Biological control systems have some distinct characteristics that must be taken into account to understand how such systems are constructed and which factors are most important in controlling system output. Grönlund et al. point out that one such important characteristic is the inherent time delay associated with regulation of protein abundance through transcriptional mechanisms. They used mathematical analysis to explore the effects of such delay in combination with system noise on regulation and efficiency of two model enzymatic processes. In one model system, an enzyme produced a substrate that was used as the substrate for another enzyme, a system where the amount of the product of the first enzyme may fluctuate greatly if the activity of the first enzyme is not tightly controlled. Usually one would find direct product inhibition of the first enzyme in such a system in combination with transcriptional regulation, such as through an allosterically regulated repressor. The authors showed that, if the time delay was not considered, transcriptional regulation alone was sufficient to keep tight control of the concentration of enzyme product while maintaining good throughput. If the time delay was included in the model, however, control was greatly enhanced by having direct product inhibition as well. In another model system in which two components are synthesized independently but consumed together (for example, in a dimerization reaction), the optimal system will balance the abundance of the two components in amounts that allow efficient production of the product, but not in such large amounts that the proteins are lost through degradation or nonproductive mechanisms. Here, in the absence of time delay, strong feedback is effective at low cost, but if time delay is present, strong feedback generates oscillations in system output. Thus the optimal system is one in which more moderate feedback is used at the expense of maintaining larger amounts of system components.