Our client continues describing how staying close to operations helped with their problem solving (see part 1):

“When we began our lean journey a few weeks ago, we decided to start on the production floor with a specific value stream. We wanted to see results quickly, so we chose a limited scope for our initial efforts: in-house assembly and test operations.

“We started with 5S and then changed the layout to accommodate our plans for flow and pull. Our supervisors did a great job of implementing a simple visual management system so we all know how the schedule is progressing relative to takt time.

“What we didn’t expect were the parts shortages. Our Materials organization just didn’t seem able to deliver parts to support our production schedule.

“After our Gemba Walk and a review of our observations with the management team, we developed a hypothesis for what was causing our parts shortages: a mismatch between the daily pull process we implemented in assembly and our current MRP-based logistics processes.

“While we managed assembly to a level production rate based on customer demand, delivery of parts to assembly was managed by our MRP system – which in turn was driven by a master production schedule that was different than the level production rate set by Assembly. Also, lead times didn’t match the new lean assembly operation on the floor so the timing of parts delivery did not match what was needed on the floor.

“Our team now faced the question: How do we fix this problem? Many in Materials were confident that the MRP system could be adapted to support lean. They wanted to change the Master Production Schedule to match the level production rate in Assembly, while the lead times in the system could be adjusted to match the new lean processes. However, the outside change agent that we were working with, LEAN Affiliates, recommended a different approach: stop using MRP for scheduling delivery of parts and instead use a pull process across the value stream. They described a methodical approach for “uncoupling” MRP so that the ordering and movement of parts is managed as part of an integral pull process along the entire value stream. This made sense to many of us since it eliminated the use of two parts logistics systems – MRP and a pull process – but the effort of uncoupling MRP represented a radically different approach than what we were doing. We worried about the risk of disrupting production and missing customer shipments if anything went wrong.”

[Editor’s note: This story began with a perceived problem with the capacity of the AGV system, but after observing operations for a few hours this management team has discovered an underlying problem: the use of two fundamentally different logistics systems for signaling the delivery of parts.

Next post: Our client decides how to move forward.]

Before the holidays, we talked about staying close to operations, and one specific habit to help do that, The Gemba Walk.

Since then, a client agreed to share his experience with us:

“My staff was reviewing the performance of an automated guided vehicle (AGV) system we implemented earlier in the year. Although initially it reduced the ‘stock to line” delivery time for parts, during the past few months delivery times were creeping back up again. Some thought we needed to increase the system’s capacity by adding a vehicle. But at the same time we knew that our volumes were steady and couldn’t understand the need for more capacity.

“So our Director of Materials and I decided to observe the delivery of parts to one of our lines during one afternoon. We parked ourselves in two different locations for four hours and watched what happened. At the end of the day we listed our most significant observations:

1. Each AGV trip was handling only a fraction of its capacity of parts. Sometimes we would see an AGV pulling only a small box of hardware (fasteners, clips, brackets, etc.)
2. The production lines the AGV system was supporting was down for over an hour because of a parts shortage.
3. While the line was down due to a parts shortage, the AGV system continued to deliver parts to the line (but apparently not the parts that were ‘short’).
4. A team lead (one of many associates who came to us to ask what we were doing) told us that since they implemented their kanban/pull system within their assembly line, the AGV System was not delivering the proper quantity or type of parts to support their schedule.
5. Several of our Associates approached us and expressed how frustrated they are with the continual parts shortages, and the reluctance of Materials to fix the problem.
   
6. There was more WIP than planned in the AGV drop zones for the line. Each drop zone overflowed with approximately twice the WIP they were designed to store.

“We then reviewed these observations with the plant manager, supervisor and team lead to create a hypothesis that would explain these observations.”

[Per our client’s request, we did not divulge the company name on the blog, but they are open to direct conversations about their experiences. Contact us to arrange this.

Next post: Our client develops a hypothesis that explains their observations.]

I made a day trip to the Assembly Technology Expo in Chicago last week. What an adventure. After all, here’s an event dedicated to the latest automated innovations for manufacturing: Over 600 leading suppliers were there, from AGI Corporation (tooling and automated handling for PCB assembly) to Zierick Manufacturing (interconnection devices). I’m always fascinated with the latest technology for manufacturing, and it was fun to put my business card in all of those fish bowls and hope that I won something.

But as I looked over the Technology Pavilion and saw all of that equipment from all of those large companies, I couldn’t help but wonder…

What if American industry spent just a fraction less on material handling equipment, and just a fraction more on improving their process flow?

What if American industry spent just a fraction less on “Machine Vision Systems” (it looked really cool), and just a fraction more on teaching their people with how to see waste?

What if American industry spent just a fraction less on automated storage and retrieval systems, and just a fraction more on creating level pull and reducing inventories?

I like cool new technology just as much as the next engineer, but I sense that most companies could make much more progress by shelving their next automation project and spending that next dollar on processes, management systems, and developing people.

Manufacturing performance is optimized when management recognizes that there are five golden metrics – five measures of value stream performance that really matter – and all other measurements are subordinate. Subordinate metrics may also have a great deal of importance within the factory, but optimizing performance to them is only valuable if it results in improvement to the five golden metrics.

[Editor’s note: Part 1 of Bill Waddell’s post may be viewed here.]

The big five are: Total Cost, Total Cycle Time, Delivery Performance, Quality and Safety. Activities and efforts in manufacturing that result in improving one or more of these performance measurements, without degrading performance to any of the others, are good performance. Activities and efforts that result in improvements to any other subordinate metric that do not result in improvements to any of these five, are meaningless.

The first and most significant is Total Cost, but not in any traditional sense. First, the only meaningful measurement of total cost is on a cash basis. All money spent on manufacturing must be summarized and the total compared to the previous period – not to a flexible budget or a plan. What matters is whether the total cash spent on manufacturing was more or less than it was in the previous period. It is important that this cost figure is exclusive of all allocations, and that it does not exclude S,G & A costs. The only exceptions are major capital investment spending and adjustments for Accounts Receivable and Payable. While these exceptions must be added back in to create a total lean accounting based income statement, manufacturing performance should be measured as if payment were made at the time materials and services were delivered, and payment was collected at the time finished goods were shipped to an outside customer.

The second metric is Total Cycle Time. This is calculated by studying the major purchased components and determining the total days on hand of each one. The total days on hand is the sum of all of the component in the plant regardless of its form – still in its original purchased state as raw materials, embedded in assemblies or sub-assemblies in a modified state as Work In Process Inventory, or embedded in a finished product. This total days on hand figure is divided by the planned shipments per day for all products that require that purchased component. For example, if there are 5,000 of a component in the plant in all of its various forms, and one each it goes into two final products that are each projected to ship 100 per day, the cycle time for that component in 5,000/200 = 25 days. The Total Cycle Time for the plant or for an individual value stream within the plant is the cycle time of the component with the greatest cycle time.

It is important to note that only “C” type bulk items can be excluded from this calculation, and that the Total Cycle Time is not an average cycle time, nor is it weighted in any way for the cost of the component. This is a measure of operational performance – not a financial metirc.

The third metric is Delivery Performance and it is simply the measure of the percentage of customer orders that shipped when the customer requested them to be shipped. It should not be modified in any way to accommodate company policies or shipping promises. It is purely a metric of manufacturing’s ability to meet customer requests and requirements.

Fourth is quality and this will vary by company, but it must be quality in the eyes of the customer. As a result, customer returns or warranty claims are typically the basis for this metric. It is not a summary of internal quality metrics, such as first pass yield. It is important to realize that internal quality metrics are only important to the extent that they provide information management can use to minimize cost, improve flow and meet customer quality. The only measurements of quality that truly measure operational performance are those from the customer’s perspective.

Fifth is Safety, and the standard metrics of accident/incident frequency and severity are usually sufficient.

These five measurements are manufacturing’s ‘bottom line’. All manufacturing efforts must be aimed at improving one or more of these performance indicators, without degrading performance to any of the others. All other performance metrics are subordinate to one of these, and are useful to management to the extent that they provide information necessary to improve performance to one of the Golden 5.