Yes, Vince, this is Steve. I mean as we said in the prepared remarks, and as Ryan said, we obviously looked at the balance sheet dated March 31, we looked at our lease impairment reserves and our bad debt allowances and decided or made the determination that it was unlikely to assume that COVID would not have some impact on the future. And once you make that kind of -- from an accounting standpoint, you make that potential impairment analysis, then you move on to like, well, how do you quantify it and you have to probability weight different outcomes. We've got this government stimulus that we don't know how the immediate $1,200 checks and the actual checks are going to continue to play out. We've got enhanced unemployment benefits that currently last until the end of July, but there's -- we have no knowledge of whether those will be extended. So we just had to use all of the information that we had available to us, along with the actual results that we've been seeing kind of post the quarter and make our best determination about potential future outcomes. As Ryan said appropriately, we're not going to be right, right? Because we don't know the future. But at the time, our best estimates were to take these incremental reserves, and they were basically applied against the valuation of the lease merchandise that we have on the balance sheet and then the amount of the AR reserves or the AR that we have on the balance sheet. So it's not reflected as a percentage of how many customers we think are going to go bad or anything like that. It's just -- it's an overlay against the reserves that we would have had anyways based on the information we had available to us when we were making the decision. And then obviously, part of the reserve is a CECL reserve, which refers specifically to our Vive business, which is a second look credit provider that partners with banks, and we had to adopt CECL as of January 1. As you know there, we had about a -- historically, we ran about a mid-teens provision expense under the previous provisioning method, and after adopting CECL, we expected that we'd be kind of in the low to, call it, mid-20s, and we were after we took the day 1 retained earnings charge from the adoption of CECL. Because of macroeconomic factors related to that CECL calculation, we ended up taking additional reserves in the quarter that put us into the low 30s percent from a provision expense, and that is heavily weighted on future expectations of unemployment rates and GDP. So that was about $7 million of the additional incremental reserves that we took.