Harnessing Big Data with a Systems Thinking Approach – (A Harley Davidson Case Study)

With 90% of the world’s data created in the last two years, what can we expect our data vaults to hold two or even twenty years from now? Today we measure our lives in peta-bytes but by 2020 estimates show a 2,300% increase in the bits and bytes that will define our lives. 35 zeta-bytes to be exact. How then can we as a society leverage the intrinsic value of so much data without getting bogged down with its complexity?

Around the turn of the century, we experienced a similar moment of euphoria when retail outlets opened ‘virtual stores’ and sold products to online buyers. A famous IBM TV ad once depicted an overwhelmed young company whose products went from a few online orders a day to hundreds of thousands. In many respects we have come full circle and are back at the starting gate of yet another era of unprecedented growth only this time instead of millions of orders, the focus is on zillions of data points.

In 2000 CEOs focused primarily on IT integration and supply chain strategies to fulfill a surge of orders. Their managers implemented the latest e-commerce packages, leveraged the cloud to reduce costs, broadened and compressed their global supply chains, and trained their workforce to adapt new work flows. Success was determined from a customer’s positive experience, measured primarily by the number of accurate and timely deliveries.

Today, the paradigm has shifted away from a transaction centric one to customer centric. Companies no longer wait for customers to buy but instead develop sophisticated algorithms that can compare a specific customer’s purchase history with multiple data sets including credit rating reports, recent purchases, and most extraordinarily, their genuine propensity to buy based upon the web pages they most commonly visit. Surprisingly, web behavioral data has become a powerful data complement that can offer unprecedented efficiency benefits to both the merchant and the consumer. Customers receive compelling suggestions, while stores inventory the products their customers will most likely purchase. It’s a win-win for both. Issues of privacy remain a sticking point for some individuals, but, as the benefits to the consumer improve, even these issues are expected to become less significant.

Striking the optimal balance will be tricky especially when the journey also involves flogging through mounds of unstructured web data. One approach being talked up within academic circles is systems thinking.

MIT’s SDM Conference – (sdm.mit.edu)
At a recent Systems Design Management (SDM) conference at MIT called “A Systems Approach to Big Data: Going Beyond the Numbers”,  Senior Lecturer J. Bradley Morrison greeted a packed audience with a refresher on Systems Dynamics; the study of how all the various components within a company (people, materials, contracts, etc), for example, interact and react together to create a product or service. Morrison’s ‘Back to the Classroom’ exercise offered new insights on how the principles of ‘systems thinking’ that today help companies scale their global operations can also be applied to leverage the new era of big data. His explanation is also testimony to the incredible versatility of ‘systems thinking’ and systems design management principles.

Morrison divided ‘Systems Thinking’ into various key areas. First off was ‘Dynamic Complexity’, which evaluates reactions when a smooth-running assembly line becomes inadvertently interrupted; for example, when a supplier’s product fails and an alternative source is unavailable. According to Morrison, unexpected manufacturing events can also have a direct affect on a company’s moral and effectiveness. The reverse is also true where systems that operate smoothly can greatly improve on what Morrison refers to as the ‘Mental Model’.

Another key area is ‘Stocks and Flows’, which Morrison dubbed humorously as  ‘Bathtub Dynamics’.  Similar to balancing the water level in a bathtub with running water, systems thinking can help calibrate inflows (i.e. inventory-build up) versus outflows (i.e. sales). The depth of the bathtub is determined by a company’s internal competitive advantage. These advantages vary widely but with regards to the alignment of systems thinking with big data, Morrison focused on skills training as a key differentiator.  He highlighted his points with a case study from a US motorcycle manufacturer, Harley Davidson.

Harley Davidson Case Study
In the late ’90s, Harley Davidson implemented lean manufacturing systems throughout its operations. Management leveraged their strong union relations to encourage employee input. The response was overwhelming. After numerous meetings, participating employees elected to improve the rotor area on the shop floor. Soon new signs went up. Space allocation was optimized, and the new employee-driven initiative became a reality. Management was pleased with their progress. The improvements paid off with an increase in productivity from 70% to 94% without the need for additional floor space. All in all the project reflected a success story until a common syndrome called ‘process degradation’ set in.

Like an ambitious diet plan, the idea reached its goal only to become unsustainable thereafter. Unaddressed issues such as an understanding of who was responsible to maintain the new process wedged away the achievements. The collaborative efforts to engage and integrate the surrounding workforce were weak and gave way to a ‘do-it-yourself’, ‘if-and-when-you-can’ approach. Despite the obvious benefits, workers returned to their old habits inhibiting further progress.

Who was to blame? …management, labor, or both?

Improving productivity with limited resources is a common problem with every company. That is why CEOs leverage technology, timely intel, and training whenever and however possible. Of these three, Morrison points to training as the greatest challenge and the most commonly ignored. Even when training is available, the type of training that he recommends is not classroom-style but rather on-the-job training.

“Learning a new skill is one thing but learning how to replace one’s old habits with a new skill is quite another,” Morrison  explained. “Workers need the opportunity to ‘change their own mental model’ before the true benefits from increased productivity can be fully realized.”

According to Morrison, managers should give their workers the opportunity to learn a new system on their own terms, regardless if it requires allocating extra time during a shift or work day — even as much as 50% more time. Unless workers are given a chance to appreciate the time saving benefits on a personal level, they will more than likely return to their old habits and simply ‘add-on’ the new changes rather than adopt them for their intended benefits.

Looking ahead…
In the next few years, new skills training will involve some form of data analytics integration. As data sources swell in every part of a business, relying on a specialized team to manage the company’s data needs will become unsustainable, especially when experts tell us that big data and data analytics, done right, depend upon the seamless collaboration and exchange of data from every corner of the company. Visionary CEOs will require every employee to learn how to collect, disseminate, compare, and use data from multiple sources. Soon-to-be, ‘unsilo’ed’ departments will depend upon each other in an entirely new manner, since the data they collect will determine the value and quality of data for the rest of the company.

Just how CEOs balance this data exchange while injecting behavioral changes among their ranks will become a number one priority for years to come. …and yet will CEOs have the foresight to allow their employees to experiment with best practices on company time? As we learned from the Harley Davidson case, those leaders that do allow their employees to adopt new behavioral changes on their own terms will more than likely achieve measurable, sustainable advantages. On the other hand, those who follow the herd by, for example, hiring more data scientists to solve their data issues, may lose an unprecedented opportunity to transform their workforce. At this juncture CEOs would do better implementing a systems thinking approach today that will allow every employee to eventually become a specialized big data provider/user for the company.

© 2014 Tom Kadala

‘Systems Thinking’ – Your Next Competitive Edge!

Imagine for a moment that a friend followed you with a web cam and recorded every moment of your typical work day.  What could you learn from so much data?  Probably not much, unless you matched each video frame with a related task. Once you did, however, you could pinpoint areas for improvement by comparing your activities on video with the expected minimal requirements (time and money) to complete each task.

The clinical engineering term used to describe this comparative analysis is called ‘systems thinking’ where choices for a set of outcomes are optimized using benchmark data. For non-engineers, ‘systems thinking’ could be described as an exercise in time management or as an example of how a person should best maneuver while driving through rush hour traffic. Surprisingly, if you are not an engineer but know how to drive effectively in traffic, you may be more of an intuitive expert on ‘systems thinking’ than you realize.

So, what is ‘systems thinking’ and why should CEOs view it as their next competitive edge for years to come?  

At a recent conference on ‘Systems Thinking for Contemporary Challenges’ held at MIT, thought-leaders, CEOs, and entrepreneurs, (some representing various Fortune 100 companies), shared their thoughts and experiences. At first, I wondered why so much attention was being given to a decision-making process that appeared so intuitive. It was not until I realized that the clinical term, ‘systems thinking’ actually has two very different meanings.  One definition applies to how an individual must think to solve problems, while the other applies to how groups of individuals must think collectively to find solutions. Then it became obvious to me that the latter was the principal reason for the conference.

Another way to look at this important distinction is to split ‘systems thinking’ into two separate definitions, one for the individual within a company and the other for a group of integrated companies that work together on large projects.

Definition #1 – An Individual learns to think in terms of systems
The first definition focuses on an individual’s ability to use ‘systems thinking’ for self-improvement and measures the potential effects from a group of employees that improve at the same time.  For example, the web cam data exercise stated earlier would have given you a frame-by-frame glimpse of your daily routines and potentially exposed hidden areas for improvements. Now, imagine if everyone in your company analyzed their daily activities frame-by-frame too?  No doubt, the sum of their improvements would translate into a significant productivity boost for the entire company.

Definition #2 – A Group of Systems learn to think together
The second definition looks at how a group of companies can effectively work together as a ‘network of systems’ to complete huge complex projects such as the production of a fleet of fighter jets at Raytheon or the design of new wind turbines at GE.  As a fragmented bunch of contracted companies, these entities would find it impractical to use a web cam to video their collective daily activities the same way we proposed in Definition #1. Instead they would apply proven ‘systems thinking‘ tools and methodologies that are specifically designed to coordinate and optimize the collective efforts from multiple companies.

How ‘Systems Thinking’ Works…
‘Systems thinking’ begins by breaking down processes into their minimal components, even down to a molecular level, if need be. The data representing the flow of information from people, machines and business objectives are thrown into the same soup and mapped onto a Design Structure Matrix (DSM) that visually connects the dots among people, activities, priorities, and time tables. Even management is treated as just another series of systems and data points. Industry tools such as TRIZ and ANYLOGIC are commonly used to identify patterns and determine optimal interactions from one group or system with another. They also highlight critical path areas caused by any number of factors such as supply chain bottlenecks, limited use of shared resources, or a realignment of priorities.

When seen from close range, flaws and inefficiencies that were once hidden are suddenly exposed like a knitted fabric with a faulty stitch. When changes are implemented, the same ‘systems thinking’ methodology used to unearth the problem in the first place is ‘recycled’ and reassessed using more current data.

Looking Ahead…
If ‘systems thinking’ is something you feel that you have been doing all along but never knew that it had a name, you are not alone. Many professionals unwittingly apply the basic principles of ‘systems thinking’ to improve their time management at work or at home.  However, the ‘systems thinking’ discussed in Definition #2 goes much deeper. It evaluates companies as though they are systems operating within other systems and applies innovative methodologies that can spot hard-to-find problems or solutions.

Companies that already subscribe to ‘systems thinking’ ideas designate one person or team to offer company-wide recommendations, but past experiences have shown that a greater emphasis on a participatory effort from a wider range of individual inputs can be more effective, especially when it comes time to implement any changes. As the workplace becomes increasingly automated, an employee’s role will also change and require a better understanding of ‘systems thinking’. Companies would do well to invest in various levels of ‘systems thinking’ training for their entire workforce or hire employees who already have a degree or experience in ‘systems thinking’.

Not everyone needs to receive a graduate degree to get hired, of course, since as shown earlier, the majority of staff can be trained in ‘systems thinking’ at the individual level (see Definition #1). Management or specialize staff members, on the other hand, can opt for degree-level programs that use sophisticated tools to evaluate groups of systems (see Definition #2). Currently the number of institutions offering degrees in ‘systems thinking’ is limited, but as the demand for training is expected to increase in the coming years, many more options will become available.

MIT’s Masters Program
For those of you who are anxious to get started or are looking to realign their MBA degree should consider MIT’s System Design and Management Program (SDM), which offers a Master’s Degree in Engineering and Management.  This degree program is flexible with 13 to 24-month career-compatible options comprised of on-campus and live, synchronous, at-a distance classes.  Students work with their peers on problem sets that in many cases can be immediately applied to their existing companies. Many students who attend the program are sponsored by their company. Of course, you do not need to wait for your employer to sponsor you.  If the timing is right to advance your education, you might do better by taking your own initiative.   Aside from getting a leg up on this new and exciting trend,  you will also have much to gain personally, collectively, and professionally.

###

– Appendix –

Client Case Studies using Systems Thinking to Improve Customer Satisfaction
As stated by one speaker, ‘systems thinking’ is a balance between science and art. To that description, I have arranged the following ‘systems thinking’ case-studies presented at the conference accordingly.

Science – using Machine Data
General Electric
At the event GE’s VP and GM for Technology and Sciences, Gary Mercer, spoke on behalf of GE’s Aviation Division.  Mercer explained how GE views its aircraft as magnets for collecting reams of machine data, to the tune of 18 million date points per month.  This data is used in various capacities to improve on aircraft design, manufacturing processes, and ultimately the passenger’s experience (in that order). Their software and analytics platforms also referred to as SAGE, allows GE’s 3,000 technologists to share the aircraft machine data and publish their ideas on a common platform.

John Deere
Another example came from a member of John Deere’s power systems team, Genevieve Flanagan.  Based on Flanagan’s ‘systems thinking’ analysis, John Deere inserted more probes in their tractors that would collect additional ‘machine data’ on a per customer basis. The ‘machine data’ is transmitted back to a central data bank for analysis and like GE is used to improve the product and customer experience.  Based on a set of algorithms, the data bank also alerts the user when maintenance such as an oil change is needed.  In this manner, the decision to change the oil or any other component is based on a client’s specific usage patterns rather than solely off a generic instrument reading such as the mileage from an odometer.

Art – understanding Client Needs
TIBCO Software
At the event TIBCO Software’s Executive VP of Global Field Operations, Murat Sönmez, cited two examples of how his company’s platform software allowed his clients to analyze their data to identify, design, and deploy ‘system thinking’, solutions dynamically.  His first example involved capturing profile data from Las Vegas gamblers for a client hotel. As soon as a guest would arrive at his client’s hotel, the hotel’s IT system would apply an algorithm that reviewed the guest’s past gambling experiences and established their most likely tolerance level for losing money.  Prior to reaching their threshold of ‘unhappiness’, the guest would receive a text with a special offer, such as a pair of show tickets. Appropriate staff members would be alerted instantly to ensure prompt delivery of the tickets and any other necessary amenities.

His other example involved preemptive measures taken for two major airline clients to minimize a passenger’s negative experiences during a luggage loss claim. Prior to landing, the airline’s systems would apply an algorithm that automatically communicated with a passenger via text and offered them instant remedies such as an approved check-in number at a hotel, a link to enter a delivery address once the luggage arrived, or a credit for purchases at a popular clothing store. In both examples large amounts of data had to be captured, analyzed, and acted upon involving numerous internal departments and partners, so that a one-remedy experience could be delivered to each guest/passenger in a timely manner.

To truly appreciate the value of ‘systems thinking’ from these two examples, imagine each client incident as a series of baton passes in an Olympic track relay.  Then, multiply this one event by thousands of different relays and baton passes where each relay represents one customer who must be treated according to their specific profile that is based upon the results generated from a dynamically, self-adjusting and self-correcting set of algorithms. Having all of these benefits perfectly determined, coordinated and delivered using the same staff as before and with minimal training requirements, if any, is an excellent testimony of the extraordinary capabilities from applying ‘systems thinking’.

CISCO Systems
Another example came from CISCO’s VP of Enterprise Smart Solution Engineering, Phil Sherburne.  Similar to the last example, CISCO learned that their customers had a 6 month threshold/tolerance for adopting new technologies and as a result adjusted both their R&D and engineering implementation processes to follow in tandem .

To continually fill the R&D pipeline with relevant projects, Sherburne used ‘system thinking’ analysis to discover missing links between his sales force and client’s needs.  The results favored a new type of sales force, one that he referred to as ‘honest brokers’ who would focus less on selling products by product line and more on solving client problems that included CISCO products.  Also in the works is a new video conferencing service that Sherburne hopes will further engage their clients into an ongoing solutions oriented conversation.