GB560 Unit 5 Failure Mode Review Example

Six Ways the Returns Process Fails, Costed and Ranked: A Unit 5 Failure Mode Review

[Student Name]

Purdue University Global

GB560: Designing, Improving, and Implementing Processes

Unit 5 Failure Mode Review

[Instructor Name]

[Date]

The seller, its returns center and every count and cost are composite and illustrative; the methods and sources cited are real.

Method

Base. 75,400 returns a year, about 1,450 a week.

Frequency. Annual counts from the customer service contact log, the problem cage register, refund correction records, repeat return records, card network chargeback reports and loss prevention case files.

Cost per occurrence. Built up from labor minutes at a loaded $48 an hour, amounts written off, card network fees and repeat shipping, as stated for each mode.

Ranking. By annual cost, frequency times cost per occurrence.

This shares the logic of a classic failure mode and effects analysis, which rates severity, occurrence and detection on ten-point scales and multiplies them into a risk priority number (Stamatis, 2003). Ranking by annual cost instead suits a business process where every failure carries a price.

The Six Failure Modes

1. Shoppers ask where their refund is after day three. 13,874 contacts a year, 18.4 percent of returns. Cost per contact $6.10: about five minutes of an agent's time, $4.00, plus $2.10 of telephony and overhead. Annual cost $84,631.

2. Returns with no usable authorization. 8,143 a year sent to the problem cage, 10.8 percent of returns. Cost $9.40 each: about 9.6 extra handling minutes, $7.68, plus $1.72 of storage and tracking. Annual cost $76,544.

3. Refunds for the wrong amount. 1,583 a year. Cost $31.00 each: 12 minutes to correct, $9.60, an average $18.40 of overpayment not recovered and a $3.00 card fee. Annual cost $49,073.

4. Units graded sellable but defective. 890 a year come back a second time from the next buyer. Cost $44.00 each: $18 of repeat shipping, $10 of handling and $16 of goodwill credit and lost resale value. Annual cost $39,160.

5. Chargebacks filed before a refund posts. 467 a year, when a shopper tired of waiting disputes the charge with the card issuer. Cost $58.00 each: a $25 network fee, 20 minutes of staff time, $16, and $17 of disputes lost on average. Annual cost $27,086.

6. High-value units returned with parts swapped. 83 a year. Cost $310 each, the average value of the missing parts. Annual cost $25,730.

Total: about $302,224 a year.

More than a third of what the process loses each year comes from two failures that are not errors at all, but side effects of making shoppers wait.

Removable or Detectable Only

The ranking says how much each failure costs. It does not say what the redesign can do about it. The final column of the review asks whether a failure can be removed by changing the design, so it cannot happen, or only detected and corrected after it happens.

Removable by design: modes 1, 2 and 5. Refund-status contacts and chargebacks both come from the wait between dock and refund; a faster refund removes their cause. Returns without usable authorization come from the authorization barcode scanning poorly; matching by the carrier tracking number instead would remove most of them.

Detectable only: modes 3, 4 and 6. Wrong amounts, missed defects and swapped parts are errors of judgment or deliberate fraud. A redesign can add checks that catch them, but it cannot make them impossible.

The three removable failures cost about $188,261 a year, 62 percent of the total. That is where the redesign in Unit 7 should aim first. Frei (2006) argues that many service costs are created by the customer's experience of the process, such as anxious calls and disputes, rather than by the work itself, and modes 1 and 5 fit that pattern exactly: they disappear when the shopper no longer has a reason to worry.

Why Keep the Lowest-Ranked Failure

Swapped parts rank last at $25,730 a year, and a mechanical reading of the table would leave them for later. The review keeps them for two reasons. Each case costs $310, far more than any other single failure, and the cases suggest deliberate fraud, which the company treats differently from an honest mistake because unchecked fraud tends to grow once it works. Juran and Godfrey (1999) caution that cost rankings should be read alongside severity, and the design keeps a manual loss prevention check for every unit above $400 for that reason.

Limits

The cost per occurrence for each mode rests on stated assumptions, such as five minutes per contact and $16 of goodwill per repeat defect, and would change with better data. The counts come from a single year. Two failure modes may overlap: some chargebacks follow a refund-status contact that went unanswered, so removing the wait would reduce both together.

What Carries Forward

The redesign should remove the wait that causes modes 1 and 5, replace authorization barcode matching to remove mode 2, and keep or strengthen the checks that catch modes 3, 4 and 6. Unit 7 maps each change to the failure it targets.

References

Frei, F. X. (2006). Breaking the trade-off between efficiency and service. Harvard Business Review, 84(11), 92-101.

Juran, J. M., & Godfrey, A. B. (Eds.). (1999). Juran's quality handbook (5th ed.). McGraw-Hill.

Stamatis, D. H. (2003). Failure mode and effect analysis: FMEA from theory to execution (2nd ed.). ASQ Quality Press.