With the acquisition of Car.com, Autobytel needed to increase the level of sophistication of its lead routing engine. When we were just a single site with a single dealer network, lead routing was fairly easy. Times evolved, the number of affiliates grew, and we acquired several of our competitors. Now we were managing leads across multiple dealer groups, sometimes facing multiple contracts with the same dealer across multiple sites, often with different terms and territories.
We were now managing 300,000 leads per month, and our existing system was showing it's age. For every acquisition, we were cloning tables and running multiple instances of the same application. To add to the complexity, we were also selling leads on the wholesale market directly to the OEMs. There had to be a better way.
We started with the core concept that a single instance of our lead routing engine (now dubbed a yield management system) would handle the workload. It had to be sophisticated enough to understand the contract variances for dealerships across territories as well as across companies and affiliates.
We crafted a system that could look at an individual dealer's run-rate. If he was on or above target for meeting his monthly goal, then we would attempt to route the lead to another dealer, or perhaps the same dealer under a different contract. If all else failed, we would check for a need on the wholesale market. Affiliate leads would be rejected at this point, but internally generated leads were always accepted.
One of our findings was that we were actually throwing away about 20% of the leads that we bought due to non/under coverage for that vehicle in that territory. This meant that on day one we had an annual bottom-line profit increase to the company of $6 million. Beyond that, we improved overall profitability of the mix, and were able to capture demand data and use that to build a sales estimating tool.