How do you test a logo with real customers?
You test it on the door or the menu, asking whether people recognise the venue and can read the name, not whether they “like it”. Friends voting on a phone are not the customer at noon. Pick a service day and show the proof to someone already in the room, at real size. Note the words they use, not a score from one to ten. A story poll is run and the oddest option wins. Then on the pavement nobody can read the name.
A note of what five real customers said in front of the proof, and the adjustment that comes from it: size, contrast or the word that was not understood. The whole thing is not redrawn for a preference.