Discussion of the steps to be taken

Measure this response as close to the interesting event as possible. 

If possible, consider the error of measurement when the response is determined. If a systematic variation which is induced by changing the experimental conditions is to be detected, it must, at least, be above the noise level of the measurement error. 

a chromatographic or spectroscopic method, the precision of which is generally known, can be used to measure the response. If the precision is not known, it can easily be determined by replicated analyses of known samples. The analytical error is rarely an obstacle in screening experiments. 

Quite often, a lot of time can be saved at this step. A response measured directly on the reaction mixture after the critical step may give the desired information without the necessity to go through the subsequent steps. For instance, a poor conversion as measured on the reaction mixture will also lead to a poor yield of isolated product, 

There is another reason why measuring the response close to the interesting event is beneficial, viz. that any unnecessary manipulation will increase the experimental error, thus lowering the probability of detecting influences of the experimental variables on the response. 

---