I've been toying around with Cars Sold / 1,000 Visits correlation lately.
Please note, I am by no means a math or statistical whiz nor do I pretend to be one on interweb forums 
Here's a real life example:
Dealership A (name has been redacted to protect the reticent)
November 2013 | 14,262 Visits | 299 Cars Sold = 21 Cars Sold / 1,000 Visits
May 2014 | 22,198 Visits | 464 Cars Sold = 21 Cars Sold / 1,000 Visits
The big jump in traffic from November to May mostly thanks to turning on a PPC campaign.
This particular store consistently performs in the 21-24 cars sold / 1,000 visits each month regardless of traffic fluctuations.
I don't know if this is good, bad or indifferent, but I like seeing that it's consistent.
Please note that this is combined new and used sales, and doesn't take in to account traffic source (internal traffic is filtered though).
If anyone is willing to calculate their own Cars Sold / 1,000 Visits I'd be quite intrigued to see your results.
