There are few things to take into account:
1. We're working with real dealers behind the scenes, calling them if we have to. We don't use automated systems (not because it's impossible, but because it leads to a poor consumer experience). So there's a cost aspect to it.
2. Since dealers can see each others' offers, there's some competitive dynamics at play here too. The nth dealer to join often thinks that s/he has a worse chance of winning the deal because other dealers got there first. The effect varies proportionally (possibly exponentially) with the number of dealers participating. We don't want to promise 5-7 dealers if we can only realistically get 3-5.
3. We split-tested the value propositions for our plans, especially around # of dealers. What we've got now seems to work pretty well in terms of driving people to pick the plan we want them to pick (Plus).
4. We've found that with 3-5 dealers, buyers generally reach the local (in terms of geography) minimum w.r.t to the price of the car. Adding another 2 dealers doesn't get the buyer much and only wastes the dealers' time.
That said, we will continue to monitor this and test it. Thanks for your feedback!