Our Sites

The costs of inefficiency

Addressing efficiency and the cost of design change

Discovering inefficiencies upfront in a robotic welding workcell such as this can prevent hidden costs later on.

U.S. manufacturers can compete in the world marketplace if we simply learn a few important lessons. Chief among these lessons is efficiency and knowing the cost of change.

For too many years the U.S. built monolithic manufacturing factories while the rest of the world's factories remained lean and mean, often out of necessity because of lack of capital or labor infrastructure. Many of what used to be model manufacturers have realized that they were far too bloated and inefficient to compete on an international level, and the result has been the loss of thousands of jobs.

A recent consulting project exposed a common example of inefficiency in U.S. manufacturing. A study of the welding and fabricating processes of a heavy-equipment manufacturer concluded that the parts were being welded as many as three or four times in the same place.

First, the assemblies were tacked together in a tack fixture. Then many of the welds were prepped before robotic welding to prevent the weld from dripping off or burning through. Some of these weld preps were in the same place as the previous tack welds, causing redundant work. The robot then welded the part, sometimes over the same tacked or prepped area (more redundancy). Finally, a finish welder wrapped the corners, performed other tie-ins, and made welds that the robot couldn't make reliably, often in the same places that had already been welded two or three times.

The processes were studied thoroughly with the goal of eliminating redundant activities and balancing the amount of work between human and machine; specifically, balancing the prewelding and tacking, the robotic welding cycle, and the manual finish welding operations. The company was able to reduce the required labor in the department significantly and reassign the personnel to other jobs.

There were other opportunities for savings at this facility. One simple example was a heavy gusset (about an inch thick) that had a large hole machined in it before welding. After this component was welded to the main assembly, the assembly was taken to the machining center for more drilling and reaming. The company discovered it was possible to put this heavy gusset on the assembly without a hole in it and make the holes in the machining center when the main assembly was machined, thus eliminating one handling of the part.

Because costs such as these are not itemized or categorized, they are absorbed in the overall cost of manufacturing and remain hidden in the cost of doing business. They are accommodated in labor standards, and everyone just comes to expect that this is how long it takes to build this widget. However, these costs of redundancy and inefficiency are real, and they can amount to some surprisingly huge dollar values.

Breaking Away From Routine

It is not always easy to uncover inefficiencies because a number of dynamics are at work, discouraging their detection. The attitude that "this is the way it's always been done" gives the assumption that years of history must mean that the process is being performed in the best way. Also, because various processes are performed by different departments or individuals, it is not so evident that redundant work is being done. In many cases, people simply are too busy to think about diving into any project in enough detail to uncover leaks in the dam.

To expose these easily hidden costs, it is necessary to understand fully your manufacturing processes in intimate detail. Some good time study software packages are available for PCs and PDAs that help give a clear picture of where exactly your operators are spending their time. This tool should not be looked upon as a method of keeping tabs on the workers, but instead as a tool to find out where every minute of every hour is going. This kind of inquiry can uncover those precious minutes or seconds you need to remain competitive.

Challenge yourself right now to just sit and stare at the wall for one minute. You'll see that a minute is a long time, and much could be accomplished on your factory floor in one minute. If your industrial engineering efforts can uncover one minute or 10 minutes or 100 minutes of redundant work per part, the time savings could be quite significant.

For example, an operator making small components could use that minute to check the piece in a jig to verify that its tolerances are fit for consumption by a robot. Or the process itself may be foolproofed by adding one step that adds only a few seconds to the cycle time. Don't always be dead set on minimizing cycle times at any cost, because that's not necessarily where you're wasting your money. Significant savings are available in catching inefficiencies early in the process.

The Cost of Change

A common dynamic at work in manufacturing has to do with the cost of change, and some see it as a fundamental law at work, like the law of gravity. The law of cost of change goes something like this, although the actual numbers may vary.

Assume a change must be made to an existing product, including engineering, drawing, and design changes. To change a design or feature in the drawing or design stage costs, say, $1. The same change made in the prototype stage costs $10.The same change made after the product is in production costs $100.The same change made in the field after the product has shipped costs $1,000 (some sources say as high as $10,000). This rate of cost increase is exponential, not linear. In other words, it makes a lot of sense to catch these flaws early in the process.

Now apply this law to the hidden costs of inefficiency discussed previously. Rather than applying the law to a product cycle beginning with design and following all the way through to the customer, consider that this same law of the cost of change also dictates the cost of changing a component part or product as it travels through your plant during the manufacturing process.

To identify the redundant, non-value-added work that is being invested in a part may require an upfront investment of time and money, but the results can be significant. If a small investment in some industrial engineering studies reveals that two minutes per part can be eliminated, the savings amounts to two minutes per part made over the product's life cycle. You may manufacture thousands of parts over a period of years, each part taking two minutes less labor than required before the study was done. Now you can see the exponential return on such an investment of time.

For example, a component part that is inconsistent or out of tolerance causes many problems for automated and robotic welding processes, as well as for assembly processes. The further along in the manufacturing process the change is made, the more expensive it becomes to make the change. To simply make better parts may seem more expensive, but fixing the problems caused by poor part design as the part travels through the production process is more expensive than making the part right the first time.

Let's apply the law of cost of change using some specific dollar values. The amount of time needed to make a change before the part goes into production includes attending a meeting to discuss the change; making changes to the part print; and changing the program that runs the machine that makes the part. For three hours of work, figure $180.

To make the same change after the component part is in production will cost exponentially more. Several people, several hours of work, rerouting parts, and carrying old inventory would come to about $2,000, not including the actual cost of the obsolete part, which could add thousands of dollars more.

To correct poor tolerances after the component gets welded to the main assembly and into the customer's hands will easily cost 100 times the cost of catching the unacceptable tolerances at the earliest stage. Figure you would have to dispatch service personnel ($3,000), pay airfare and travel expenses ($2,000), order replacement parts in small lots at a premium price ($2,000), revise drawings and procedures ($1,000), and reprogram CNC machines or robots ($2,000). This adds up to $10,000, at the minimum.

The Real Price of Quality

Although many books, seminars, and articles have stated this, what many manufacturers don't seem to realize is that quality does not cost money; it saves money in the long run. Poor component part quality probably is the most common reason for poor performance of a robotic welding cell or downright failure of a robotic welding installation. High repeatability and proper fit-up of component parts are the primary keys to success in robotic welding, followed closely by part design, fixture design and function, and personnel skills and training.

Addressing each of these issues upfront and in detail may seem like it would cost your company a lot of money, but this is only because these activities, such as training, cost-saving teams, meetings, and quality programs, are highly visible and typically budgeted for and closely scrutinized. The fact is that these foundational activities can prevent more expensive costs of dealing with compromise later on. If they are not brought to light and itemized, the costs of poor quality may remain hidden in daily procedures and overhead, looming like an iceberg under the water.

Efficiency is about measuring twice and cutting once. Nothing is taken for granted. No process is assumed to be the ideal process until it's examined under the microscope. Everyone must be involved in optimizing processes, especially the shop floor personnel, who are the ones that really know how the machines operate and how the parts should fit together.

Change is not the enemy. Today's manufacturers must be wired to make changes quickly and efficiently. Efficiency is borne out of effective changes made early in the process. Invest the time and money upfront, and you will see that those exponentially rising costs can be reduced as quality rises and throughput increases.

Jim Berge is the owner of Berge Robotics.