Fowler Associates Labs

Throughput Analysis:
How Static Impacts the Bottom Line, Part II

Detecting a static-control problem can be only half the battle; you still have to define the problem and sell its solution to management.

Stephen Halperin
President, Stephen Halperin & Associates
Elmhurst, Illinois

Editor's note: This month, Steve Halperin tells us how to determine a static problem's magnitude and location, and even how to present findings to management with a maximum likelihood that your presentation will result in support for static control.

The raw data obtained in the initial phases of throughput analysis is summarized in the device Utilization Data spreadsheets. This collection of information forms the foundation for any detailed analysis of static problems. Further processing of this data is now required before it can be put to effective use.

The analytical portion of our plant study is based on data sorts of all items listed in Table 3. The first sort is based on the material value of the suspected ESD-sensitive device losses and is shown in Table 4.

This is an optional step since the true value of any ESD loss includes burden costs. As an indicator, it's better suited to showing that a problem exists than determining how to attack it. Still, while a single data sort based on material costs isn't recommended as a sole criteria for decision making, it provides an important view of estimated static impact. Some organizations use material costs as an initial indicator when these costs are high compared to related burden costs.

Sort the list in Table 3 from highest to lowest item-cost deviation using column four, "Deviations: Cost X Units," as the primary sort key. Once sorted, classify the results in an "ABC analysis" this way:

Class A: Top 10% of listed items
Class B: Next 30% of listed items
Class C: Final 60% of listed items

The result is shown in Table 4.

Generally, the top 10% of the items listed in an ABC analysis--Class-A losses--accounts for 50 to 80% of dollar losses in ESD-sensitive material. The next 30%--Class B losses--accounts for some 20 to 40% of deviation-material costs. The remaining 60% of items usually accounts for less than 10 to 20% of losses. In our example, the analysis by material costs shows the following percentages.

Class A: top 10% of listed items represents 54.9% of total estimated material losses. Class B: next 30% of listed items represents 34.1% of total estimated material losses. Class C: final 60% of listed items represents 11.0% of total estimated material losses.

A second (and more useful) data sort incorporates both material costs and the related burden expense. Using the "Total Material & Burden" cost per item as the primary sorting reference, sort the entire list from highest to lowest total item cost (burden plus material costs as shown in Table 5).

Table 5 shows our data sorted in this way and includes percentage calculations for both the material cost and burden for each item.

The dramatic impact of burden expense is evident when Tables 4 and 5 are compared. If we classify Table 5 using ABC criteria, our loss distribution now shifts.

Class A: the top 10% of listed items represents 23.6% of total estimated material & burden losses.

Class B: the next 30% of listed items represents 56.8% of total estimated material and burden losses.

Class C: the final 60% of listed items represents 30.5% of total estimated material & burden losses.

The only item in Class A represents the largest single potential loss--$62,750--while the three Class-B items represent a combined loss of $121,860. These four items alone account for almost 70% of static's potential impact on this facility.

We now can gain a concrete idea of static control's potential value to the facility. We have identified the inventory items that are most likely to cause static losses, and we also know roughly where the losses are taking place because we have tracked the items to their requisitioning departments.

Reporting to Management

It is always wise to keep management informed, and in this case, they should have been forewarned of an ongoing analysis and briefed on its methods. Most companies insist on ongoing project status reports. The initial analysis findings could form one of these, to be delivered before beginning the investigation of causes. A useful report would include the following information:

1. Estimated ESD material loss in dollars
2. Estimated burden loss in dollars
3. Estimated combined burden and material loss.
4. A summary of the Class A, B and C loss estimates, including burden and material-cost data.
5. Losses as a percent of ESD-senstitive device and assembly purchases.
6. A list of requisitioning departments experiencing the ESDS-sensitive-device activity and an estimate of the deviation attributable to each.
7. A description of the action to be taken to confirm suspected ESD losses in terms of general procedures, anticipated time and resources required.
8. Comments regarding production-volume objectives vs. volume attained, rework activity and field-service problems that may be reflected in ABC-analysis data.
9. A request for management support during the ESD casual investigation.

Obviously, almost all of this information is available through manipulating the device-usage tables. Further information can be obtained by creating the Purchasing, Inventory, and Deviation Cost Analysis illustrated in Table 6.

The summary of our illustrated data is shown in Table 7 and forms the basis for the interim management report. The most telling fact displayed in Table 7 is that 37% of ESD-sensitive items purchased are unaccounted for, and that of these 1/5 are of a single type of part.
This shows how throughput analysis can help locate trouble spots in an operation. In this case, a close-up inspection of the usage of items one should uncover is one of the plant's major sources of deviations.

For presentation to management, the summary of this information should be objective and concise.

Confirming Losses

Class A and B items represent the smallest item quantities used and yet have the highest potential financial and quality impact on the company.

Therefore, the actual handling of these devices, in addition to inventory factors, packaging and other logistical elements should now be reviewed in detail to confirm the cause of the deviations.

The interim-throughput-analysis report and device-usage spreadsheets point to several interesting areas, aiding the ESD professional in deducing causes and planning corrective action. These examples show a few thing to look for.

Sixty percent of the devices listed in Table 7 have ESD sensitivities of < 2,000 V. Sixty three percent of estimated ESD losses are associated with four items having sensitivities < 1,500 V. Given the foregoing, the company's maximum allowed environmental ESD voltage should probably be reduced significantly.

The rework operation is using 70% of all the ESD-sensitive deviation items. Thus the area of major concern for loss analysis is in or ahead of the rework operation (i.e., in manufacturing or the supply chain).

Approximately 20% of deviation utilization is in field-service. This might support one or more of the following deductions: 1) the product is unsafe for use in customers' ESD environment; 2) latent defects are being engineered during production and testing; and 3) finished-goods packaging doesn't protect the product during shipment or storage.
Obviously a facilities evaluation for static-charge generation is a wise idea. There are several documents, methods and services available to define the cause of static loss (Ref 1).

Communication with Management: The Bottom Line

Management wants the maximum ROI possible at an equal level of quality. Consequently, it is not enough to estimate ESD losses. One must be able to project a reasonable return for the time and money spent solving ESD problems in order to gain management support.

Without knowing the cause of a given problem, nobody can estimate the cost of solving it. However, this writer has never seen static control generate less than a five-fold ROI in one year. In most cases, proper static control yields a much higher ROI, but 5:1 is a minimum. Given that the problem's cause is accurately defined, relatively inexpensive solutions almost always exist.

Taking a conservative stance, if we suppose that no more than 80% of the estimated losses can be eliminated with proper static-control technology, then this is a reasonable projection:

80% of $265,740 = $212,592.

20% of $212,592 = $42,518.

Using these figures, management can justify at least $42,000 to deal with this problem, depending on company policy and attitudes toward quality control. However, astute managers realize that product quality and efficient production always return more than "even money."

Reference
1. Halperin, "Facility Evaluation: Isolating Environmental ESD Problems," EOS/ESD Symposium Proceedings EOS-2, 1980.