BLOG
BLOG
Diana Fierro is a customer success manager with Drishti. She is responsible for regional operations in Mexico and is based in Jalisco.
Manufacturing is time sensitive — pressing deadlines, forecasts and scheduling requirements are always at the fore of daily business decisions. Mature lean manufacturers push to achieve efficiency levels that may seem trivial to the outside observer, but shaving seconds off of cycle times equals big savings in the long run. Efforts made to reduce time to assembly are part of the job of engineers on the line. These initiatives are important as continuous improvement in business means striving for perfection, knowing full well that perfection is impossible. It is essential to set realistic goals based on clear expectations. If your goals aren’t clear, followed and adjusted based on performance it can be disheartening for the people doing the work.
When we forecast times for a new or changed process, we usually start off with a hunch based on known variables. Then we go through a period where we measure performance and use those findings to adjust or maintain expected cycle times. From benchmarking, a clear expectation is created whereby we can tell whether or not assembly times are within acceptable tolerances on the production line and troubleshoot additional problems when they are not.
Five workers were set at identical workstations and tasked with producing 10 of the same basic patch cables. The instructions were clear, and the workers were all trained and had similar levels of experience. The materials were provided and they set out to work. The prescribed time to complete the 10 patch cables was 40 minutes (four minutes total time per unit, accounting for batching, epoxy mixing time, etc.). These were new cables, 99% similar to other products, with the exception of a different end fitting.
Four of the workers took approximately 40 minutes to complete the tasks at hand. The fifth worker took 25 minutes to complete the cables. All 50 of the cables passed their final inspection and were found to be within specifications.
When this happened, the line manager reported these findings to the engineering department. A curious thing happened. Instead of attempting to discover why the one worker had been so fast, they attempted to figure out why the other technicians had been “so slow” despite the fact that the prescribed time was achieved, even though the workers technically all built their cables within or below the expected time. The assumption was that something was wrong with the four workers who produced the cables in forty minutes rather than examining the positive outlier critically. The four technicians with the higher times were visibly demotivated, as well; after all, they were now being asked what made them "so slow.”
But just because 25 minutes was possible does not mean that the new prescribed time should be moved. In fact, if the prescribed time were officially moved, the facility would have seen a sudden spike in high assembly time issues, and engineers would potentially have been running down problems that don’t actually exist — a huge waste of time. If everything is over your tolerance in a given data set, it’s possible that your tolerance is set in the wrong place and that benchmarking needs to be reviewed.
Part of what Drishti does is allow for positive and negative deviations to be discovered. Often in manufacturing, we are so focused on fixing or preventing what is wrong that capturing value is overlooked. But when a positive value is highlighted (like the fifth worker taking 25 minutes), we still need to approach it in a methodical manner.
The line worker that achieved the short assembly time was impressive, but not necessarily representative of the production line. The worker represented, perhaps, what “could” be possible for the line but not necessarily what “should” be expected.
The correct action for the organization in the story would have been to critically examine the assembly process of the “25 minute” technician to see if the methodology could be captured that allowed for such a quick assembly time. Once that was established — and found within standards — then the engineers could train the other workers on the line and gradually move the prescribed time based off of new benchmarking.
Drishti does a great job benchmarking assembly operations. With our patented action recognition technology, we are the first solution that can measure the efficacy of physical processes at scale. Even with Drishti, though, setting realistic targets is firmly in the hands of your subject matter experts. The human owns the PDCA process. If your staff is trained, materials are available and the process is firmly established, but you still have an unusually high percentage of operations above your expected time, it may be time to reevaluate.
If you want to know more about how Drishti can help with benchmarking or measuring processes, check out our solutions page.
