The big data challenge of transformation for the manufacturing industry
New ways of creating operations and business interactions have emerged from the volumes of data and the human desire to derive insights from all that data. CEOs and other C-suite executives often find themselves discussing or strategizing the importance of data and analytics in their planning sessions. They look at data products and the monetization of data, and they are worried about the disrupters such as the Ubers and Airbnbs of their respective industries.
Manufacturing companies are looking at similar desires and needs. The entire industrial sector is going through a major transformation these days, and that change is fueled by analytics, cloud technologies, the Internet of Things and volumes of data. Manufacturers are looking at these technologies alongside socioeconomic pressures such as cost cutting, globalization and localization, workforce changes, and regulatory and environmental constraints.
The formidable dark data challenge
Manufacturing can be a complex and highly process-oriented operation in which a large volume of data is generated and somewhat consumed throughout these processes. Processes such as design and simulation, build and production, sales and distribution, utilization and deployment, maintenance and service, and market and demand are data heavy and becoming nonlinear and multidirectional. Add to these processes the rapid expansion of sensory data that is quickly bridging the physical and digital gap.
Today, more data is produced within the four walls of manufacturing than ever before, and yet the ability of manufacturers to analyze this data and derive insights out of it is woefully lagging. According to a recent study by McKinsey & Company, “The Internet of Things: Mapping the Value Beyond the Hype,” much of the data collected from Internet of Things sensors today is not used at all. Consider one instance in which less than 1 percent of the data generated by 30,000 sensors on an offshore oil rig was used to make decisions. Another way to look at this example is that the other 99 percent of the data being generated could be dark data, from which a company cannot draw proper insights or make predict predictions.
A platform strategy for value-add efforts
This scenario warrants serious focus by enterprises in general and IT in particular to fortify their data and analytics platform and technology stack. With the right data and analytics platform strategy, manufacturing companies can shift their resources toward real value-add work such as operationalization of data products, developing new algorithms, collaborating on developing data products and so on. The approach, however, needs to be much different from the approach that many organizations have traditionally taken in their data warehousing initiatives.
In a new world of managing massive volumes of data, no longer will there be one big physical, monolithic data repository. The three Vs of big data are forcing us to think within a multifaceted framework in which a purpose-built data layer, along with the ability to logically combine and interact with data in multiple ways, is required. This requirement has exponentially increased the importance of the data lifecycle and data governance. A platform well suited for this polyglot big data journey needs to have several features:
- Advanced analytics, including machine learning, cognitive computing and Apache Spark
- An open ecosystem for incorporating third-party tools
- Automated deployment
- Comprehensive data governance
- Flexible integration and deployment
- Rapid data ingestion
- Self-service, role-focused tools
- Support for data of all types
- Team-based collaboration among all roles
Whether you’re in manufacturing or any other sector, you can advance your polyglot big data journey with the IBM Watson Data Platform. See how it enables data-driven professionals to collaborate in a simpler way and quickly find new and unexpected insights that deliver business-changing results.