The datafication of supply chain operations creates a strong need for analytics that can make sense of all of that data. When you break down analytics conceptually, they fall into 3 categories: Descriptive, Prescriptive and Predictive.
Descriptive Analytics: Looks at historical information and answers the question “What has happened?”
Prescriptive Analytics: Looks at historical information and answers the question “What actions should I take?”
Predictive Analytics: Looks at historical data and using statistical models answers the question: “What could happen?”
Descriptive Analytics – Understanding the past
Currently, the majority of analytics systems focus on descriptive analytics. They compile and transform the raw data into usable information and then visualize it for easier consumption. Many of the BI tools have the capability to perform prescriptive and predictive analytics, but they require a substantial amount of contextual understanding of the underlying data set and custom coding. Some examples of descriptive analytics within Supply Chain Operations are:
- Pick accuracy
- Cost to serve
- Employee performance scores
- On time delivery
- Volume / through put
Descriptive analytics will likely remain the majority of analytics because they provide an excellent scorecard for your business and insights as to how you are performing.
Prescriptive Analytics – understanding what to do
The next stage of analytics is to apply rulesets against the descriptive analytics to “prescribe” actions to be taken on the information. For example, the system would send action items to supervisors to acknowledge employees that are exceeding performance expectations and to address employees that are not meeting expectations. With the broad adoption of mobile technology, the ideal system would actually push the action items to each person based on their ruleset configuration in near real time.
The challenge with descriptive systems is that they require the user to sift through volumes of information and then determine what actions to take. By codifying the rulesets into a prescriptive analytics system, it reduces the effort required by the user to determine what actions need to be taken. In today’s fast moving operations environments, simplifying the process of determining which actions to take through prescriptive analytics models can have substantial rewards.
Predictive Analytics – seeing the possible future
The final evolution of analytics is to use the historical data to predict future possible outcomes. Big data and modern software tools make this much more achievable than in the past. Here are some possible predictive outcomes based on currently available data:
- Labor Forecasting– correlating historical volume cycles and current employee, you can forecast your staffing needs.
- Equipment breakage– material handling and vehicle telematics open a whole new opportunity. By analyzing vehicle telematics and comparing it to part failures, it is very feasible to predict a vehicle breakdown before it happens.
- Employee Turnover– analyzing employee performance trends correlating with time and attendance can identify employees who may quit.
New data sources continue to come online and the amount of data available grows exponentially. The challenge all industries face is how to make use of that data without suffering from over complexity and paralysis of analysis. It is important to understand both the underlying data and the real world problems you are trying to solve.
Easy Metrics focuses on simplifying this world of data and providing easy to use analytics for Supply Chain Operations with the goal of helping its customers improve efficiencies and reducing costs.