Batch inserts usually increase throughput by reducing the number of write IOPS, fsyncs, etc. They usually aren't associated with a 10x _CPU_ savings, which is the finding here.

There are a million 'db best practices' you can go implement blindly, but the point is that this methodology – determining the bottlenecking resource and then profiling to determine exactly what is consuming it – will _reliably_ yield huge wins, whereas implementing 'best practices' on gut alone is a very inefficient way to improve performance.