For analytics to become truly effective in global manufacturing settings, veracity and consistency are critical requirements. Based on a recent interview with an executive at a large global apparel company, several insights regarding improvements in global consistency of data became important. Too often, companies seek to collect data for analytics, but do not consider that the data will be used as a tool for investigation into root causes, or as a predictive tool, or as a tool to communicate improvements. Rather, data is collected and viewed as a means for simply capturing what happened, at a given moment in time, and hopefully glean insights into why an event happened. This will never work – particularly as we move into an era of global outsourcing and contract manufacturing in low cost countries such as Vietnam, India, Bangladesh, Malaysia, Indonesia, Cambodia, Sri Lanka, and others in Asia.

Low cost country factory problems should be taken seriously, as they create risk for the brand, impact regulatory changes, and impact union management.  Three key points stand out as consistent reasons for problems in most low wage factories.

  • Union management issues
  • Safety issues are not well understood.
  • Benefits, salaries, and regulatory changes around labor

Driving consistency in analytic meaning is at the forefront of building a more reliable and predictive approach to capture a defined set of factory incident indicators. For instance, a simple incident such as a “strike” has a number of various possible meanings. A strike means something different in China then it means in Vietnam, and the reasons behind the strike will also be different. But to understand this requires deep knowledge of the situation. To truly understanding what is triggering strikes and protests, and what differentiates a “protest” versus a few hours of discussion between management and workers, a higher level of context and understanding must be captured in the analytical data collection mechanisms. The same goes for fatalities and management allegations – it is often difficult to understand the context in these situations without being there. For example, fatalities may include individuals who were killed off-site in a traffic accident on their way to work, something that does not reflect the lack of compliance on the factory’s part. On the other hand, a majority of injuries in factories occur because of violations in Standard Operating Procedures, where a worker violated the SOP or management failed to properly educate workers on the SOP. People were not aware of safety issues, and often there was not a culture of safety in the factory. This is a common element, that needs to be more formally captured in the data.

In this respect, seeking a relationship between audit work violations as a predictor for work disruptions would be a productive avenue for global sourcing managers and analytic exploration. Further indicators of a “culture of safety” include regular safety training for new and existing employees, a general awareness of safety, available training and educational materials, workers all wearing safety equipment voluntarily (not because they are told to!) and an open environment where safety and problems are discussed openly. Emphasizing the basics of safety through simple observations such as whether people walking up and down stairs are holding the handrail is important. Such minor violations are called out by co-workers to remind them instills a culture of safety.

A good example occurred in a low cost country factory. The factory was seeing an increases in safety incidents, and a conversation with the factory management team quickly made it abundantly clear the reason for their struggles. Quite simply, safety was simply not a priority, and the statement made by management was that “injuries just happen, and are part of manufacturing” was used. If this attitude prevails within the management team, then injuries are going to happen! This attitude needs to be reverse:  the goal of factory safety is to NEVER have an injury (vs. injuries are “normal” in this industry).  This will lead to a different mindset, a different culture, and a different set of  approaches that will drive down and in the short term eliminate injuries altogether.

The final element in factory problems, changes in labor regulations, will often result in a disruption, particularly if there is a poor relationship between the union leader and the factory management. Moreover, management must react early to regulatory changes, and must anticipate that the situation will cause a problem. Putting up a poster on the wall, without properly explaining and communicating what the regulatory change means, is not sufficient. If workers are not informed properly, then misunderstanding of policy changes will occur, and if a union is involved, will quickly lead to rumors and a strike. Factory managers need to change with unions much earlier, and not wait for them to react. In other cases, a majority of strikes occurred simply because workers were asking for a wage increase or bonus increase, or there were misunderstandings about bonus policies. Management must have people who are willing to sit down, and spend the time to have a two-way communication channel with workers and unions. This entails answering questions workers may have about how policies will affect them. Simply putting a suggestion box on a wall doesn’t work! Having a two-way communication will reduce the number of most strikes and protests, and there is empirical evidence to support this. In addition, a human resources person who has expertise in industrial relations, who is dedicated to the role and knows how to negotiate with a union representative, is a critical component of factory management.

2 Responses

  1. CleanMark Labels

    August 13, 2017 @ 7:56 pm

    There were some excellent points in here—especially about understanding the context behind the analytics, which should be included. This does not seem to be talked about often.

  2. handfield

    August 18, 2017 @ 2:47 pm

    Thanks for your comment. Unfortunately, not many people take the time to understand the meaning and consistency of data in analytics…

Some HTML is OK

or, reply to this post via trackback.