The point is, if the motor is running at nominal load, the energy saved when using a variable frequency drive (VFD) is about 3%, for 75% load is about 25%, for 50% load is about 75%. So to calculate how much energy saved by the VFD, you have to check how long the pump/motor will run on under nominal load conditions and how big is the difference between nominal and operating load. Now, this x% has to be translated in kWh and then in money. This money you have to compare it with VFD price and see how long it will take to get your money back from this investment. This is if you run the pump on x% load. Now add the time the pump will run on nominal load and you will see how long will take to have a profit from this. If a VFD costs 5.000$ and saves 100$ in a year it's no point to buy it. You will need to replace it even before you will have your money back from the investment.
Lots of applications that will save much of energy by installing a VFD (Primary, secondary air, exhaust/flue gas extractor fan, cooling pumps for sure, motors used for automatic adjustment of parameters). Anyway, in case of the plant/motor is used for power balance regulation when the load is likely to increase/decrease more frequently, you can install a VFD for saving energies. If your plant is running under nominal speed a long time, the energy saving with the motor is a minor problem for you. The big loss will come from other points like electricity consumption, motor efficiency, maintenance etc. Fitting a VFD for such motor will not help too much.
Interesting to buy a VFD? Shopping on GoHz now, 2 hp, 3 hp, 10 hp ... and more.