Sampling

A common quality control tactic for business operations involves intensive review of a portion of the work completed.

For example, a manager may review one of every 100 resolved customer service requests, to assess what types of requests customers make, how accurately they are processed, whether the staff follows guidelines and procedures, and so on.

To support such reviews, PRPC supports statistical sampling of work items that reach a certain point in a flow. Using the standard utility activity Work-.RandomSelection with appropriate parameters, you can route a portion (such as 2 percent) of the work items that reach that shape to a new assignment, appearing on a supervisor worklist or a quality control workbasket.

Random, impartial selection makes the results of the quality control review statistically valid (depending on sample size).

In other situations, quality control review can focus on work for certain important customers, complex or challenging work types, work items that involve high risk or high money amounts, or other non-random criteria.

The Work-.RandomSelection activity can also support other functions. For example, your flow can send half of arriving work to one team and half to another, selected randomly.

Definitions