It is critical that the right data are measured. If this does not occur, any improvement effort would likely fail. The team would ideally measure variables that will have a direct impact on the outputs. In choosing the type of data to collect, one should consider the time/cost of data collection, how those data tie into customer expectations, and the ability to obtain accurate data.
After a decision is made by the team on what to measure, the team also needs to provide an operational definition, and the source of data (eg, LIS, log book), prepare a plan and refine the process as needed. The operational definition is especially important when multiple people are involved in data collection. In the earlier example regarding chemistry stat turnaround time, "received" in the lab could be defined in multiple ways. Some laboratories would define the "received time" as the time the sample physically arrived in the lab (via courier or pneumatic tube) but some might define it as the time the sample is accessioned in the LIS.
The team would then proceed to write up a data collection plan. Ideally, the period of data gathering should be reflective of the normal workflow of the laboratory. A time period that includes major holidays (Thanksgiving, Christmas, and New Year), or is at the peak of an outbreak (eg, H1N1 virus), or during student summer vacation, if this is a CLS teaching laboratory, might not be reflective of the normal operation of the laboratory and will lead to skewed results. The period to measure will vary depending on the project scope and the available resources. Generally, 2-3 weeks of data should be sufficient.