Track online car sales in Google Analytics with no new tech or complex data integrations. Yes, it's possible and no you don't need to implement a Digital Retailing solution!
Ever since I got into the automotive digital marketing business, figuring out how to track sales on dealership website has been the "holy grail" for automotive marketers. E-Commerce websites have developed sophisticated analytics and tracking capabilities that allow businesses to understand exactly what leads to online revenue. By my estimation at least half of the features in Google Analytics are only useful if you're tracking sales. Until recently, I thought this kind of capability was going to remain out of reach for automotive websites until Digital Retailing was broadly adopted. Keep reading to find out what changed my mind and how you can take advantage of it...
For years I've been talking about the need to find ways to measure quality and not quantity in order to make decisions about what marketing efforts are working. A while back, a consulting client told me they were on board with the need to measure quality, and asked if I could implement a single quality score in their Google Analytics account to measure their marketing efforts. My response? "Good question. I'm not sure, but I'll try."
After starting the project, I quickly discovered / remembered that Google actually has a feature called Session Quality Score. When it first came out, I immediately discounted it for automotive use because it doesn't activate unless your website is firing "E-Commerce Events" - shopping cart adds, checkouts, sales, etc. This time, likely fueled by my desire to be able to go back to my client with a solution, I thought about it again and had an epiphany : I could develop an algorithm to "detect" a sale on automotive websites, and then fire virtual sale events into Google Analytics to unlock all of their E-Commerce features (Assisted Sales, Attribution, Time to Sale, Visit Paths to Sale, Revenue Generated per Campaign...the list goes on and on.)
I set to work implementing the algorithm, which is heavily influenced by shoppers identifying themselves as a real person (i.e. not a bot) via their behavior, and also repeatedly demonstrating interest in the same car. I also implemented a mechanism for filtering out employee traffic (hint : they use the website a LOT more often than actual buyers.)
The algorithm has now been implemented for a large dealer group (~25 stores) for about a month, and we're seeing amazing results. The sales reported by the algorithm are within 20% of actual sales, and we've recently made some improvements and tweaks that I think will bring that number within 10%. When the standard has been "nothing" for years, a model with 80-90% accuracy has proven to be incredibly useful.
Here are some screenshots of the reporting that this tooling enables...
Choose Wizely and reach out to firstname.lastname@example.org if you'd like to learn more!