Zebra recently announced that it has acquired Profitect, a leading provider of prescriptive analytics for the retail and consumer packaged goods (CPG) industries. That has led many people to ask us: what is prescriptive analytics? And is prescriptive analytics the same thing as predictive analytics?
Knowing that business analytics can be categorically complex (even for data scientists), we’ve asked our in-house experts Tom Bianculli and Guy Yehiav to explain the benefits and use cases in the most simplistic way possible…
Your Edge Blog Team: Prescriptive analytics and predictive analytics sound similar. Are they the same thing?
Guy: No, they’re not the same thing, but you’re right, prescriptive and predictive do sound alike. There are actually four types of analytics starting with Descriptive, Diagnostic, Predictive and Prescriptive. In simplest terms, descriptive analytics is “what happened”, diagnostic analytics is “why did it happen”, predictive analytics is “what will happen” and prescriptive analytics is “what should I do”. Technically all four types analyze large volumes of data to identify business trends and “events” that could impact business decisions.
However, predictive analytics requires users or workers to understand and know how to interpret “the future”. Whereas, prescriptive analytics are analytics for everyone (including those at the edge) with a focus on future performance through identifying controllable factors and providing actionable opportunities that deliver results.
Your Edge Blog Team: So, what are the differences between each of these types of analytics?
Tom: The best way to describe the differences between each analytical approach is to consider how and when the extracted business insights will be used to inform decisions or actions:
- Descriptive analytics is the one of the most common types of business analysis conducted today. Business performance is analyzed to understand why certain business processes or actions worked and others didn’t. These insights can help you inform future decisions and actions.
- Diagnostic analytics takes descriptive analytics one step further using techniques such as drill-down, data discovery, data mining and correlations. This form of analytics helps you to understand why something is occurring, which leads to smarter decision making.
- Predictive analytics is akin to forecasting in the sense that you are leveraging past business trends to anticipate the probability of certain scenarios occurring, ideally helping to estimate the likelihood of a future outcome based on historical data patterns.
- Prescriptive analytics is one of the most advanced forms of business analytics. It uses machine learning and pattern detection rules and algorithms to identify anomalies in a company’s operations – and then prescribes a corrective action to optimize the outcome. That last part is what makes prescriptive analytics so valuable: it can intelligently prescribe the action for someone to take to optimize for a particular business outcome.
For example, Profitect’s prescriptive analytics solution will mine a retailer’s data to find inconsistencies that could impact sales or margins and then automatically notify stakeholders of the potentially disruptive issue using a simple description. But it doesn’t stop there. It also provides easy-to-follow guidance on how to address the inconsistency, whether further investigation is needed or a clear-cut fix is defined. It also identifies potential cost-saving and growth opportunities within the value chain, making prescriptive analytics a win-win.
Your Edge Blog Team: It’s not enough to have a predictive analytics solution then?
Guy: You’re right, prescriptive analytics goes beyond predictive analytics to give you the reason for those anticipated events and what to do about them so the outcome is optimized. Profitect’s prescriptive analytics solution even goes one step further by triggering a workflow that allows you to track subsequent actions taken by your team to respond to the issue or take advantage of the opportunity, which ensures accountability.
Your Edge Blog Team: So, in a way, prescriptive analytics is to the future what descriptive analytics is to the past as far as extracting the reason behind an outcome. Except, with prescriptive analytics, you’re in a position to take proactive measures to mitigate the risk or maximize the opportunity. With descriptive analytics, you’re making decision in reaction to something that’s already happened with the hope that you can replicate something that worked well or avoid repeating a mistake. Diagnostic analytics tells you why, but doesn’t provide any further actions. And predictive analytics can give you a heads up about what may be coming, but can’t tell you how exactly to leverage that information to your advantage?
Tom: Exactly. Prescriptive analytics offers very pointed guidance on what you should do in any event and why, as well as what could happen if you don’t follow recommended actions – and why.
Guy: Prescriptive analytics also translates the data into a description, this way you eliminate the personal and political biases that affect the way you read a report. Using plain-text removes any biases, ambiguity, or interpretation, and coupled together with a prescriptive action, increases efficiency and effectiveness. While Profitect’s solution can identify problems and suggest actions, a continuous feedback loop also identifies best practices that can be replicated across the enterprise.
Your Edge Blog Team: Are many companies using prescriptive analytics today? It seems like we hear more about descriptive or predictive analytics applications at the enterprise level.
Guy: Retailers have been using prescriptive analytics for several years to capitalize on the data they capture in-store and online. The Profitect solution is currently used by some of the most recognized retail and CPG brands in the world to improve inventory and pricing accuracy, reduce out of stocks, minimize unsellable merchandise, and fix assortment discrepancies. Our customers consistently report sales lift, as well as margin and labor productivity improvement. Ultimately this creates a better overall consumer experience.
Your Edge Blog Team: That’s interesting. Can you provide an example?
Guy: An interesting example would be one of our international grocery retailers whose U.S. division was challenged with Direct Store Delivery (DSD) vendors, process gaps that led to high markdowns, and technical issues impacting product resets being scanned for loss.
They started leveraging Profitect’s prescriptive analytics solution to track inventory data, remedy these process gaps and bring millions of dollars back to the company’s bottom line. To date, they have been able to create more than 750 “patterns” (or algorithms) to look for – and successfully find – areas that impact their business.
One example includes several stores identifying DSD deliveries that were not delivered according to the order. Now stores are informed of any DSD product that has been shipped to their store, but has not achieved sales. Profitect uses machine learning to cluster stores and compare behavioral consumptions and shipments. This helps identify any behavioral change which will point to a compliance or fraud issue created by the delivery company.
Your Edge Blog Team: That was a great example. But, how can prescriptive analytics be applied to areas beyond retail?
Tom: In retail, prescriptive analytics goes beyond inventory or vendor management. It can also be applied to functional roles like Asset Protection. For example, Profitect’s prescriptive analytics solution has been used to detect:
- Organized retail crime / Credit card fraud: A hardlines retailer saved $3.5 million in fraud in 2 days
- Gift card fraud: A department store identified over $30,000 in employee gift card fraud
- Pass offs: A footwear retailer recovered more than $50,000 of merchandise
- Employee refund fraud: A specialty goods retailer received $225,000 in restitution
We know that prescriptive analytics can also be used for similar fraud detection purposes at multiple supply chain touchpoints, not just at the point of sale. Imagine the impact that these analytical tools could make on combating counterfeit drug distribution. There are numerous possible applications in the manufacturing, warehousing, transportation and logistics space that we will explore in the coming months.
Guy: Additionally, due to the increasingly complex nature of supply chains, prescriptive analytics offers Collaborative Planning, Forecasting and Replenishment (CPFR) users a significant advantage over report-based systems. Using prescriptive analytics, raw data becomes “smart” tasks, distributed to the appropriate stakeholder with specific action steps to resolve. Under a report-based system, identifying who should perform what task could take a data scientist days, by which time the insight may no longer be actionable. It’s important to catch and weed out supply-chain inefficiencies and sources of waste in near-real time. Prescriptive analytics enables fast actionability.
Your Edge Blog Team: Is that why Zebra decided to acquire Profitect? To help Profitect scale their solutions and broaden the reach of prescriptive analytics to customers in other industries?
Tom: Zebra has been focused on expanding its global leadership in Intelligent Edge Solutions for some time. That is why Zebra Ventures first invested in Profitect in 2014. More recently, we have been seeking new ways to advance our Enterprise Asset Intelligence vision – to have every asset and worker on the edge visible, connected and optimally utilized. If you think about that last point – “optimally utilized” – it’s really about informing and empowering that worker to take the best next action. Offering the worker a prescribed action that will optimize their workflow is what it is all about. This is precisely what Profitect does and why we continue to invest and build out our capabilities at Zebra that contribute to advancing prescriptive analytics.
The addition of the Profitect technology and talent to our Zebra family enables us to more effectively build the "analyze and act" layers of our Zebra Savanna™ platform, which will enhance our existing Intelligent Edge Solutions. Combining the real-time data that Zebra solutions capture with Profitect's access to operational data, machine learning and analytics, we can now work with our partners and customers to empower front-line workers across all verticals, not just retail, with the focused insights they need to make smarter decisions and take faster, more effective actions.
Have a question for Tom or Guy about analytics? Share it in the Comments section below.
Editor’s Note: Learn more about how Zebra’s combined Intelligent Edge Solutions, including the Zebra Savanna IoT platform and Profitect prescriptive analytics solutions, can benefit your business.