Dantar Oosterwal, author of The Lean Machine and Partner at Milwaukee Consulting Group, looks back at his experience at global heavy-weight motorcycle manufacturer Harley-Davidson and explains how the company transformed its product development system without a burning platform to start from.
Many have heard the story of how in 1981 thirteen executives bought The Harley-Davidson Motorcycle Company from AMF and brought it back from the precipice of bankruptcy.
Yet very few people know the more recent story of the silent transformation that occurred in the company’s product development department that powered top-line growth and profitability through similarly drastic improvements and the adoption of lean product development techniques.
Harley’s transformation in the 1980s played out mostly in the public purview. By the late 1970s the company’s short-term profit potential had been exploited by AMF, whose need to sell Harley-Davidson was a direct result of poor decisions made by a management team that was distant in geography and in its understanding of the business. Poor quality and ill-conceived products caused sales to decline sharply.
No one was interested in the depleted hulk of a company. That’s when thirteen Harley-Davidson executives who understood and believed in the motor company leveraged their personal assets and purchased Harley from AMF. The road to recovery, however, turned out to be more tumultuous than anticipated.
Shortly after the purchase, an overall slump in the motorcycle market led to additional production cuts, which in turn resulted in half of the company’s staff being laid off.
By 1985 Harley-Davidson was nearly bankrupt.
CEO Rich Teerlink convinced the banks to accept a restructuring plan that included the application of “Japanese management principles”. These practices of just-in-time (JIT) manufacturing and “autonomation” have evolved into what we now refer to as “lean”.
They had been pioneered by Henry Ford, promoted by people such as W. Edwards Deming and Phil Crosby, and adopted by Toyota, while they were rejected in America.
The application of lean, coupled with the adoption of new management theories and a temporary tariff on large imported motorcycles, created a brief window of opportunity for a turnaround, and for the rebirth of Harley-Davidson.
Rich Teerlink proved to be a visionary leader for the company, and the changes he introduced proved to be very bold. His daring plan worked and in 1987, one year before the five-year temporary import tariff was scheduled to expire, Harley-Davidson announced that it no longer needed tariff protection to compete and petitioned the International Trade Commission for early termination of the deal.
But the tariffs had not been Harley-Davidson’s savior. The business saved itself when it addressed its manufacturing woes, started to create appealing, high-quality products, and revamped its management approach.
Between 1985 and 1987, it increased inventory turns from five to twenty and productivity by 50%, while reducing inventory by 75%, scrap by 68% and required manufacturing space by 25%.
Results of such magnitude have driven the continuous improvement boom and kick-started the lean revolution we see globally today, but they were unheralded in the 1980s. Even now, 25 years later, most of what we perceive as lean remains rooted in the ideas and principles as they are applied to manufacturing. When we look at the adaptation of lean to other industries, such as healthcare, the techniques originate primarily in the theories formulated on the manufacturing floor.
This, however, undermines a trove of opportunity: the impact lean management can have on an organisation is much greater than the simple application of lean manufacturing tools and techniques in a new environment. Nowhere is this more evident then in lean product development.
Harley-Davidson’s transformation in the 1980s was “open to the public” and driven by the need to survive, while the move to lean product development occurred behind closed doors, shielded from public perception, and did not happen for fear of extinction.
Following its rebirth in 1986, Harley had become very successful. From the time it went public through its initial stock offering in 1986 until the 100th anniversary of its existence as a company in 2003, revenue had increased at a compounded annual rate of 17%. Over the same period of time, profits climbed at an astonishing compounded annual rate of 36%.
So why did such a successful organisation decide to lean out its product development system, which yielded over a fourfold increase in development throughput, 50% reduction in time to market, and a quality level in excess of a 98% customer repurchase intent? And why would it wait 25 years after successfully applying lean in manufacturing?
After implementing lean on the shop floor and improving the quality of its products, Harley-Davidson focused its attention on delivering more products, specifically on increasing manufacturing output. The emphasis was on boosting production without increasing investment in plants and facilities. The near-death experience the management team had been through led to a very conservative approach to investing.
By the mid-1990s, there was nearly a two-year wait to receive a new Harley-Davidson motorcycle and the company realised it could no longer achieve the same levels of growth it was experiencing without significant investment in new plants and equipment.
In 1998, a new production facility in Kansas City, Missouri, was inaugurated. By the end of the decade, the motor company had doubled its capacity thanks to new powertrain manufacturing facilities and heavy investment in its final assembly sites in York, Pennsylvania and the new facility in Missouri.
NEW PRODUCTS, NEW CUSTOMERS
There are two main aspects to business: manufacturing customers and manufacturing products (or services) to sell to customers.
As more products were manufactured at Harley, the waiting time to receive a new motorcycle was reduced and the impediment to growth shifted from manufacturing products to manufacturing customers. The main stimulus in the manufacture of customers is creating new products: the need to increase throughput now fell squarely on new product development.
There are four fundamental attributes found in all “exceptional organisations”, which represent the essence of continuous improvement:
- Design and operate systems that highlight problems;
- Quickly solve problems and use them to improve the system;
- Create universal learning and share learning across the organisation;
- Leadership takes responsibility for the systems and development of people.
In manufacturing this is “easy” because you can see the work. When there are moving assembly lines or machines it is possible to see parts flow through the factory. You can physically see the system and readily develop ways to highlight deficiencies, abnormalities, and opportunities. In new product development there is no assembly line or physical part flow to observe.
One of the first revelations in our effort to adopt lean product development principles was learning that the phase-gate process we relied on to manage product development was very ineffective in identifying problems.
With no physical part movement, it is very difficult to see if the “line is down” in a product development environment. The phase-gate process is reportedly utilised in some form by over 80% of companies, and at the time Harley-Davidson was no exception. In fact, Harley-Davidson was considered to be one of the best, having won the Outstanding Corporate Innovator award given by the Product Development Management Association. For us it was difficult to imagine that we had a problem.
In a phase-gate model, specific work is completed in a particular stage of the process. The work is “checked” at the end of the stage in a “gate review” to verify that it has been completed satisfactorily.
The gate review is a sort of oversight and financial control mechanism whereby leadership agrees to additional spending based on satisfactory completion to a particular level dictated by the metrics of the gate. In principle, this is a very logical means of fiduciary responsibility.
Although problems are identified as a part of the gate review process when projects fail to achieve the objectives set out in the gate (check sheet items), it is generally impractical to kill and painful to delay projects. This system is not an effective indicator of product development workflow. Although the gate reviews are intended to provide oversight of the development process, they are poor indicators of success.
In fact, our revelation came by evaluating five years of gate reviews history. It was astonishing to discover there was no correlation between successful phase exits on time and successful projects. Through deeper investigation, however, we learned that there are factors within the development system (although not necessarily a part of the traditional phase-gate process) that correlate with successful projects with an r2 value of 0.98.
These events are referred to as “integration events” and later became known as “pull events”, as a part of actively managing the new product development process. Identifying integration events and actively manipulating them in the form of “pull events” to manage the system allowed the creation of a method to effectively highlight problems.
With a system in place to accurately reflect and manage the actual workings of product development and effectively highlight the conditions of the system, it became possible to create “help chains” to support the management needs of the organisation.
Help chains are a predetermined protocol for escalation and resolution of issues to address the second attribute of exceptional organisations. Through the introduction of visual management techniques (like the Oobeya) it was possible to identify the progress of each project in a portfolio in accordance to the next pull event.
Any deviation from the pull event triggered a predetermined intervention to allow problems to be addressed. Early on, there were certainly many issues uncovered as the immediacy of problems became more apparent. I have learned that there will always be issues associated with product development: the difference between good development organisations and poor ones is the ability to identify issues and address them effectively. As we became able to “see” issues in real time, we could identify the causality of problems; as we addressed them, we applied what we learned to improve the development system.
As we established our ability to see problems and we used the problems we discovered to improve the fundamental development system, we started to address the third attribute of an exceptional organisation: the creation of universal learning and sharing that learning across the business.
While seeking to create learning, we finally recognised the fallacy of traditional point-based development solutions prompted by a traditional phase-gate methodology. Thanks to coaching provide by Dr Allen Ward we came to realise the importance of set-based development techniques. And as we learned, we shifted from a “build&break” development mindset to one in which we embraced development learning cycles and the creation of limit and trade-off curves as a foundation for creating re-usable knowledge.
Although three of the four attributes of exceptional organisations have been described, it is the fourth that is most crucial to success: leadership taking responsibility for the systems and development of people.
Whether a business is successful or not depends on the involvement of leadership: in organisations where leaders are only after results or in those organisations that are willing to undertake the journey but have leaders that are not directly engaged, the efforts fail. However, in those where leaders become personally involved, often even identifying that they may not know the answers but that they are willing to get the coaching support to help them find the way, success is more likely.
THE ROAD TO ARIZONA
A poignant example of an integration event at Harley-Davidson is an early “Pull Event”. One of the more difficult aspects of the development process occurred early on in the projects when it is relatively easy to let development timing slip.
We established a pull event where the entire senior leadership was directly engaged with the development work in an orchestrated event designed to highlight issues. It allowed for open collaboration between leadership and the development team. Rather than sitting in a room to review a check sheet for progress, this pull event was focused on the actual hardware.
The objective was for each development team to have representative vehicles available at our proving grounds for the executives to evaluate and to demonstrate progress.
Due to the distance between the development centre in Milwaukee, Wisconsin and the evaluation site in Phoenix, Arizona, we needed a week to transport the vehicles by truck.
When we loaded the trucks for transport, one of the development teams was not ready. They had recently discovered a new issue with their design that required modification. As a leadership team, we assumed that we would need to cancel the project.
However, when we arrived in Phoenix a week later to conduct the evaluation, the missing project team had arrived with their project ready for review. The team members had rented their own truck in Milwaukee, loaded the parts they had, and driven all day to the manufacturing plant in Kansas City.
There they picked up the remainder of the parts they needed. They built the motorcycle in the back of the truck as they drove to Phoenix.
It was through the dedicated engagement of leadership, the commitment of the development team, and the adoption of lean product development principles – not the re-application of manufacturing tools – that we allowed our product development system to evolve.
For over 30 years, Harley-Davidson has thrived introducing a new model three out of four years. When the focus of the organisation shifted from manufacturing product to manufacturing customers, product development responded: 25 years after the lean manufacturing revolution, Harley’s lean product development revolution produced nearly five times the output in new products than it had for the previous three decades.