I found the TV ad for this device implausible. For it to work by sending any kind of signal over the house wiring would require the target organism to have some kind of receiver, responsive to the radiated signal. Three types of radiated signal come to mind: electric field, magnetic field, and acoustic field caused by mechanical vibration of the conductors, which could result from an AC magnetic field.
The ad claims that each unit covers one floor of a building. Clearly this has to depend on how the building is wired. If each floor has a single branch line from a common fuse or breaker box, then at least the required topology is satisfied. How many homes would conform to this? I believe most homes have at least the ground floor divided up into multiple separate lines from the central fuse/breaker box. Coverage would either be limited to a single branch circuit or else include multiple floors. This was the part of the ad that seemed to me the least plausible.
All of the wiring in my new (unfinished) home is BX (spiral sheathed) cable. I doubt any significant electric field would radiate from this wiring. Even with unshielded cable, unless the wire length is a significant fraction of a wavelength (say, > wavelength/20) the only significant electric field would be in the space between the conductors. If the cable length is 100 feet from the fuse box to the device, this corresponds to a minimum frequency of around 500 KHz. Shorter lengths would require higher frequencies. Can mice, cockroaches, etc. detect this? Also, since electric field is produced by voltage rather than current, and at 500 KHz the transformer out on the pole would present a rather high source impedance, the voltage signal would be essentially the same across every branch circuit served by the same transformer. It wouldn't matter which floor or branch circuit the unit was on. It could be even in another house served by the same transformer.
What about magnetic radiation? Since the currents in the two conductors are in opposite directions, they produce opposing magnetic fields. Magnetic radiation would be detectable only very close to the cable, such that one conductor is significantly closer to the receiver than the other conductor. Only the portion of the cable which carries current to the device would radiate any magnetic field signal. To maximize coverage, the device would therefore have to be as far as possible from the fuse/breaker box. While many organisms are sensitive to magnetic fields, I believe only their navigation is affected. I have no idea whether they can detect AC fields.
The last possibility I considered was acoustic radiation from the conductors due to the force from their magnetic fields. The conductors repel each other with a force proportional to the current squared (correct me on this, somebody, if necessary). If the device works at all in accordance with the claim that the house wiring delivers the signal to the receiving organisms, this seems to me the most likely modality. Since coverage would depend upon how far the device is, along the wiring, from the fuse/breaker panel, it would also account for some of the inconsistent results reported by various users.
I recall a childhood experience which seems to support this hypothesis. I was in the basement using my father's old, rather large, electric soldering iron. My mother was upstairs using either the vacuum cleaner or the electric mixer, I don't remember which. I could hear the electrical noise from the appliance as a sound emanating from the soldering iron! It was unmistakable because I could hear the appliance upstairs varying in speed and simultaneously hear the sound from the iron. I wasn't about to take my father's soldering iron apart to investigate (hey, it still works 50 years later), but I suspect that the heating element is coiled and the current through it was modulated by voltage noise across the power line. The impedance of the transformer outside the house would have permitted the relatively small ripple current through the appliance to produce this voltage signal. The modulated current in the heating element produced a modulated magnetic field which produced vibration, which produced the sound.
So how much sound would emanate from the house wiring? Probably not much, and it depends on the current signal. If the device injects a current signal into the power line, it must be at a voltage sufficient to pass this current through the impedance of the line at that frequency. This impedance, for high frequency sounds, is rather high if it's determined by only the transformer on the pole. But that's never the case - it depends on other equipment connected to the line. This would be another reason for inconsistent results from different users.
.