Biochemical reaction yield and enzyme promiscuity

Reaction yield, i.e. the molar percentage of product over substrate, is often mentioned by chemists, but never by biochemists. My guess is that many enzymes are not perfectly efficient, but have a range of reaction yields.
In The hitchhiker's guide to the galaxy a ship is hidden thanks to the "somebody else's problem" principle, namely people will ignore something problematic that isn't their problem. The reaction yield of enzymes is not something often discussed. The reason is pretty self evident: differentiating between low abundance products would be a minefield of pesky technical issues. So it is somebody else's problem.

Catalytic promiscuity on the same substrate

There are some cases where catalytic promiscuity on the same substrate is known, that is, there is a poor reaction yield. Human serine racemase fails to stop serine undergoing β-elimination half the time, the reason being that there is little selective pressure to optimise it as a result the enzyme has a reaction yield of 50%. Another example are some sequiterpene synthases from plants and fungi, which convert a single compound farnesylpyrophosphate into a variety of different compound, a fraction of which are known to be bioactive —evolutionary plasticity plays a part, but that's a tangent for another time. In both cases, like most reactions, an unstable intermediate is formed which is forced in a direction that it would not spontaneously, but the enzyme residues do not always stop it. So a near-zero uncatalysed rate is surely a factor —but it is not the main factor.

Selection is surely a main factor. Low-turnover wasteful promiscuous activities are tolerated either because they are too small and/or their suppression would affect the main activity too much.
Enzyme evolution is never straightforward. In secondary metabolism suicide enzymes are sometimes found: even though the cellular cost of, say, a 200 aa enzyme is more than 200 ATP, not much product is needed, so that extravagant expenditure does not dent the cellular economy or is cheaper than having a series of extra enzymes. Less dramatically, in secondary metabolism there are several examples of enzyme reactions that, were I to design them with the standard cellular biochemical logic, I would adopt more, but straightforward, steps. Enzyme efficiency correlates with the requirement of the product, so these insanely complicated steps are kept because they do not need a high catalytic activity. So if I were to place a wager on what enzymes have the lowest reaction yield, I would place it on enzymes with low catalytic efficiency (≈low selection) and the most diabolical reaction (≈zero uncatalysed rate). Ironically, these enzyme are enzymes to avoided at all costs, so there is no sane way of knowing…