The numbers suggest this is not the case. They needed 17 extra employees for 68 existing staff members. Naively, one would expect that for 68 x 8 = 544 man hours per week, one would need 544/6 = 91 (rounded up) employees. Yet 68+17 = 85, which is six employees less!

Presumably some shifts require managers/bookkeepers/etc not required on other shifts. The overnight shift is less likely to require a receptionist dedicated to walk-in visitors and phone calls for example...

I'd be interested in how long they've had 68 employees for; was it 68 two years prior to the trial, or even 51?

I wonder if there's a balance somewhere here too. Perhaps have employees work 7 hours at a slightly reduced rate.

This trial alone is probably difficult to draw solid conclusions from, other than to say if these employees work less they produce less (albeit not proportionately it seems).

Ouch, I didn't stop to do the math. I had assumed the naive addition turned out to be too small, but this makes it hard to understand what surprised them!