If I may borrow from Jim Collins for just a sec, at CLASSIC.COM we have a “big hairy audacious goal” (BHAG): to help our users find any car, anywhere, and price that car in real time.
The ‘find any car’ part of this equation is a relatively straightforward problem to solve: aggregate vehicle data from all over the world, organize it in a manner that makes sense to us humans, and make it easy to search for. Of course, there’s quite a bit of technology that goes into solving that problem, but we’ll get into that in a future post. In this post, we’ll focus on the second piece of our BHAG, which is to ‘price that car in real time’.
Why the status quo is what it is
In the past, the pricing of vehicles in the classic and exotic car world has been the strict domain of experts, professional appraisers, and people with access to proprietary data. This has been the norm for many reasons, including the fact that there is no central registry for all vehicle transactions, no standardized vehicle taxonomy, and to some extent, an inherent interest by a few old-school market participants in keeping the information close to their chest. The end result is the status quo: a market that only some dare to participate in. In fact, we fundamentally believe that the market for classic and exotic cars would be far larger and better for all participants, if access to information was more transparent.
But even if we created the central registry, created the most commonly used vehicle taxonomy, and provided incentives for industry players to share information – things we’re actively working on – that still leaves us with the question of how to price classic or exotic cars, in real time.
Solving for the 80%
To be clear, our goal is NOT to solve 100% of the pricing problem (we define 100% as the exact price that someone would/should pay for a certain car). The goal is to solve 80% of the problem, by isolating all of the subjective value drivers of classic and exotic car valuation, such as emotional aspects (nostalgia, fomo, etc), celebrity provenance, maintenance provenance, certain modifications, and so on. Of course, those subjective value drivers can cause an individual vehicle to sell at several multiples of its market value, but in general, looking across 10’s of thousands of data points, we believe that subjective value drivers represent only 20% of the pricing equation (more on this later). The other 80% boils down to largely objective value drivers that we can attempt to control.
The idea is relatively simple: Comps. That is, comparable data points.
In essence, our plan is to continuously gather millions of vehicle listings (data) and normalize those listings across a number of Objective Value Drivers that we can control. After normalizing those listings, we can then generate a list of Comps for any vehicle, anywhere.
With enough closely related Comps, we can provide a very precise estimate of the market value of any car, anywhere. And as the software becomes more precise, more people will want to contribute their data. The volume of listings will increase, and then the software becomes more precise (yes, a circular reference known as a data network effect).
To that end, Comps on CLASSIC.COM now provide our users with a real-time list of Comps for any given vehicle in our database. I wouldn’t call this the end result (by any means), but it’s definitely a step in the right direction – baby steps.
Comps on CLASSIC.COM today
As an example, you can see a list of Comps for this 993 Carrera 4S, which sold for $109,200 in March 4, 2022.
In that particular example, as of this writing, the system found 5 comps with a relevance score of 100%. This means that for the Objective Value Drivers that we currently have for this model, the system found 5 listings that match exactly to the target. From the 6th listing onwards, the match score starts to decrease based on Value Drivers that fail to match to the target, ie, less relevant comps.
As you can imagine, certain things matter more than others. For example, exterior color generally matters more than interior color – and that is weighed into the algorithm. But it gets complicated: the transmission type in a Ferrari F355 is a huge determinant of value, but not so much if it’s a 2nd Generation Toyota 4Runner. Therefore, expect relevance scores to change overtime as the system learns what matters where and we continue to tweak the formula for each model.
Currently, the Objective Value Drivers we are tracking are: Taxonomy (Make, Model, Generation, etc), Attributes (Colors, Transmission Type, etc), Conservation (Original, Restored, etc), and Location.
We don’t have these backfilled for all listings, but that’s something we’re working on.
Future iterations of Comps will include other Objective Value Drivers such as: Mileage (is the comp relevant to the target with respect to mileage), Recency (how recent is the comparable data point), and Condition (what is the condition of the comparable data point relative to the target)
But it doesn’t stop there. To really get to the end result (ie, price that car in real time), we’ll need to grab all that data, sprinkle on top the real-time supply and demand dynamics for a specific model, add some relevant macro-economic indicators, and voila; we’ll have an ESTIMATED MARKET PRICE.
Solving for the 100%
I mentioned before that our goal is to solve 80% of the equation. That’s because we understand that classic and exotic car valuations, unlike commodity cars, are impacted by subjective value drivers that have little to do with the market: nostalgia, celebrity provenance, the real or perceived quality of the maintenance or restoration history of a specific vehicle, awards, and much more. But looking across hundreds of data points (and even just a dozen in cases of low volume markets), you can isolate the impact that subjective value drivers have on specific models.
So who’s going to solve for the remaining 20%? That’s where dealers, auctions, and appraisers will continue to play a role. Our software is not going to replace any of those market participants. It will merely provide a level playing field to help millions of buyers, sellers, or any intermediaries make better decisions.