i should openly disclose we've done some studies in the context of a commercial product, inside www.andrewalliance.com, but i thought it was anyway worth reporting our outcome. There is no way to universally quantify the benefits of pipetting improvements. Let me take one example. If you have a protocol that manipulates magnetic beads, whatever pipetting improvement you take will be dominated by the uncertainty/lack of reproducibility introduced by the beads themselves. If you have a process where all other parameters are well under control, and reagents are uniform and of constant quality over time, you could see an improvement that can be up to 10x (from 20% to 1% or so) by comparing bad pipetting techniques with good pipetting techniques (we see this with glycan analysis and testing). We quantitate the improvement by comparing our robots to manual users, and assuming our robot is equivalent to the quality of a calibration expert. So, it finally depends on your biology, reagents, and the intrinsic reproducibility of your process.