Benefits Administration Testing: Part 1
A Systematic, Complete and Efficient Approach
Posted on 1/11/2007 (originally published in print on 8/1/2004)
by Marc Heinz
Introduction
The purpose of this two-part series is to describe a comprehensive and reliable testing method for both new and significantly reconfigured benefits administration systems. The desire to dig deeply into this subject stems from a need to ensure a production system that delivers predictable eligibility and event processing results in any scenario that may occur in the database. This goal may seem either impossible or extremely difficult considering the complexity of the system and the seemingly endless ways an employee’s data may change. When building a test plan, it may seem as if the only practical approach is to gather the team members for a protracted brainstorming session where the group tries to conjure up every far-out, yet plausible, scenario. Usually this brainstorming approach yields a list of test conditions that seems to gain validity in a simple yet direct relationship to mere length. The sense of validity is false because while a team brainstorm or other similarly unstructured approach may cover many or perhaps even most processing scenarios, some will certainly be overlooked. Untested scenarios inevitably take place as a production system matures resulting in a Ben Admin module that may seem, at times, to have a mind of its own. Processing results such as mysterious plan terminations, inappropriate option presentation, and other undesirable outcomes may gradually surface. Sometimes unexpected or unusual processing outcomes are relatively benign and have little impact to either employees or data quality. Too often though, the impact is costly in terms of both the resources necessary to correct the configuration and the effort necessary to restore confidence in the system.
I am sure you get it. The idea is simple. Either test the system fully or pay the price later. Describing the nasty consequences of incomplete testing and preaching about quality is easy, obvious and trite. Fortunately, what follows is much more useful. As above, the overall goal of this two-part series is to describe a comprehensive and reliable testing method for benefits administration. The two sections will include, by section:
Part 1:
Introduction
To continue reading this article you must have a current VP1 Subscription.
