Safety first: Learning when not to replicate

In a manufacturing environment, communicating bad news and sharing knowledge about specific accidents and incidents can ensure they are not replicated with more sinister results. Stan Kwiecien explains how the application of knowledge-management methodologies can lead to a safer workplace.

Ford Motor Company has a history of sharing and replicating ‘best practices’ among the company’s manufacturing and business-support activities. Beginning with manufacturing operations, naturally occurring communities of practice have used a common process to exchange proven business-process improvements. The best-practice-replication (BPR) process requires that community members collect, vet, re-distribute and provide feedback regarding these business improvements.
As well as describing the genesis of BPR at Ford, this case study outlines how the methodology has been applied to improve worker safety at the company.

There are now nearly 60 BPR communities of practice at Ford, although remarkably only 15 are directly related to manufacturing. These were easily identified and quick to launch, covering obvious areas like metal stamping, power-train operations (casting, machining, engine and transmission assembly, and testing). The process actually began to evolve in the mid-1980s within vehicle operations (body construction, paint and final assembly).

For decades, productivity improvements were shared openly during events or conferences. Typically, industrial engineers or operations managers would get away from their plants for a few days to meet with their peers. Each attendee had to bring information on recently implemented productivity improvements. Handouts ensured attendees could take information away with them and implement similar activities for themselves. We would head home full of good intentions carrying stacks of great ideas. In reality, few of us methodically applied our knowledge. Certainly, there were some ideas that were so simple and clever that they were applied immediately. Those that were less easy to implement were filed under ‘opportunity’.

The vehicle operations general manager at the time saw that there was an abundance of innovation taking place, yet felt frustrated that knowledge sharing regarding these successes was restricted to these conferences and events. He had a vision to rectify the problem and assembled a team to fulfil it. The group included one person from each of Ford’s 18 assembly plants. The team travelled to each of the plants to examine their operations with the aim of learning how they addressed each process. In doing so, the group learnt that not every location performs assembly processes in the same way. All deliver the process intent, but some make it look simple. The team proved that no-one was best at everything and yet everyone was best at something.

Armed with quantitative and qualitative data, and knowledge of how processes can be applied in different environments, the teams continued on their travels. They could now review the operations at a particular plant and suggest improvements based on what they knew worked at another location. We did more than simply highlight opportunities; we stayed long enough to help with the implementation. A matrix was created and kept that identified the process-improvement opportunities, where they had originated and where they had been ‘replicated’. We also captured the value of the replication. This process flourished within vehicle operations stamping, body, paint and final assembly. The size of the team diminished and soon, its sole function was soliciting examples of proven practices, preparing the documentation, mailing or faxing it to the plants and attempting to collect feedback on the implementations. This was indeed an administrative nightmare.

The intranet provided us with a much needed solution. In 1995, this was still a new and evolving technology at Ford. It allowed us to enter all the descriptions into a template, which was displayed as a ‘picture sheet’ that I instantly sent to all plant representatives. A bonus was that the feedback was also collected online. Most importantly, the activity feedback reports were automated and any interested parties could look at them. This closed the loop and made the process fully robust. Best practice replication became Ford’s first web-based transactional database application. It was relatively simple and quick to build since the process, templates and reports were well defined.

The same methodologies would be applied beyond manufacturing. As stated earlier, manufacturing had a history of sharing knowledge and identifying who the people to go to were. This same approach could now be applied to support activities, such as material planning and logistics, quality, human resource, plant engineering, marketing, information technology and product development. The leadership responsible for these activities realised that this process and supporting web-based application would enable remote locations to learn from these improvements. The leaders recognised that they had to set expectations and build relationships where they did not already exist, and that the entire process of sharing and replicating knowledge relied on trust and relationships.

As knowledge management became a recognised discipline that could effectively improve the core competencies of an organisation or enterprise, various companies and academics identified that a well structured BPR process was a powerful element of KM. Following invitations to speak at conferences and the publication of our experiences in books and periodicals, a number of corporations benchmarked the Ford BPR process. Indeed, a few companies chose to license the use of the Ford BPR process rather than start from scratch, including Royal Dutch Shell, Nabisco, Kraft Foods and the US Navy. The license agreement entitled them to learn from the detail and intricacies that make our BPR process flourish.

The health and safety of our employees has always been of paramount importance, especially in the manufacturing arena. Our manufacturing strategic priorities are safety, quality, delivery, cost, morale and environment (SQDCME). Each of our manufacturing locations around the world uses these criteria and associated metrics to make up their balanced scorecards. Note that safety is the top priority.

Good practice in safety is a common topic within each of the manufacturing CoPs. For the most part, these were created in response to specific incidents or accidents that had occurred previously. The safety leadership recognised that a more focused and proactive approach was required to make a significant improvement to Ford’s safety record. The rate of lost time due to accidents was slightly better than the industry average, but still not good enough. Our goal became ‘zero lost time and zero accidents’.

Occupational health and safety (OHS) approached the Best Practice Replication Support Team in 1999 to determine if we could modify the BPR process and system to support and improve communication about significant accidents or near misses. The objective was to increase awareness and therefore prevent replication. Not repeating a bad thing has greater value than replicating a good thing. Nothing is more valuable than life, good health and a safe working environment.
Historically, the Central Station received news of significant incidents by phone or e-mail where staff would forward the call or message to a specific distribution list.

Often there was a time delay, or the initial communication did not fully describe the significance of the event. This information was passed on, but as it was not codified or stored, little actual data was collected, which made it difficult to analyse incidents and identify trends. OHS and the BPR team designed a process that was a derivative of the mature BPR methodology. The focal points were the designated safety engineers at each manufacturing or parts-distribution location. We established criteria to define what constituted a reportable accident or incident. This included any death, whether it was due to an accident or natural causes, any incident resulting in hospitalisation or anything that was a high potential near miss. As we designed the process, it became apparent that sharing knowledge of hazardous conditions that had thankfully not resulted in an accident could prevent a similar event occurring but with more sinister outcomes. The new process would cover the reaction to an accident, take measures to prevent it from recurring and identify potential hazards that could make it happen again.

We called this derivative of BPR the global preliminary incident reporting system, or GPIRS, pronounced ‘jeepers’. The normal replication process involved information being submitted as a draft, the community gatekeeper then reviewing and approving it, and notifications being sent the following day.

For GPIRS it was important to ensure that information flowed immediately. Once information about the incident had been entered, the system would instantly send the notification by e-mail. A template similar to that used by BPR captured information about the incident. The use of drop-down menus here ensure data was collected consistently. Data is of more value than information; data can be translated into knowledge while information is open to interpretation. The automatic e-mail would include all the pertinent information about the incident and contain a link to the BPR system. Knowledge therefore flows via e-mail and is captured in a database.

There was some initial reluctance to openly broadcasting ‘bad’ news. We coached the safety engineers and middle management and convinced them that if we properly reviewed and addressed the bad news from one location it would reduce the likelihood of accidents happening again elsewhere. We were able to raise the awareness of potentially hazardous situations. One benefit that we had not expected was the improvement of safety talks. Each supervisor of hourly workers
has to conduct periodic safety talks. Previously, they were given textbook safety talk sheets. Now they use examples of real incidents or near misses. The workforce can directly relate to the conditions described and the human aspects when they realise that one of their colleagues was involved. Awareness has definitely improved. When I visit a plant, it is comforting to have one of the hourly workers approach me, ask if I am visiting, and advise me to keep within the safe walking zones and adhere to the safety procedures.

With time and improved quality reporting, both in content and relevant criteria, the database was populated with significant statistics that can now be analysed. This analysis has helped the OHS organisation prioritise and address the most frequent and serious types of incidents. We had long known that any interaction between a pedestrian and a forklift truck carrying production materials would rarely benefit the person on foot. Data collection from GPIRS led to a revised policy and strategy regarding forklift and other material-handling vehicle traffic patterns. Each location was given a directive to implement strict guidelines to designate clearly marked areas where people could walk without fear of encountering material-handling vehicles. Barriers prevented such vehicles from entering these areas. Within the production areas, specific walk paths were marked indicating where it would be safest. These areas were marked in yellow since the areas were shared with forklift and burden-carrier traffic. Certain areas, such as the shipping-receiving docks, where painted red, which indicated that a person should not enter these high-traffic areas. The long-term strategy is to completely eliminate the presence of forklifts in the production areas. As assembly areas are designed to support major product launches, the configuration allows materials to be fed from the perimeter, by conveyor systems or wheeled into place using portable racks. Each visitor to any manufacturing location must first view a safety video that explains these concepts in detail before being issued a visitor’s pass.

In addition to the GPIRS notification process, we established a standing meeting every Friday for each division to review the GPIRS incidents for the week. At this meeting, the division-safety engineer invites the appropriate facilities or process engineers to review incidents where engineering action is required to reduce the potential for reoccurrence. Examples include control-panel revisions, conveyor modifications to eliminate pinch points and other areas where the existing design or specifications need improving. When corrective action is identified and verified for effectiveness, the original preliminary-incident report (PIR) is upgraded to an immediate-corrective action (ICA). The BPR system now pushes the ICA to the field-safety engineers (with distribution limited by division or region as applicable). The safety engineer is now responsible for reviewing the ICA with the appropriate engineering, maintenance or operations personnel. They are the ones ultimately responsible for implementing ICA directives. The safety engineer is also responsible for following up and providing feedback regarding compliance to the corrective-action directive.

OHS staff members are also better able to target communication with all safety engineers. They use GPIRS to announce changes in policy or alert them to issues that relate to commonly used equipment and tools. These alerts require the use of the BPR-feedback process to signal compliance or applicability. A prime example of this was an instance of pneumonia caused by Legionnaire’s Disease at one location. Investigation showed that the presence of the legionella bacterium was more widespread than originally thought. Although present at low concentrations in most commercial and domestic hot-water-storage systems, certain conditions could generate high levels and the potential for infection. By understanding this, specific procedures to test, contain and control the legionella bacterium were communicated via GPIRS. It became a requirement to frequently test and use GPIRS to notify employees of any situations where the concentration was above a still safe threshold. Awareness, while avoiding overreaction, resulted in a systematic approach to controlling this potential source of infection.

The pure intent of GPIRS is to notify people with direct responsibility for safety of all high-profile accidents and near misses. The process also includes a follow-up procedure and compliance review.
Every day, incidents that result in a visit to the medical facility are logged in another system, occupational-health-and-safety information management (OHSIM). This process starts with a visit to medical. The attending nurse or physician logs the incident in OHSIM. This in turn triggers the requirement for an accident investigation by the person’s supervisor. Depending on the severity of the injury, various techniques are used to complete the investigation.

One effective toolkit, 8D, was initially designed to resolve quality issues. This is an eight-step disciplined approach (hence 8D) to problem solving. The process is used following major accidents for complete problem identification, root-cause analysis, resolution and confirmation. The steps involved are:

D1 – Form a cross-functional team of problem solvers;
D2 – Define and fully describe the problem;
D3 – Identify, verify, implement and validate interim-containment actions;
D4 – Identify and verify the root-cause and escape point;
D5 – Identify and verify permanent corrective actions;
D6 – Choose, implement and validate permanent corrective actions;
D7 – Perform meaningful activities to prevent recurrence;
D8 – Provide recognition to team members.

This 8D process is also a web-based application. Each accident investigation is assigned a tracking number and linked back to OHSIM. However, OHSIM only captures accidents and not near misses.
We plan to integrate GPIRS and OHSIM to better contain high-potential incidents before they become accidents. The GPIRS process aims to transfer knowledge. OHSIM and the related problem-resolution techniques intend to gain a true understanding of health-and-safety issues. This understanding then becomes ingrained in our policies, strategies, standards and everyday work processes.

The emerging use of a Six Sigma process – define, measure, analyse, implement and control (DMAIC) – is also applied when a pattern of incidents and accidents indicates that the issue is not clearly understood. A Six Sigma project often takes weeks to complete, but by the end of it, no stone has been left unturned to ensure all possible solutions have been investigated.
The key points of the approaches discussed above are:

1. Openly acknowledge that a problem exists;
2. Communicate with the appropriate people, push information to them and don’t expect people to search;
3. Collect data, analyse and then use it;
4. Use common, well-defined processes and instil a discipline to use them;
5. Include cross-functional representatives when problem solving;
a. No one person is an expert at everything;
b. People can get too close to their function and may not see the simplest solution;
6. Use the data and the resulting knowledge to determine improvement actions;
7. Prove that the solution is effective before rolling it out;
8. Focus on the people who have the responsibility over implementation;
9. Expect results that improve the business, in this case the safety of the workforce.

This said, how do we know that improvements are being made? We use the following metric to indicate one measure of safety performance. It is recognised by health-and-safety professionals and is expressed in terms of the number of accidents that result in lost time from work per 200,000 man hours. Bear in mind that Ford has manufacturing facilities in 25 countries on six continents. Our manufacturing workforce is about 80 per cent of the approximately 340,000 people employed at the company. Ford produces passenger cars, trucks, engines, transmissions, castings and forgings, and metal stampings of all kinds at its 110 wholly owned, equity-owned and joint-venture plants. Also consider that a large plant may have a population in excess of 5,000 people who in some cases spend half their waking hours at work. The manufacturing environment can be hazardous. Safety must be the top priority.

Although we not yet achieved our target of zero accidents and zero fatalities, the improvement shown above was not a chance result. Data for 2003 shows continued improvement. Increased awareness through the use of knowledge-management applications like GPIRS, and the processes and toolkits provided by OHSIM, 8D and Six Sigma have all worked together to yield results. There is not one solution, but a portfolio that must operate in harmony to achieve our goals. The knowledge gained from past accidents and incidents has been widely distributed and implemented to create better processes and safer facilities. Accidents must not be replicated they must be prevented.

Stan Kwiecien is best practice replication deployment manager at Ford Motor Company. He can be contacted at skwiecie@ford.com