Back in 2017, I had two really wonderful things going for me. Number one, I had the privilege of doing business with thousands of dealers through the better part of the previous decade. So we had a database. With millions of used vehicle transactions that we'd seen come and go through dealer's inventories. And I was also very fortunate, back then, as I am today, and that is that I had the benefit of some really brilliant people, some really brilliant data scientists. I'm proud to say that in the room today is David Rice, who heads up a team of a couple hundred people here in Austin who are heads down coding. And David leads that team of these brilliant people. And those two things working for me, we went to work. And the first thing that we did is we took a couple million used vehicle transactions that we had in our database. Vehicles we'd seen come and go through your inventories over the past decade or so. And we broke them up into three categories, ones that achieve the highest our ROI, the medium ROI, and the lowest ROI, very often negative ROI. So we put them into those three buckets, high, medium and low ROI. And then what we did, the data scientists did, is they looked to see if they could find characteristics of those vehicles while they were in the dealer's inventory, that had statistical correlation to their ultimate return on investment when they sold. And what they discovered is that there are, in fact, many characteristics that we can identify about a vehicle on a dealer's inventory that does have statistical correlation to its ultimate ROI on its sale. But among all those characteristics, there were three that were found to be most highly statistically correlated, predictive, if you will, of the vehicle's ultimate ROI performance when it sold. And I'm going to introduce those to you right now, but I'm going to tell you not one of them is going to surprise anybody. Each one individually will make intuitive sense to you. So the first condition that we can identify of a vehicle while it sits in a dealer's inventory that proves to be highly predictive, reflective of its ultimate ROI is simply how right you own it measured by its cost to market. And I think that makes sense, right? Because when you own a vehicle right, it's much more likely you'll make a larger gross profit on it. And gross profit is one of the three elements of ROI, along with how much should you have to invest and how long did you have to hold it. So I think that makes just intuitive sense. The second highly correlated condition is simply it's market day supply and that too makes sense, right? Because the more right or I'm sorry, the higher the demand, lower supply, the vehicle, the less likely you are to have to discount it or negotiate or negotiate as much. And again, that too often leads to a higher gross. One of the three elements of ROI. And then the third feature condition of a vehicle we could identify wallets and inventory is simply its level of popularity in the retail market measured by its retail volume. So each one of these individually made sense. And then what the data scientists did is they melded them or blended them together into what they call an investment score. Now, I think what's really important to note here, but don't want to go deep into the weeds here now, but it's important to understand that those three factors individually and then collectively, while they're blended into what we call the investment score, the weighting of those three factors is not one third, one third, one third for all cars. The weighting of those three factors depends on a variety of factors. So on any given car, the weighting of cost, the market, market day supply and retail volume will vary. But with that said, we created a scoring and what we call an investment scoring system on every vehicle. And let me explain how that scoring system works. It's on a scale of 1 to 12. So a vehicle in a dealer's inventory that might score a 12 on a given day would be a vehicle that we can predict with almost absolute certainty that when it retails, it's going to generate among the highest ROIs. That would be a vehicle that would be characterized by a low cost to market. You're in it really, right. Car's got low market day supply, high demand, short supply, and it's a high volume retail seller in the market. On the other end of the spectrum, a vehicle in a dealer's inventory on a given day that might score a one would be the opposite, would be a vehicle that we could predict with almost absolute certainty when it sells, it's going to produce little, if any, and likely a negative ROI, be a vehicle would be characterized by a high cost to market. You're in it for a lot of money, maybe too much money. It's got high market day supply and it's got low retail volume. So starting back in mid 2017, behind the scenes in our database, you didn't see it in your applications back then. But behind the scenes, the database, we put the investment score in every vehicle and then we watch when it's sold to see if our investment score was in fact reliably predictive of its ultimate ROI. And throughout '17 and well into '18, we kept tweaking it and tweaking it, the weighting of those three factors until we got it to a point where it is frighteningly accurate, frighteningly accurate, that we could predict the likely ROI outcome of that vehicle when it sold while it's sitting in a dealer's inventory. So once we got it to a point where we felt it was well, sufficiently accurate. And again, we're at that point we're scoring, you know, couple of million vehicles every day and watching their outcomes across roughly around 12,000 dealers or so. There was a day that I asked the scientist, I asked them to do something. When they did it, and I saw what they did, I asked them a question and they gave me the answer to the question. I knew that by mid 2018 that we would in fact change the way that every dealer manages used vehicles. Not overnight. I can assure you, not overnight. But we will eventually change the way that our industry manages used vehicle inventory, perhaps again. I also knew at that point in time that I wasn't ready to ride off into the sunset. So what is it? What is it that I asked the data scientist to do that led to that question that led to that conclusion? Well, one day I said to them, I said, you know what? I said, "We're scoring a couple of million cars for 12,000 dealers every day." I said, "What I'd like you to do is roll them all up." And what we did is, is we created four categories. The vehicles in a dealer's inventory on a given day that score ten, 11 or 12. We call those platinum cars. The ones that score seven, eight or nine, we call gold; four or five or six got called silver. And the vehicles that scored a one, two or three on that 12-point investment scale, we call bronze cars. So I said, "Roll them all up as if they're just one big dealership, one big enterprise, and show me what their inventory looks like in these four precious metal buckets." And this is what they showed me. This is what they showed me. Now, there probably are many things here that are worth questioning and talking about. But there's one thing on this screen that should jump off and hit you right between the eyes is being patently irrational, patently irrational. And if you haven't already noticed what that is, it's your pricing on your bronze cars, the ones that you own for the most amount of money that have the highest market day supply and the lowest retail volume are the cars that you are most proud of. As if the longer we keep them around, the better they get. Like, maybe they're fine wine or something, and we just simply know that's not true. And on the other end of the spectrum. These platinum cars, the ones that you owned for the best money, have the lowest market day supply and the highest retail volume. We're pricing them as if they're distressed merchandise like they needed to be gone yesterday. Now, this makes no sense to anybody. Nobody can look at this and put a rational explanation on it. I couldn't believe it when I saw it. So then I asked the fateful question. I said, I don't believe it. And they said, I assure you this is it. And I said, okay, then do this. I said, "Unravel it, unbundle it." This is what I would call an inverted pricing profile. Right. Inverted. Because what we would all agree what a rational pricing profile would look like is that we're most proud of our best cars and least proud of our worst cars, our bronze cars. So I said, unravel this, unbundle it. This represents roughly 12,000, some-odd dealers at the time and I said, tell me how many of these dealers have a rational pricing profile, one that is not inverted. And when they came back with the answer, I couldn't believe it. I could not believe it. They came back and they said "Dale, the 12,000 some-odd dealers this represents, there is not a single one, not one, not one that has a rational pricing profile." Well, today I know something that not one of you, not one of you today, if you saw your vehicles through this new investment lens, which we could show you, not one of you would have a rational pricing profile. How could this be? How could you ever say that every dealership does anything the same way alike, let alone something that makes no rational sense? How could this possibly be true?