Dan Jacob, Research Analyst, LNS Research
Manufacturers have historically struggled to generate meaningful Quality Management Analytics. With the advent of the Industrial Internet of Things (IIoT) and Big Data Analytics, it’s now more important than ever to have a long term Quality Management Analytics strategy.
How Important is a Quality Management Analytics Strategy and Why Now?
Numerous organizations have identified quality metrics as a top challenge, and have identified quality improvement as one of the most important IIoT use cases. In a recent LNS survey of 636 respondents, 34% cited “Quality metrics not effectively measured” as one of their top quality management challenges, right behind the #1 challenge of “Disparate quality systems and data sources” (38%). A separate LNS survey identified quality improvement as the 4th most frequently identified top use case for companies looking to adopt IIoT (23% of 252 respondents).
So, quality management metrics are important, but what necessitates a strategy now? The chart tells this story. It compares feedback from 1,733 respondents over the last 6 quarters (LNS Research survey, data extracted 12-4-2015). It depicts a rapid swing from a disinterested majority in Q3 2014 — 58% either didn’t know about IIoT or weren’t planning to invest — to a motivated majority in Q4 2015 with 68% planning to invest or investigating investment.
The trend analysis is clearly pointing to broad adoption in a fairly short time period. However, the challenge for senior management and quality professionals is that while a significant portion is interested in driving quality improvement initiatives using IIoT, organizations lack the core analytics needed to address basic needs of today.
Let’s be clear that quality management analytics don’t need IIoT to be valuable. However, IIoT and new analytics capabilities make fundamental and advanced analytics broadly accessible and therefore a disruptive force.
A Quality Management Analytics Framework
LNS Research recommends using a generic framework for quality analytics that captures Descriptive (What happened), Diagnostic (Why it happened), Predictive (What will happen) and Prescriptive (What action to take).
This framework aligns well with quality, as can be seen in the Quality Analytics Framework below, and allows capture of both Big Data as well as traditional analytics. Traditional analytics are based on structured data in a traditional database, data historian, or data cube whereas Big Data involves multiple data sources of structured, semi-structured, and unstructured data with a high degree of volume and velocity. Analytics maturity increases from left to right. Some common analytics include:
- Descriptive. Cost of Poor Quality, Count of Open Nonconformances, Process Capability (Cpk), Mean Time Between Failures (MTBF), or Supplier Defects per Million Opportunities.
- Diagnostic. Big Data Analytics approaches such as data mining and semantic mining, as well as traditional methods such as Ishikawa, Fault Tree, or Bayesian Belief Diagrams.
- Predictive. Big Data Analytics approaches such as data and semantic mining, correlation, regression, and machine learning techniques, as well as traditional methods such as Reliability Growth Analysis.
- Prescriptive. Less widely deployed currently; techniques often leverage Predictions combined with Diagnostics and extended data models such as Service, Logistics, Supplier, Financial, etc. to determine appropriate actions.
Developing a Quality Management Analytics Strategy
Unfortunately, there are no shortcuts in analytical maturity. Each class of analytics builds on predecessors, as can be seen in the following case studies:
Case Study 1 (Descriptive to Predictive): A Medical Device OEM IIoT-enabled its equipment to reduce customers’ unplanned downtime. Customers were frustrated by infrequent but unpredictable module failures mated with a traditional, reactive service model. The OEM found it could predict module failures in advance by adding new streams of machine and process data to existing QMS data. With this new information. the OEM was able to proactively call the customer to debug and fix the nascent problem before the customer was aware that there was an issue.
Case Study 2 (Diagnostic to Prescriptive): Elevator OEMs have long leveraged telematics to enable elevators to log QMS nonconformances and calls for service using built-in diagnostics. Other data such as number of door open/closes, time to open/close, floors visited, etc. were received but had little value. A leading elevator OEM applied Big Data analytics techniques, resulting in accurate failure predictions along with recommended service actions and spare parts. Grouping service activities allowed the elevator OEM to achieve previously unrealizable thresholds.
Building Support for a Long Term Strategy
It is clear from these case studies that a successful strategy will require a roadmap to build support for core Descriptive, Diagnostic, Predictive and Prescriptive Analytics. Taking these three immediate actions will provide a good start to this roadmap:
- Determine a vision for analytics. LNS recommends setting this vision with both a view of impact to the Operational Excellence model as well as potential out-of-the-box transformative benefits.
- Identify opportunities for harmonization. It is apparent both from the Quality Analytics Framework and industry case studies that a centralized but flexible EQMS data model is critical to organization-wide data interpretation and helping these organizations advance beyond Descriptive analytics.
- Benchmark your current analytics and process landscape against the identified analytics and harmonization vision to identify the most important first steps.
In our Research Spotlight on Quality Management Analytics, we provide additional guidance on how to apply LNS Research’s Digitization and Industrial Transformation Framework to deploy of Quality Management Analytics, as well as specific recommendations and guidance regarding architecture and partner selection.
Best Practices for Establishing Quality Management Analytics
Is your organization prepared for its next technological leap? Many organizations are stuck in decades-old technology approaches, generating inaccurate, dated, and disconnected analytics from siloed systems. Now is the time to set a high priority on improving quality management analytics. In this white paper, LNS Research shares new research to help you begin or strengthen your quality analytics initiative.