Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

What to look for in a prescriptive analytics platform

Improved interoperability can lead to greater connectivity that produces real-time data to enrich enterprise analytics tools—optimizing prescriptions, and ultimately producing better outcomes.

Analytics platform

[Images: Gorodenkoff/Adobe Stock]

Profile photo of Guy Yehiav

President · SmartSense by Digi

With a volatile economic outlook trending for 2023, retailers and national chains of all verticals should identify ways to maximize their existing tech stacks and equipment investments. For operators that already have substantial IoT infrastructure, the prospect of switching vendors to upgrade their tools may be inhibited by the cost of needing to rip and replace critical equipment components. As commercial IoT moves to a more open ecosystem of applications, retail operators have the opportunity to take advantage of hardware-agnostic technologies that integrate with existing hardware and software architectures (full disclosure: SmartSense offers this service).

Improved interoperability can lead to greater connectivity that produces real-time data to enrich enterprise analytics tools—optimizing prescriptions, and ultimately producing better outcomes. For example, existing Emergency Management Systems for refrigeration assets can be optimized by adding additional sensors for ambient and food temperature for full FDA compliance.

DEVELOPING CAPABILITIES FOR DIFFERENT TYPES OF ANALYTICS 

So how do enterprises maximize their existing resources to predict, manage, meet, and exceed pressing consumer demands? One word: analytics. The connected enterprise produces large amounts of data, which can be leveraged for analysis and actionable insights through four distinct types of analytics: descriptive, diagnostic, predictive, and prescriptive.

  • Descriptive analytics: provides an account of “what happened” in business operations 
  • Diagnostic analytics: slices and dices data to find the root cause of “why something happened” 
  • Predictive analytics: finds patterns in historical data to show “what will happen” in the future 
  • Prescriptive analytics: generates “what you need to do”—a guided action for a human or collaborative robot (cobot) to take to optimize the outcomes of the business operation  

Data analytics generate dynamic, actionable insights that guide management to make strategic business decisions that create real-time, satisfying customer experiences. Far too often, predictive analytics is wrongly attributed as being the latest, most advanced phase of analytics when, in reality, it’s not. Predictive analytics points to business trends and operational outcomes based on historical factors.

Prescriptive analytics, on the other hand, actually describes what action needs to be executed in the moment to optimize what will happen and deliver a desired outcome. By producing descriptive insights—via voice or text—prescriptive analytics leverages all the other mechanisms of the other types of analytics and its own added capabilities to generate accurate guidance in the moment that people can easily consume and act on.

DEFINING REQUIREMENTS FOR PRESCRIPTIVE ANALYTICS

A prescriptive analytics platform must be able to generate multiple forecasts and determine causal factors. Given those predicted forecasts and causal factors, the system then suggests the next best course of action. 

For an analytics engine to determine causal factors, it needs full access to as much telemetry data as possible: point of sale data, traffic levels, volumetric data, pH levels, inventory weight, temperature readings, humidity levels, weather data and patterns, consumer index information, GPS, truck arrivals, etc. Being able to intake and consume all these different sources of data requires an open platform that operates in a hardware-agnostic manner.

This requirement connects back to the need to maximize existing IoT capabilities and equipment infrastructure. The sum of an enterprise’s connected technologies comes from a variety of vendors and includes different categories of hardware, from sensors to cameras to mobile devices. In order to use all these different types of data, an analytics platform needs to be able to work with a variety of hardware devices, intaking the best data from the best-of-breed solutions that have been implemented across the enterprise for each specific operational function. 

Once all the relevant telemetry data is included in the analytics engine, it must be run through statistical analysis to determine which factors are casual (impact the outcome) and what the prescribed action to the human (or cobot) should be. This can help optimize the outcome to save time and money, increase operational efficiency, and maximize labor resources.

For operators looking to maximize existing hardware and technology investments during this uncertain economic environment, an AI-driven prescriptive analytics engine may be the right option to help you turn data into doing for higher levels of efficiency.


Guy Yehiav is President of SmartSense.