You are here:/The Shainin Medal: Recognizing Innovation

The Shainin Medal: Recognizing Innovation

Richard D. Shainin

Abstract

The Shainin medal recognizes innovation. It is awarded for the development of unique creative approaches applied to the improvement of quality or reliability. The Shainin medal nominating committee wants to generate more quality nominations for its consideration.

This article provides some background on Dorian Shainin and the medal named in his honor. It describes the elements of innovation that the committee evaluates in assessing each nomination. Finally, it provides two examples from previous Shainin medalists. The first is a statistical innovation that helped ensure the performance of sophisticated radios. The second innovation improved business process problem solving and optimization.

Background

In 2003, ASQ established the Shainin medal to recognize innovation in the development of methods and techniques that improve the quality and reliability of products or services. To date the medal has been awarded 11 times.

Dorian Shainin (1914-2000) was an aeronautical engineer, certified management consultant and innovator. He was a founding member of ASQ and recognized as an honorary member in 1996.

His first statistical innovation was the Hamilton Standard Lot Plot. It used sample data to determine if a lot of incoming material was acceptable or suspect: flagged for 100% inspection. The analysis was graphical and it quickly became a standard for acceptance sampling with variables data. When it was published in 1950, ASQ recognized it with the Brumbaugh Award as the most influential paper of the year.

After his discovery of the Red X® principle in 1947, Dorian’s innovations focused on finding and controlling the Red X®. The Red X® principle recognizes that system variation follows a power function, e.g., Juran’s Pareto principle. That means that no matter how many sources of variation have been discovered and controlled, the remaining variation is dominated by one cause-effect relationship (often an interaction). Discovering and controlling that relationship is essential for improving system performance. In complex systems that cause is often hidden. Dorian developed a highly structured and disciplined approach to converge on the identity of the Red X® through a progressive search. He invented 20 unique statistical tools ranging from Component Search™ through Tolerance Parallelogram™ to Pre-Control. Dorian had a passion for developing statistical methods that were easy to use and powerful. When analysis was required, it was almost always graphical. He wanted his methods to be accessible to shop floor personnel.

Evaluating Innovation

Innovation is hard work. Ideation (the process of forming ideas) is the recognition of a need and the flash of brilliance that leads to a new approach. Realization is the development work to turn the idea into a usable method. It often requires refinement and multiple iterations to make the technique simple. Da Vinci is reported to have said: “Simplicity is the height of sophistication.” It may be sophisticated but it is not easy.  Innovations must be honed and refined to reach that level of sophistication. The final step is utilization. An innovation’s impact can only be judged by the benefits derived from its use.

Consider the development of component search. Dorian was faced with a difficult challenge. His client had high rework costs on an aircraft hydraulic pump with 40 components and was unable to meet the shipping schedule. By this time, Dorian was very familiar with Sir Ronald Fisher’s work in designed experiments.  He was also looking at the world through the lens of the Red X® paradigm. He recognized that if he swapped a component between a low output pump and a high output pump, he would be testing the effect of that component as both a main effect and a piece of an interaction. But years of problem solving experience also made him aware that he needed to assess the contribution to variation from the assembly process before he swapped anything. Swapping parts wasn’t a new idea. Aircraft and auto mechanics often swapped out suspect components when trying to repair a plane or car.

Dorian’s innovation was swapping between BOB and WOW units to assess the contributions from the various components and discover interactions. The realization phase involved the formal development of rules for estimating the assembly variation, stage 1, the graphical representation of the elimination phase in stage 2 and the evaluation algorithm for stage 3.

Since its development in 1956, Component Search™ has been used thousands of times to identify the assembly process step or the component(s) that contain the
Red X®. Seventeen years later, in 1973, Dorian awoke in the middle of the night with an inspiration.  He could use the same thought process with a modified algorithm to test the influence of many variables. That was beginning of Variable Search™, Shainin’s alternative to fractional factorial experimental designs.

In evaluating submissions for the medal, the Shainin Medal Nominating committee considers uniqueness of the technique (ideation); the degree of development required (realization) and the impact of the new method (utilization). Here are two examples from previous Shainin Medalists:

Patricia Cyr – Shainin Medalist 2014

Harris Corporation designs and manufactures radios for both First-Responder and Military use. Typically, each radio is tested for numerous performance parameters at multiple frequencies in both receive and transmit modes to ensure and document performance and compliance to requirements.

During 2010, Harris was planning to consolidate 600,000 square feet of manufacturing and test equipment to a new facility. The operations team’s challenge was to break everything down, move it, and have everything set up and in production again in two weeks. The Quality team’s challenge was to verify that post-move, all test equipment was producing the same results as before the move. The last step was station validation before production could resume. The analysis had to be thorough and quick so as not to delay the resumption of production.

Using univariate statistics was not an option. The sheer number of test parameters across multiple frequencies and power bands tested would have required a small army of engineers several weeks to collect and analyze data. The Type I error associated with such an approach was unacceptable. This challenge had to be overcome to prevent a work stoppage without accepting additional risk using a greatly reduced data set.

Patti Cyr, a statistician with a background in chemical engineering, had experience with multivariate analysis from previous work at Kodak. Multi-variate analysis is most often applied in the pharmaceutical, chemical and biotech industries. Patti recognized the opportunity to apply multivariate methods to a pseudo-spectrum from the radio frequency and functionality data (ideation).

OPLS is normally used to see the influence of multiple inputs on a single output. This electronics adaptation astonished the developer of the original technique. A key feature of the new methodology is the comparison of numerous outputs in a before and after analysis to reveal areas for further investigation. It is a Y to Y analysis as opposed to a more typical X to Y analysis (realization).

The data was collected using an MSA style DOE. Selected units were tested three times each before and after the move. This allowed for consideration of impacts on both average performance and variability because of the move. SIMCA, a software package for multi-variate analysis was used. Although SIMCA uses the correlation matrix for analysis, where each variable has its data centered and scaled to unit variance, additional preprocessing was needed for the data. The responses gathered in the final test have values as small as 10-5 and as large as 108. It was found that even using the correlation matrix, the responses on the order of 108 overwhelmed the analysis. To handle this difficulty, these responses were first deviated from their target values before the analysis began.

The true advantage using OPLS was in the isolation of the move as the driving force for differences. The use of SIMCA, or other statistical software package, was essential to communicate the findings in terms the engineering community could understand and act upon.

The use of OPLS and preprocessed data allowed for the quick and systematic analysis of over 500 electronic functionality responses, focusing on the tests with the largest shifts in value. The methodology was used by three additional engineers who could successfully assist in the analysis of the data without understanding the intricacies of multivariate analysis.

As desired, and thanks in no small part to this clever application of SIMCA, the team at Harris RF Communications could successfully disconnect, pack, transport, relocate, reinstall, calibrate, and verify that all test equipment was running correctly and producing comparable test results before and after the move within the timeframe required.

This technique is also used extensively at Harris RF to validate changes before they become part of products. Although a change may be made to address a function of the device, OPLS allows for thorough analysis to determine if there are any unintended consequences in other areas of functionality. It is deemed the “control radio process” and is now part of the documented Harris standard procedure for introducing software or hardware changes to existing product lines. Within Harris, it has been applied successfully to dozens of product and process improvements across several product lines, and has allowed changes to be implemented successfully without this risk of unforeseen or undiagnosed performance shifts (utilization).

Ms. Cyr has developed a new methodology comparing the holistic change in system performance following a process or product modification. While the modification is validated there can be unintended effects which might impact the system’s performance characteristics.

This methodology has been applied (utilization) to numerous changes at Harris including:

  • Product design changes
  • Process design changes
  • Test equipment relocation
  • Test equipment changes
  • Supplier changes such as new supplier or new source lot

 

The methodology is user friendly allowing quality engineers and managers to identify potential important changes without understanding the underlying statistics.

Jane Hoying – Shainin Medalist 2008

In 2003, Jane Hoying, a senior consultant with Shainin was faced with a difficult challenge.  An important client had requested the development of a simple and efficient means to solve complex business problems that paralleled the Red X® methodology for solving complex technical problems.  Jane was tasked with the assignment.  Her management specified a few key parameters: the system had to be true to Dorian’s principles including an investigative approach that converged on hidden root causes based on evidence not expert opinion; it had to be statistically simple and statistically sound; and the analysis and communications needed to be graphical.

Jane brought 25 years of automotive manufacturing experience, a degree in Chemical Engineering and strong success in applying Red X® Problem Solving across a wide range of industries and manufacturing technologies.

Red X® Problem Solving uses strategies based on the physical nature of a manufacturing system or product. It also relies on insights gained in talking to the parts.  In a business process, while the system may have some physical elements, the key components are procedural and there are few physical objects to be measured. Jane had two key insights (ideation). She would have to find a way to talk to the occurrences and she would use a functional description of the system rather than a physical description.

Jane developed a system for talking to the occurrences that revealed which system function had failed.  She adopted function models, a simple to understand graphical method for documenting system functional relationships to business processes (realization).

The first application solved a logistics problem that had resisted previous traditional methods such as process sequencing or value stream mapping. By revealing surprising breakdowns in functions, a $1 million annual cost was eliminated and the client had a deeper understanding of how their process was supposed to function.

The TransaXional® Methodology has evolved recognizing that there is a hierarchy of functions in every system and that foundational functions must be addressed first.

Within a few years, TransaXional® had been applied at six companies on a variety of business process including:

  • Logistics
  • Production Material Control
  • Information Technology
  • SAP Implementation
  • Quality Systems
  • Engineering Systems
  • Finance
  • Accounting
  • Purchasing
  • Personnel
  • Service Operations Prototype Vehicle Operations
  • Manufacturing Operations

 

The methodology has proven effective in solving business process problems, optimizing business processes and coordinating the implementation of new business systems.  Within a few years, it has saved tens of millions of dollars (utilization).

 

Summary

The Shainin Medal has been awarded 12 times to date.  Each medalist has been recognized for an innovative method that has improved the quality or reliability of products or services. A nomination form is available on the ASQ website (search for Shainin medal).  The submission should communicate ideation, realization and utilization. Testimonials from end users are appreciated. Submissions are due to ASQ by October 1.

2018-06-18T13:50:56+00:00
Contact us

Use the form below to contact us. We will reach you as soon as possible.