Managing the data around complex machines such as airplanes, groundbreaking equipment or wellheads is challenging. The formal data (most commonly geometry, telemetry, and simulation data) may be stored in dozens of repositories. Related informal data – such as Word documents or PDFs from the Engineering team – could be just about anywhere. For the last three decades, organizations have efficiently stored these assets, but new business models such as predictive maintenance, multi-geography design centers, and the need for faster innovation have all conspired to render traditional data infrastructures obsolete. Forever.
THE IMPORTANCE OF UNSTRUCTURED DATA
Frictionless access to structured and especially unstructured information is critical to downstream activities such as customer acceptance, revenue recognition, test cell troubleshooting, product commissioning, new product design and warranty issues. Yet, even when critical formalized data assets are stored in PLMor other repositories of record, the information may not be readily accessible or complete. Often the thinking-about-tradeoffs that led up to the design are not included in those repositories. Instead, these answers are distributed across any number of servers, desktops, laptops and shared-drives, making the information practically impossible to track down.
Problems that demand a definitive technical response to a customer (e.g. warranty issues), equipment performance shortfalls (e.g., contract acceptance criteria), local contract requirements (enabling engineering to take place far away from the traditional product-domain center), and turnover of critical component suppliers, are all classic examples that quickly unveil the “big data elephant in the room;” namely, the inaccessibility of data.
WHY TRIBAL KNOWLEDGE FAILS
Faced with the inaccessibility of data teams turn to what they know- relying on time-tested tribal information of experienced personnel. However, this “old reliable” back up plan –tribal knowledge—is folly in a world where unstructured data sets are growing exponentially and product lifecycles span decades. In truth, even if the entire original engineering team remains intact, critical information surrounding the design process is needed to reconstruct calculations, identify issues and analyze results. Without a formal management, governance or data-recall facility that captures all the information surrounding a product, from design, to simulation, test, production and maintenance engineering decisions are based on conjecture and indirect validation instead of empirical data.
The transition from educated yet subjective choices to informed and quantitative decisions, and operations-based data-driven insights requires a large-scale data infrastructure that can consolidate data with a frictionless path to find, mine, manage and reuse your unstructured enterprise data.