We can derive two further insights from this. The first is that reinforcing the weakest link will eventually yield zero returns, because another link will eventually take its place as the constraint or the bottleneck, limiting the performance of the entire chain. The second takeaway is that because we cannot predict where the constraint will move, we need to constantly monitor the entire system in search of the next weakest link. Blindly optimizing a single part of the system—even if it was once the weakest link—will eventually lead to waste. This is the local optimization trap.