Kilowatt-hours per cubic metre delivered is the one number worth putting on a wall. It is also the number most often reported in a way that cannot be compared to anything.
Three ways it gets reported wrongly
Against water produced rather than water delivered. If a fifth of production leaks away, energy per cubic metre produced flatters you by exactly that fifth, and it improves every time a leak gets worse.
Averaged over a year. Pumping energy is dominated by summer peak, and an annual mean hides the months where the money is actually spent.
Excluding boosters. The trunk main is easy to meter and the district boosters are not, so they are frequently left out — and on a hilly network they can be half the total.
What good looks like
There is no universal figure, because static lift dominates and geography sets that. The comparison worth making is against yourself last year, on the same boundary, with the same pumps in scope.
A network that improves this figure by ten per cent in a year has usually done one of three things: fixed a control problem, replaced a badly worn impeller, or found an air pocket.