Yes, this is the correct approach. You need the onstage speaker to localise the source of the sound to the prop you want the audience to think the sound is coming from. If you just had an EQ'd sound in the mains then someone sitting close to a main speaker would wonder why the sound of a cellphone speaker mid stage was coming from entirely the wrong direction.
There are scenarios in which you might want your futzed sound in the mains as well as the prop cover speaker onstage.e.g If you wanted an exaggeratedly high level greater than the onstage speaker is capable of (particularly with battery powered speakers fed from IEM receivers).
You would create 2 copies of your futzed cue (with the bandpass effect) route 1 to to the onstage speaker and the copy to your mains. On the cue routed to the mains after the futz effect you would insert a delay ,e.g AU delay with 0% feedback and either 0.02 or 0.03 secs delay (20 or 30ms). The idea is that, even if you were sitting in the seat next a main loudspeaker, the sound from the onstage speaker would arrive at your ears a few milliseconds before the sound from the main speaker, which would retain the apparent source of the sound as the onstage speaker. AU delay only allows you to set delays in 0.01 (10ms) increments equivalent to around 10 feet of distance. If you want finer control you will have to use a different plug-in
It's even worth using this technique (without the bandpass filter on the cue routed to the mains) so that the full range copy of the sound is exactly in sync with the onstage sound when you fade it up, for instance you were doing a headphone bleed cue with a fast ttsss ttsss ttsss ttsss hi-hat pattern, with the clean track fading up on the mains. This would avoid a doubling of beats for the first few seconds of the sound fading up on the mains.
Mic