Life data analysts often face the challenge of zero-time failures, which are defined as failures occurring before a product is in the hands of the end customer or out-of-box failures. Processing zero-time failure data can be a complex process since most statistical distributions and analysis software packages do not handle such data well. The analyst must thoroughly understand the appropriateness and implications of including zero-time failures when analyzing life data. Incorrect usage can result in inaccurate forecast results and potentially negative engineering and financial consequences. Simply plugging numbers into a software program will not always result in satisfactory results if the background engineering analysis is not performed as well. The analyst must determine if the zero-time failure data should be used as presented, transformed according to known field exposure correlations or ignored all together.
