Prescriptive Analytics

Prescriptive analytics tools help you make decisions based on your real-world goals (“objectives”) and limitations (“constraints.”) This can be especially useful when you’re facing a business problem with multiple, conflicting goals (such as cutting spending while increasing production) and multiple constraints (such as time, distance, product availability).

Predictive analytics tools seek to find patterns in data, in order to predict what might happen in the future. For example, predictive analytics can predict who will launch which cyberattack, which experiments are more likely to prove the hypothesis, imminent machine failure, supply chain issues, infrastructure maintenance needs, and price movements—all before they happen.

Although there are countless ways to use prescriptive analytics, here are some real-world examples, with links to their stories:

• Transportation providers, such as Air France, Swissport, and Uber, use prescriptive analytics to create optimal routing, staffing, and maintenance plans.
• Professional sports leagues, including the National Football League and Beko BBL, plan their game schedules using prescriptive analytics.
• Manufacturers use prescriptive analytics to plan and manage the procurement, production, and distribution of their products.

Prescriptive analytics tools provide a detailed set of recommendations for how you can best achieve your goals, given your limitations. Although you can use it to automate decision-making, you can use it to inform your traditional decision-making processes. Its ability to explore what-if scenarios can be particularly helpful.

Say you were planning a trip. Predictive analytics can predict what you may encounter along your journey (weather, traffic, engine trouble), and prescriptive analytics can, given those predictions, identify the route that best helps you achieve your goals (fastest, cheapest, safest route), given your constraints (time, budget, speed limits).

Here are some additional examples:

• Use predictive analytics to predict supply chain issues, and use prescriptive analytics to identify the least costly way to reroute shipments.
• Use predictive analytics to predict cyberattacks before they happen, and use prescriptive analytics to identify the right investigators based on cost and skill.
• Use predictive analytics to predict imminent machine failure, and use prescriptive analytics to identify the best time to shut down the production line.
• Use predictive analytics to predict customer likelihood to buy more with targeted offers, and use prescriptive analytics to identify how many discount coupons to offer, in order to maximize revenue.

Prescriptive analytics doesn’t rely on historical data—which means you can make decisions for the future, even when it doesn’t look like your past. To use prescriptive analytics, you need to know three things:

• The goals you need to achieve (“objectives”)
  • For example:
    • Minimizing product costs
  • The limitations you’re facing (“constraints”)
    • For example:

• • • Minimum production of a given product
    • Required manufacturing time and cost of a particular machine
    • Raw material inventory

• Finished goods inventory capacity

• The questions you’re asking (“decision variables”)
  • For example:
    • In which order should we produce which products?
    • In which manufacturing facilities?
    • On what product lines?
    • In what quantities?

With this information, the prescriptive analytics tool can generate a detailed action plan for achieving your goals, given your limitations.