In predictive analytics, one learns from the past to predict the future. With prescriptive analytics, you will use the past to take control of the future, as you will optimize the outcome of your actions toward a certain predefined goal,” adds Steven Noels, chief technology officer at NGDATA.

To further illustrate this relationship, Noels uses the example of customer satisfaction. Predictive analytics will enable a company to foresee the degree of satisfaction that a customer will feel and flag those customers who are prone to churning. While this capability is useful for developing a better understanding of a customer, it does not provide advice on the best course of action to take to retain those who are unhappy, or even to identify which customers are worth pursuing.

Related Articles

Allego Launches Version 5.0 of Sales Training Platform

The latest release of Allego's sales learning and coaching platform delivers configurable dashboards and improved management and tracking of assigned learning.

NGDATA Releases Enhancements to the Lily Platform

The customer data platform's new Group DNA and OneAnalytics will help companies deliver more relevant customer experiences.

Amplero Upgrades its Artificial Intelligence Marketing Platform

The Learned Insights function aims to help marketers "nano-segment" their customer base to focus outreach at scale.

SAS Launches Machine Learning Platform SAS Viya

At SAS Global Forum 2017, the company unveils new products and hosts a talk from an Orlando Magic executive.

SAS Global Forum 2017, Day 1: New Releases Make Data and Analytics More Accessible

Data Step Debugger, SAS Graphics Accelerator, and Edge to Enterprise Analytics make analytics accessible to all skill levels.

InsideView Strengthens Integration With SAP Hybris’s Sales Cloud

The native Sales Cloud offering now cleans, updates, enriches, and standardizes account data.