I believe a dealer needs more than 1 attribution model and those models should be broken down by New and Used. The overall concepts I'm posting are all generalities. Analytics Attribution helps dealers make marketing decisions, talking to your customers grades those decisions. Analytics Attribution and KPI's are a monster and different at every dealership.
Talking to a customer, I believe the first thing you want to establish is the "heaviest influence" that brought them in the door. Family/friends, former customer, internet, etc.. First and foremost I want to know if the starting point was the dealership or a vehicle shopping journey. After that I want to know the heaviest influence inside the starting point, all I care about are things that made an impact. I agree with the charts above that it's usually around 6 influences on the internet, just need to weight them on their role and monitor duplicity. I also believe the dealership should be extracting information about the whole process, but you can grow into that. (i would be writing all of these insights down and storing them in a shoe box. AI will be coming to automotive at some point and historical customer insights will give you a competitive advantage).
From there I believe you can combine analytics attribution and customer attribution to build correlations and look for signals. Signals could be dealer weaknesses or a waste of money among other things. It brings together a clearer picture for grading and discovering new opportunities.
I would never list my website as a source, I would judge it on its own merit. This is about as basic as I would go in trying to even tackle attribution, over time you can expand into where the data takes you or what you want to discover. Any sense of trying to get close to a true attribution model without your customer's input is guessing, IMO - you're lacking intent and value. I also believe a true attribution model is impossible to attain, you just want to get as close to it as possible so you can make the right decisions faster.
Start easy then go micro.
