I like Latexman's suggestions. Unless your electricity is free and your performance is not important, the performance spec is a requirement. If it's at all practical, consider having the manufacturer guarantee the results.
Disclaimer: I work for a major mixer OEM. I've seen the incredible advancements in mixer efficiency, and I've also seen tremendous waste and difficulty due to archaic specifications and poor assumptions.
Some things to keep in mind:
1) New mixing designs can make immense energy savings over previous designs. The big savings come from proper impeller sizing, proper RPM, and proper sizing.
2) Some new mixing designs look good but the performance claims are phony. My company makes extensive use of scale testing to determine which ideas work and which are just pretty.
3) Scale tests of mixing are an excellent way to determine real-world requirements and size the full-size equipment accurately. The only exception to this is aeration, which is not scalable, but can be tested in situ.
4) Mixer sizing strategies are usually based on horsepower or torque. Horsepower is cheaper to build (bigger motor/smaller drive and impeller) but torque is usually more energy efficient.
5) Mixer impellers generate vortices that cause bending loads on the agitator shaft. This is normal. The bending loads must be accounted for in the overall design - make sure the vessel will not deflect too much in operation. Steer well clear of a mixer manufacturer that is not forthcoming with this data.
6) Some mixers are built with commercial gear reducers and others use mixing-specific designs. The bending load I mentioned before can overload the output shaft of a commercial reducer. Also check on the maintenance intervals - some use sealed/greased bearings that are to be replaced annually. (That can be a big waste of your time and money). For heavy duty loads or minimal maintenance, a mixer-specific reducer is usually justified.
7) If your tank is not yet specified, keep in mind that making the tank too tall (mixer shaft too long), can be very costly. It can also drive up the loads on the tank nozzle.
Scale tests might seem like overkill, but they can optimize the design. Mixers look like (and really are) simple devices, but mixing dynamics are not, and scale tests are often very necessary. CFD is insightful but it cannot consistently predict everything you need to know. (Beware the vendor with CFD but no test data to confirm the validity of the analysis) If you're serious about the highest product quality and the lowest energy consumption, the scale test can offer massive improvements in the design.
David