comment 0

Knowledge Management Fails

Why doesn’t more information lead to better decisions, reduced redundancy of content creation and more collaboration?

The Silo

• Data for one business division or project is in one PDM system, while data for other projects is in different PDM systems. And good, universal information on company policies or lessons learned from tests are unavailable to those on other PDM systems.
• Strict segregation of data by job category can limit the ability of others to use it. “That’s all in the reliability group’s database.” And only reliability engineers have access to the data, though design engineering or quality may need access to it to do a better job.
• Each business division does its own thing, limiting access to information that is universal to only their group, resulting in duplication of content across divisions.

Do Another Study

• When someone is faced with an old problem, the common answer is, “We’ll put together a commission to study it.” The fact that there were other studies that outline solutions is considered irrelevant, leading to additional, wasted studies instead of implemented solutions.
• Assumptions that newer is better leads to rejection of prior studies on the topic as being valid today, so you get redundant studies that give similar answers.
• The tendency to go in with a goal in mind and perform studies until you get data supporting the conclusion leads to excessive data analysis versus actual results.
• When people don’t know what to do, they often state that it needs more study, and then hyper-analyze the data, hoping that they will get the wisdom to know what to do from the data – instead of a decision maker making the decision.
• Data that has been analyzed is often so tightly associated to the original project or product that information about how a specific engine performed or material did in quality testing is missed when others want to know its specs.

All of these situations result in an increase in the amount of data to be managed without improving the quality of information available to make decisions by.

comment 0

Hanger Queens and Kings and Information Technology – Updated

When I worked in productionization, that step in the product development process when you figure out how to build cheap production versions of a costly prototype, we often accumulated a storage area filled with units called “hanger queens”. When your prototypes cost hundreds of thousands of dollars, you don’t want to get rid of them. Their sheer cost and constant presence is why they were called “queens”.
There were several options for these units once production was in full swing. One option was to decommission them and destroy them, since they were obsolete. Another option was to try to find out why they failed. This consumed engineering time and testing resources that could often be better used to support production. And with each iteration of the production design, the hanger queens became more obsolete.
A third choice was to use them as experimental units for repair processes, such as a new way to remove ball grid arrays or patch printed wiring boards. This worked until production methods were no longer close enough to make it a valid source of test data; for example, if you switched to a different adhesive for some components or lead-free solder, the old hanger queens were no longer valid as a way to test your new thermal profile for removing chips.
The fourth choice was to let repair technicians use them as practice, with the spare hope that the unit would become usable. In some cases, engineering couldn’t be persuaded to do anything except test, attempt to modify and then test some more.
Decisions on what to do with each unit didn’t resolve the underlying problems that caused a “hanger” or storage area full of “hanger queens”. Why did we end up with so many hanger queens? Why weren’t engineers willing to let them go? And how does this apply to the IT realm?

Problem: Cost to Keep Troubleshooting

Hanger queens consumed engineering time, test labor and time on test sets. Yet there was always someone with a new idea to try to get one working. The objective was to decipher the root cause of a problem. Yet the problem might be unique to that unit and not relevant to the production design. The reason hanger queens continued to hang around was because there was no cut off point in the development process.
There wasn’t a strict cut off that said after a certain degree of testing and troubleshooting, we will stop and support production. There was no cost-benefit analysis as to where engineering and test personnel were better utilized, so the programs often defaulted to the engineers’ wishes.

Solution: Cost-Benefit Analysis

To avoid this problem, know the value of your people and their efforts. Determine when development is over and when you will focus your energies on production to maximize the quality of the final product. Know when to stop the development and focus on bug-fixes and production support. You should also determine what types of problems are not worth solving so that you can focus on those that are.

Problem: When Do You Quit?

There were initially no set criteria for determining when the effort is futile. Testing and troubleshooting continued on units that eventually became scrap, becoming wasted sunk costs.

Solution: Decide When Enough Is Enough

This was curtailed when management said that a unit could pass through the repair shop X number of times. If it didn’t work after X repair and test cycles, it was scrap to be sent for operator training or recycling. This lowered the repair shop’s costs, since they no longer spent operator time and spare parts on units that were later tossed. Know when to give up and walk away so that you focus your efforts on more profitable endeavors.

Problem: Continually Sinking Costs

When expensive prototypes exist, they represent a significant capital cost. Maintaining them becomes less important when production comes online. Yet engineering and IT talent can become consumed in supporting the familiar processes and products over the cheaper, simpler versions. The reason these nearly useless projects exist is simply familiarity and a dollop of pride. Engineers need to avoid emotional investments in projects.

Solution: Know When the Ship Is Sunk.

Adoption of pet projects for enterprise wide implementation or emotional investments by key personnel in a project leads to less than ideal asset and labor allocations. Take the emotion out of the decisions about which projects to pursue and problems to solve. Set a dollar limit on all ongoing projects that will kill it, regardless of whether it is the first attempt to salvage it or the tenth. Regardless of how much the hanger queen to cost, and how much more you poured into it in the form of sunk costs, you need to know when the queen has sailed and sunk so that you stop pouring money into a lost cause – freeing it up for the future.

comment 0

The Future of Industrial Engineering

A December, 2015 XKCD comic poked fun at our future not looking like that promised in comics, movies and books over the past few decades. Where’s my flying car? Where’s my hover board? Then there’s the irony that we have an internet with computers connecting everyone to masses of information but not a central computer running everything, though “the cloud” is evolving to shift us back toward that paradigm.

One facet of the sci-fi world of tomorrow that is arriving today is the use of drones as on demand robotic delivery. I was in an IIE holiday party when the topic came up. Someone wondered if this would make many engineers obsolete.

Actually, it won’t. A fair number of industrial engineers work for firms like UPS, Fedex and trucking companies scheduling logistics and planning routes. Drones simply add a third dimensions to the logistical challenge. Adding on demand delivery using drones, as Amazon has proposed, creates a brand new opportunity for industrial engineers already working on operations research, scheduling and optimization.

Will drones replace traditional delivery methods? In a word, no. Drones don’t replace trucks for delivering your new dishwasher, carrying cars to the dealership to be sold, carry oil by the barrel to refineries and power plans. Drones may allow for the rapid delivery of items already manufactured to the destination, but it doesn’t replace in the in place infrastructure for delivering raw materials en masse.

The same is true for 3D printing. While 3D printing lets you make things out of mostly plastic feedstock, 3D printing doesn’t eliminate the infrastructure required for delivering fuel to power plants, carrying people to their destinations, transporting materials to factories that make the plastic feedstock that gets used in 3D printing. 3D printing’s greatest immediate impact on industrial engineering is the growth in traditional job-shop type planning and management, as well as demand for producibility and ergonomic design, for any printing shop more than one person in a basement making one-off items.

What do I see in the future, beginning with 2016, as new technologies continue to evolve? New demand for old skill sets, without making the still common industrial engineering jobs obsolete. These new technologies actually increase the demand for the expertise of industrial engineers.

Making the Connection

I was showing the James Burke “Connections” TV show to my son when he asked what was the point. Was it the history of technology? Was it an explanation of how something he talked about in science worked?

The overarching lesson from the show actually is: the big advancements in science have historically happened when, to quote John Stossel, ideas have sex.

Two people in unrelated areas are drinking together and talk and ideas come together. Or the student of one talks to an expert in a different field, and an existing product gets put to a new use. Or someone sees someone in one field sees the mechanical process similarities and says you could use this process we use for your product, too.

Those who spend years doing detailed trials with plenty of error do make progress in science. The light bulb is one proof of this, and the herbal medicines that were adopted by pharmacology are another. However, many of the big advances are cross-discipline transfers of known applications and solutions to new uses or someone discussing a discovery and others saying “we could do this with it”.

How can we use these lessons today?

• Your greatest advances are likely to come from cross-discipline idea exchanges, not the continued study of existing field experts.
• Many of the suggestions to make things better, faster or easier will likely come from those who apply related expertise to your process.
• The grind of research isn’t to be ignored or neglected, because it produces the sheer volume of data needed for others to know what works and doesn’t work.
• Don’t neglect the mistakes and accidental byproducts – those may be the new materials that are more useful than the products you are making incrementally better via slow trial and error.

• Punishing all mistakes as failures instead of learning from them or trying to find ways to use what you discovered will hurt in the long run. It will also hinder exploration that leads to new discoveries.

Discovery and invention requires the free exchange of ideas and learning from everything that happens, good or bad. If only many managers and team leaders did, instead of siloing information based on job type or ignoring input from those who don’t have as many credentials.