Poor power factor is a measure of "out-of-phase" reactive current. At unity PF (1.0) the voltage and current are in-step, rising and falling at the same time during the 60 cycle oscillation. This means V*A=VA and VA=W. With inductive loads (most common for industrial plants coming from motors, transformers, etc... ) the current lags behind the voltage. In other words, when the voltage rises, the current takes time to build in the inductive load. This means that a sine-wave of the voltage and the current do not lay on top of each other, but rather are offset from one another. Leading PF is exactly the opposite, with the current leading the voltage (caused by capacitive loads) This means that real power (actual W or HP) is derived from the voltage and current at the same instant in time along the waveform. Measuring the current in a reactive load will be a total of the in-phase current (real) and the out-of-phase (reactive) current. At a PF of 1.0 there the real current is 100% of measured and the reactive current is 0% of measured. For a fixed load (W, HP, any measure or work) as the PF lags (or leads) the reactive current rises but the real current stays the same. All this means is that regardless of the PF, the energy used is the same but the amps rise. The reactive amps are out of phase, so they cannot perform any work. They do take-up capacity in your wiring, transformers and motors and it does effect voltage drop. Many installations install PF correction capacitor banks or possibly synchronous motors (which can be over excited to supply the reactive current to the system). The proper sizing, design, and installation can be very complex costly. If you are not experiencing overheating of conductors, or tripping of circuit breakers, the cost may not be worth it. Another factor is that your electric utility supplier may or may not charge of poor power factor. If they do, this cost should be a part of your ROI calculations for possibly adding PF correction devices.
Hope that helps.