Predictive analytics for car dealers is software that predicts which past customers are ready to buy their next car, so that the dealer calls them first. It is a part of the dealer management system.
A dealer's own records already hold the clues. The DMS knows when each customer bought a car and how it was paid for. It knows how many times the car has been serviced and when the loan will end. The software looks for the patterns that earlier buyers showed before they bought again. For example, many people buy a new car soon after a loan ends, or when the old car has started needing costly repairs.
Say a dealer in Ludhiana has 4,000 past customers and six sales executives. They cannot call everyone every month. Without software, they would call in the order of the list, and many calls would go to people who bought a car only a few months ago. The software instead gives each customer a score for how likely that person is to buy in the next few months. The executives call the highest scores first. Each call is recorded in the dealer's CRM, and the result teaches the software which scores were right.
A score is only a likelihood, not a promise. A wrong guess costs the dealer one phone call, so the dealer can afford to be wrong often, as long as the top of the list is right more often than a random list would be.
The same records show which customers are ready to sell their old car. The dealer can then offer to take it in part-exchange. It later sells the old car to another dealer at an online car auction, or lists it on a used car marketplace.
This is different from a CRM. A CRM keeps the record of every enquiry and every call. Predictive analytics decides who should be called.