August 2012 | Commentary | IT Matters

Discovering the Value of Analytics

Tags: Logistics I.T.

Michael Watson, Ph.D. is World-Wide ILOG Supply Chain Lead, IBM,312-529-2871

Many companies are building analytics strategies, which use data to facilitate better decisions. To develop improved analytics strategies, consider the three different types of analytics: descriptive, predictive, and prescriptive. Each type uses data in a different way to provide a different type of value.

Descriptive analytics: Using data to improve how you describe or report on your supply chain. Obtaining a lot of data does not guarantee the ability to better understand what is happening in the supply chain. A good descriptive analytics strategy lets you easily use the data to drill down to meaningful details.

Descriptive analytics strategies typically include a business intelligence (BI) tool that provides this drill-down capability. It also often includes a real-time visibility tool that alerts supply chain managers when it is time to take action.

For example, if a supplier's shipment misses its scheduled vessel, the BI tool alerts management to take action, such as using air freight if arrival time is important, or waiting for the next ship if inventory is available to meet upcoming orders.

Predictive analytics: Using data to better predict what may happen next in your supply chain. Examples include forecasting future demand patterns based on past demand patterns; using internal and external data to estimate rates on various transportation lanes; and consulting customer order patterns to predict which items move together and which move in the opposite direction.

Being able to accurately predict outcomes allows you to develop smarter strategies. Predictive analytics relies heavily on statistical techniques. While traditional statistical analysis used only the data available within the organization, predictive analytics now merges your data with externally available data to develop better results.

Prescriptive analytics: Using data to recommend a course of action. This type of analytics is based on using a mathematical optimization engine that sorts through millions of options to determine the best scenario. Routing trucks to make a set of deliveries is a good example. If you know the deliveries and the cost of the trucks, an optimization engine can find a solution that is 10 to 15 percent better than what you could find manually.

One strategic example includes using network design optimization to determine the right number and location of plants or warehouses, and how products should flow through the supply chain. Another involves using inventory optimization techniques to help determine where to buffer inventory in the supply chain, and how much safety stock to keep of each item.

The trend toward prescriptive analytics and optimization involves embedding optimization to help make decisions closer to real time. For instance, ocean carriers can use optimization to determine how to best return empty containers to demand points. Retailers can supplement their replenishment systems with optimization to better balance inventory levels, promotional pricing, full truck pricing, and service levels.

Which type of analytics is best? When you develop an analytics strategy for your supply chain, you will most likely need a balance of all three approaches.