I'm concerned that this idea could actually backfire, and cause less participation rather than more.
If I remember correctly, GIMPS won a prize ($100K???) for the first million digit prime (or maybe 10 million digit prime), and part of that prize money went to the person who found the prime and the rest was held in reserve to award to future finders of even larger Mersenne primes. Each person who finds a new world record Mersenne prime at GIMPS therefore gets a share of that original prize. For argument's sake, let's say that smaller prize is $5000.
So while we might be offering a small "lunch-money" prize, it's entirely conceivable that the secondary prize money being handed out by GIMPS might be larger than what we're offering. If people are attracted to the prize money, the prize offering here could end up sending more people to GIMPS than PrimeGrid.
I'm also not entirely comfortable with the idea of having money involved in distributed science projects. It seems like the potential exists for funded projects, led by universities or corporations, regularly handing out small monetary prizes for participation. That would be an incentive volunteer projects such as PrimeGrid could never match. That could potentially be a disaster for projects like us.
My lucky number is 75898524288+1