I can say I agree with Joe that this seems like a far fetched number but each individual dealer can check whether this is true for them or not IF they use their CRM correctly.
Many studies seem to be supporting this "trend". Of course companies like ATC are very supportive of these studies, no matter how loose or inconclusive the data could be, as it places a large gray area into the actual performance of classified website and it's relation to walk-in traffic. Which is difficult to track.
I'm not taking any sides on the messaging although anyone who knows me would know I'm incredibly biassed on this topic. Reasons: 1) a lackluster relationship with ATC & VinSolutions during my dealer days and 2) Autotrader is a competitor. However, I believe I can add to the conversation by pointing out technical problems with walk-in tracking.
Surveys
Post sale surveys are fantastic for trend discovery. If your General Manager(s) has not been exposed to this kind of data it will change his way of looking at the store. Asking the right (non-leading) questions is difficult to do, so this is where a third party can assist in crafting the survey questions. One part is extremely difficult and that is asking marketing questions. Usually the people taking the survey have or are in the process of buying a car from you; they're biased to you. They're also just trying to get the survey over with because marketers typically ask too many questions.
In my experience, most people clicked off what they knew. We found things with stronger brand recognition were cited in responses more, but people couldn't recall actually using that service after drilling into their response deeper (asking them outright). Human psyche plays differently when taking a survey vs. actually buying a car.
CRMs
We all know our staff is either the worst at obtaining the proper information or way more efficient at collecting the right info than the customer is when they submit Internet leads to us. Imported databases and the time when the core of the CRM was originally built also factor. Then you have the practice of how the CRM is used within the sales process. For the record the "accuracy" of the retail CRM is based on:
- -The customer
- -The staff
- -Legacy data (imports) from prior CRMs/ILMs and DMS systems + any list-purchasing endeavors
- -The CRM itself ....fixable
- -The process ....also fixable
I'm going to skip the first three....
Most CRMs available to the car business were built in departmentalized silos. This is especially true of systems built before 2003 (pre-Internet-acceptance-in-the-car-biz years). They concentrated efforts to work with DMS systems that are also silo'd data centers: Deal silo, Service silo, and the Inventory silo are the 3 traditional silos CRM companies built around. Since the Internet lead became such a viable part of business, technologies have been developed to better connect the customer who calls himself "Alex," but is really "Alexander" when signing a contract. The flexibility of today's technologies are a lot better at decreasing the amount of duplication simply because it has had time to evolve in comparison to some older systems being sold today. I can speak about this for ages, but will stop to not bore you.
The dealer's process can be different based on how certain systems are utilized. Do they even care about de-duplication enough to really do something about it? Or, even at its most basic, is a CRM or DMS even used? And there are still a ton of stores (not so much groups anymore) using ILM systems that are foreign to the staff on the floor.
Anyway, if CRM data is used (and I can guess which would have been in this case) to make definite claims, it is full of holes.
Man-in-red-hat tracking
A minority of stores use a way of counting lot traffic by physically doing it with administrative staff (maybe even a manager). Someone watches the lot and puts a tick down for every customer who walks-in with a description of the customer's appearance along with who upped that customer. Then it is the job of the sales agent, who did the upping, to assist the administrator with a name and status as to what happened with that floor up. Overall, this produces some of the most accurate floor traffic counts! And if managed properly it can make some strong connections to information already in the CRM. However, this method isn't widely used so it isn't a viable way of seeing national trends.
What else?
I guess you can try to connect ADF data to DMS data. You can hire companies like Polk to mix registration info in (looks like they did some of that).
Back to my opinion and it is a WTFer:
At the end of the day can we really say people are skipping past submitting Internet leads in droves before walking in? If that's the case why are we seeing increases in Internet lead activity (including chat) from where I'm sitting? Is Autotrader saying more people are visiting the store, leaving, and then submitting Internet leads afterward?