Driving Product Reliability Through Data-Driven Analysis

Alanna Quimpo
Sep 11
2 min read

CAE Integration can bring GSR to your business and revolutionize your customer's experience.

How this all began for us . . .

There was a 8 channel phone card and the ringer would burn the same spot on every board after 2-3 years in the field. Warranty costs were skyrocketing with no solution in sight. The point of failure was where the ringer circuit pin went into a chip adjacent to a ground. And the spacing of the chip pins violated the creepage and clearance distance of the ringer voltage. When asked about this, the chip maker said it is fine at one point, but the traces must diverge within 25 thou to get to their creepage distances. After analyzing the RMAs in GSR, the one commonality found was that all of the units were coming out of Louisiana and were getting sold into Texas. Louisiana is humid and Texas is dry. So what was happening is that over the course of a couple of years, humidity was soaking into the material and changing the conductivity of the board, hence the minimum creepage distance. The correction was to re-spin the board to fix the layout per manufacturers recommendation. However, what was to be done for the boards already out on the field? The field remedial action was to drill out the burned area of the circuit board and replace the offending trace with a Teflon insulated wire.

What Happened to all those fixed boards?

The boards with the bore holes never came back for servicing. No more warranty costs, more profits coming into the business.

What does this process look like?

With basic data collection in place, the Reliability Improvement Cycle empowers engineering teams to systematically identify, investigate, and resolve product failures using a mix of statistical and scientific tools.

The process begins by categorizing return data (RMAs), which allows for trend identification via pivot tables and custom queries. Engineers then focus on high-impact failure types by:

Analyzing patterns in assembly dates, personnel, and return timing
Reading RMA notes for environmental or use-case clues
Using product life data to identify trends like infant mortality or wear-out
Drilling into cross-product/component dependencies and version histories

From there, a statistically significant sample of failed units is selected for root-cause analysis. Once patterns emerge, miscategorized failures are reclassified, giving precise failure rates and cost impacts.

Corrective actions may include design changes, manufacturing or training updates, or customer usage guidance. Results are tracked over time using tools like GSR’s Plot Part Operating Life, enabling engineers to validate improvements and refine future investigations.

When data is missing, interviews with involved parties help recover insights—and reinforce the value of detailed recordkeeping. As teams grow accustomed to the process, the required data and questions become more intuitive, accelerating product reliability gains.