Understanding why results are as good or bad as they are ('2Cy'), may:

• be the first step towards improvement or optimization,
  
• provide evidence that target limits are too stringent for that analyte and measurement method,
  
• lead to tighter confidence/prediction/tolerance intervals and lower detection limits, hence a wider scope for specific hardware in food safety, pharmaceutical analysis, anti-doping research, environmental analysis, forensic science, metrology, &c.,
  
• enable one to operate closer to the target value in process analysis,
  
• reduce development times, because various aspects such as data pre-treatment, sample selection, factor selection and outlier detection can be consistently handled, and
  
• help to develop self-calibrating instruments.
  

It stands to reason that these benefits add value to an instrument.

One may also find inspiration from the following authoritative editorial (dowload, =81 kB):