Your card turns up on their lock screen while they are around the corner.

This is the normal outcome, not the failure case. Geo-fencing is cheap to copy, so any street with two good operators on it ends up with both messages landing on the same pavement. Planning for that from the start is what separates a fence that keeps working from one that gets neutralised in a quarter.

What settles it is not who bids higher. Neither of you is bidding. It is which message is worth more to the person holding the phone, and a member who already has a balance with you is holding something the other store cannot match on the spot. That is why the copy points at the card rather than at the competitor: a discount they would have to sign up for is a worse offer than points they already own.

The tell that you are doing it wrong is a message that only makes sense if the reader knows who you are competing with. Write it so it would still read well if it arrived while they were standing in your own car park.

Point at what the member already holds, not at what the other store lacks.

This is the play most likely to run into a rule. Several jurisdictions restrict advertising directed at another licensed retailer, and many set a minimum distance from schools and treatment centres. Check before the pin goes down.

Brands already pay for placement in your shop: the shelf, the counter card, the budtender training. A geo-fence is the same kind of inventory, and it is the only piece of it that reaches somebody who is not in the room yet. A brand running a drop can have the window around your door for as long as the promotion lasts, and you take it down when it ends.

Keep the reward yours. The brand supplies the reason to come in, the member price and the points stay on your card, and the message should read as your store telling a member about something good rather than as an advert you were paid to forward. A member who works out that the second kind is happening stops reading the first kind too.

Run it as its own zone rather than editing your standing one. A separate zone can be pointed at a subset of cards, switched off on the last day of the promotion, and reported on without disturbing the fence you rely on all year.

The brand brings the reason. The reward stays on your card.

The coffee shop across the road, the record store, the tattoo studio, the venue. None of them sells what you sell, and all of them have a queue of people standing still within a block of your counter. Each side points a fence at the other and writes a message that is worth reading to somebody who came for the other thing.

It works because the trade is symmetrical and costs neither side anything. It stops working the moment one side writes a message that is really an advert for itself. The useful version names the neighbour: a member who hears "grab a coffee across the road and come see the new drop" believes the second half more because of the first.

Name the neighbour. A recommendation reads better than an offer.

Your standing fence talks to everybody, which means it mostly talks to regulars who were coming anyway. A second zone on the same address, pointed only at cards that have gone quiet, reaches a different person entirely: somebody who used to shop with you and still walks past.

The message can be blunter than your standing one, because the reader has nothing running. It is the one fence where naming the gap is fair, and it is worth writing as a person rather than as a promotion.

Same pin, different audience. The segment is doing the work.

When a second store opens, the members who should hear about it are the ones who already live or work near it, and you cannot tell who those are from a home address. A fence on the new store, pointed at every existing card, finds them by walking past it.

Leave it up longer than a launch campaign. The first month catches the people who happen to pass; the six months after that catch the people whose routine changed.

Geography finds the right members faster than your list does.