In their recent case studies, a group of data analytics collaborators from Microsoft and two universities discussed where predictive analytics is having the biggest impact and shared new insights on the value of predictive/planning analytics across industries and along value chains.
Here are some key findings from their case studies:
- Airlines are particularly interested in predicting mechanical failures in advance so they can reduce flight delays or cancellations.
- Some airlines are able to predict the probability of aircraft being delayed or canceled in the future based on relevant data sources, such as maintenance history and flight route information.
- A machine-learning solution based on historical data and applied in real time predicts the type of mechanical issue that will cause a flight delay or cancellation within the next 24 hours. This allows the airlines to take maintenance actions while the aircrafts are being serviced, thus preventing possible delays or cancellations.
- Similar predictive-maintenance solutions are also built in other industries. For example, in the oil and gas industry, companies are tracking real-time telemetry data to predict the remaining useful life of an aircraft engine and employing the data to predict the failure of electric submersible pumps used to extract crude oil.
These cases help highlight a few general principles:
- The value derived from the analytics piece can greatly exceed the cost of the infrastructure. This indicates that there will be strong growth in big data consulting services and specialized roles within firms.
- Big data is less about size and more about introducing fundamentally new information to prediction and decision processes. This information matters most when existing data sources are insufficient to provide accurate or actionable predictions due to small sample sizes or coarseness of historical sales (e.g. small effective regions, niche products, new offerings, etc.).
- The new information is often buried in detailed and relatively unstructured data logs (known as a “data lake”), and techniques from computer science are needed to extract insights from it. To leverage big data, it is vital to have talented data engineers, statisticians, and behavioral scientists working in tandem. The term “data scientist” is often used to refer to someone who has these three skills, but in our experience, single individuals rarely possess all three.
Predictive analytics also has the potential to be applied in ways that disrupt existing processes and create radically new applications. For example, machine-learning models that take massive data sets as inputs, coupled with clever designs that account for patient histories, have the potential to revolutionize how certain diseases are diagnosed and treated.
Another example involves matching distributed electricity generation (e.g. solar panels on roofs) to localized electricity demand. Huge value is unlocked by equating electricity supply and demand with more efficient generation.
The immense value of predicting demand more accurately, creating better pricing, and performing predictive maintenance easily justifies large firms’ investments in big data infrastructure and data science.
The value of radically new applications is challenging to understand and naturally causes speculation. It is reasonable to expect losses for many firms, due to uncertain and higher-risk investments, with a few firms earning spectacular profits.
Two of the principles we have listed above focus on the need for (a) consultive thought partners to help organize the analytics infrastructure and (b) data scientists and analysts to bring new data combinations and relationships to decision-makers.
A truly competitive enterprise needs a contemporary, on-call set of analytics that is supported by big data and everything else and is based on a streaming architecture.
Our Austin, Texas location keeps us at the epicenter of planning analytics and data-driven innovation and keeps our global capabilities and services at the forefront of our industry.
Email us at kcerny@mia-consulting.com or call us at 512-478-3848 to start a friendly, productive conversation.