What if the data are wrong?

ask jeff shiver what if the data is wrongQuestion: I’m an engineer, and Deming said it best: "In God we trust; all others bring data." I want to drive decisions based on data. It doesn’t seem that the team has been acting on data, and the maintenance group is working on whatever they think is important. Data makes it easy to make informed decisions; how can we argue with it? I’m learning that our problem is that we don’t have a sufficient system for producing that data. We use SAP for our CMMS, but culturally we don’t get good information in breakdown work orders to be able to reach any conclusions. I know we can work on training and getting good information in the system, but plant leadership wants results, and working that angle is going to require time for training and time for us to have enough information to discern what the problems are. The other system we use requires adjustments to be made to it for the data to be valuable. I’ll give you an example: I’m looking at a group of 10 pieces of equipment, and the lowest machine availability is 70%. As I dig deeper as to why it is 70%, it is because of how the data was logged. The machine wasn't shut down properly – the operator shut down the outfeed conveyor, resulting in it being logged as a jam instead of as no scheduled production. It makes it tough to analyze the data.

It’s a wonder this place is running like it is. I’m struggling to gather data that means anything. What suggestions would you offer for prioritizing our efforts? Do we prioritize by volume? We have eight packaging lines, and four days a week I have a team of 15 who are focused on improving uptime for specific equipment on the lines, but how should we prioritize their efforts? These four days a week are days where the lines are running. The other three days, one of the lines is down for maintenance while we perform PMs and corrective work on that line.

Also, what suggestions do you have on setting direction for percentage of time spent on PM optimization, building BOMs, spending time on the line with operators, watching the equipment run, and operator training, given that we have two days a week where the lines are running and this is our time to make these improvements?

Your time and wisdom are appreciated.

Brian, maintenance process leader, TX

Answer:

Hi Brian,

I can’t agree with you more regarding data if it’s good and not misleading data. Therein lies part of the problem: You need time to immerse yourself in the data to understand what is occurring. I know you recognize that downtime reporting systems can leave much to be desired, especially when they rely on the human component. Add to that the choices people make to disguise realities and assign blame where it doesn't belong by listing erroneous failure codes or "cooking" the reports. At one site I was recently at, the production management did just that and gamed the system by taking lines “out of production” to hide downtime. At that site, the downtime could have resulted from late packaging material deliveries, a lack of warehouse space to place finished product, or operations staffing shortages, as examples.

As a process leader, you are challenged to straddle the fence, with one leg on the side of the plant leadership, which wants production results today. It’s how you both are measured, truth be told. Because of the demands of the here and now, we often forget that we must have a strategy to go the distance. What does tomorrow look like? Like more of today? That is where your second leg falls – on the side of building the future state. You must deliver results in the short term and in the longer term.

I’m reminded of a post I saw on social media recently. It's a video of an artist drawing the same three images in different amounts of time. In the first image, the artist can take one hour; in the second, the artist has 10 minutes; and in the third, the artist has 10 seconds to draw the image.  Naturally, the image crafted in one hour is way more detailed that the other two. Time and focus improved the detail and contrast. It’s not different for you than it is for the artist. What will be your legacy?

As you mentioned, it starts with education. How do you train people to achieve desired results, such as avoiding incorrect downtime reporting? In my experience, it always comes back to leveraging training and accountability to get results and solve problems. Start by training the operators on the issue you have identified. Hold them accountable for the reporting. I know this can be easier said than done, especially when it requires buy-in on the production leadership side.

With prioritizing your efforts, I like to focus first on where you are bleeding. Don’t try to improve all lines at once. Learn from one and transfer the knowledge to the others. Aim for precision on both the operations and maintenance sides. Add metrics that highlight what you are trying to achieve. Figure out a way to elevate the metrics visibility to that one line. What gets measured gets done.

When you identify the specific line you'll target first, consider putting representatives from production and maintenance together to conduct a line-level FMEA to address all issues, not just the asset reliability issues. Consider packaging materials, staffing availability, warehousing, and so on. Then chart the results into a four-quad grid that has a difficulty axis and a value axis. From the line-level FMEA and matrix, build a strategy and start the execution. Keep the end game in mind. Remember that reliability is not a maintenance thing.

What’s keeping you from leveraging operators to improve your reliability? If you are already doing this, what were the challenges you faced to make the effort a success?

Email me, and I will respond or place your questions with my answers here on "Ask Jeff."

Talk soon,
Jeff Shiver, CMRP