As long as I can remember, I have always had a problem with the concept of isolated yield. I mentioned this in the past, and I will say it again: this fuzzy metric is driving me insane. Isolated? By whom? By a first year undergraduate student or by a postdoc? We work at educational institutions and there is a lot to be said about how students grow in confidence and experience. This always has a direct effect on their technique.
The point is, there are subjective and objective metrics in science. While isolated yields are clearly subjective in nature (there is a real person behind the number), conversion and selectivity (I primarily refer to chemoselectivity here) are parameters that exist irrespective of our meddling with reactions. Personally, I am in favour of not considering anything subjective when it comes to process control and optimization. This means that our students need to have a good grasp of monitoring reaction selectivity. If you multiply selectivity by conversion, you get assay yield, which is just perfect as it decouples us from our inadequacies and focuses on what’s of interest to anyone who wants to read our papers – I refer to what actually happens in our reactions. There are many analytical tools available to our students and I want to focus on one of them tonight: GC and its use in catalysis. Choosing the right internal standard is the first step: one needs to ensure that the corresponding molecule is not being chewed up while the reaction of interest is taking place. Once you have the right internal standard figured out, there is a tendency to assume that things should be fairly straightforward. Not so fast. Here is a great paper we discussed at our group meeting earlier today. When I visited the UK a number of years ago, I became aware of this short, yet highly educational piece by Professor Fairlamb. The central message is simple: the fact that you have a nice GC vial containing a mini-worked-up aliquot does not mean that the reaction has stopped. This paper emphatically demonstrates that colloidal palladium cannot be seen by a naked eye, and that no matter how clear your vial is, your analysis might be in jeopardy. I love Figure 1: you can clearly see that the Sonogashira reaction still goes on in the GC vial as you wait for your sample to be processed. I wonder how many other reactions display this behavior during analysis…
http://www.sciencedirect.com/science/article/pii/S0040403904005763
amphoteros 
There is a similar, common problem with TLC-analyzing reactions performed at low temperatures: A good technique (a long TLC capillary inserted through a 16-gauge needle stuck through the septa under milld positive pressure of argon) prevents oxygen from getting into your sample, but it won’t prevent the reaction mix in the capillary from warming up before you spot it on the TLC… {Another tangentially related problem is analyzing heterogennous reactions (i.e. powdered KOH + 18-crown-6 in benzene, with a polar substrate, or tri-phasic phase transfer reactions) – a significant fraction of the starting material or product can be stuck in the inter-phase, on the surface of the solid or the glassware, so the sample of the reaction mix that you take is not really representative.}
In cases like this (for kinetic study experiments) it is better to take fewer data-points, but run series of fully worked-up small scale experiments, set up in parallel, rather than trying to sample a reaction which cannot be easily quenched…
Interesting! I actually used to use that needle trick myself… Thanks.
It is also infuriating when somebody reports a 5% difference in isolated yield as “improved”, especially on the sub-millimolar scales typically used in methodology papers. That’s the difference in a lost drop while loading a column….
Agreed…