“Imagine if we could access 100%…” said the science guy in the film Lucy. Well, I know what happens when my computer’s CPU hits its limit. It stops working.
Utilisation is a concept mainly stemming from consulting, where some work is billable. The target rate usually sits around 85-90%. For product-based companies it’s a spongier concept, but though it may not always be measured, the tradeoffs are relevant.
There’s solid mathematical modelling behind these rules of thumb. The Kingman Equation demonstrates that your backlog waiting time dramatically trends to infinity once you exceed 90% utilisation. You can edge higher in utilisation more safely if you reduce the variability of the work.
An ex-Google leader once told me that the 20% time initiative was effective in keeping workflows fast as well as nurturing innovation. Forcing utilisation rate to stay at 80% was the reason.