Richard A. Galanti
Management
First of all, in terms of -- we actually expected a little spike, simply because sometimes you -- even though we didn’t make any formal announcement over the change in policy, it was out there pretty quick. Looking at some local newspapers around the country, one headline that I remember that I like the most was “You better bring everything back now because things are changing” when in fact that’s not the case. Everything purchased prior is grandfathered in. According to Jim, he said he saw a little spike in the first couple of weeks. Less than expected and we’ve actually, just in terms of looking at actual returns in a given week versus sales that week as kind of an ongoing percentage, we have seen a very slight dip in that number in the last month. But that inarguably should not relate to the return policy change because very little, if you think about it, if today is May 31st, the earliest change we made in mid-February, so mid-February to mid-April, mid-May, we’re only 75 days. The most impacted item purchased is 75 or 80 days old. It can’t be -- it’s still -- anything out there that’s bring-backable is still bring-backable, if you will. Returnable I guess is a better word. So in that regard, I can’t explain why this month is a little different, other than maybe people think hey, they’ve got this new return policy, I can’t bring it back. So really that’s going to be more impactful over the next 12 to 18 months, I think. In terms of what we saw differently, keep in mind again historically, it was a fairly -- a more simplistic approach that didn’t delve down to the item. When we were doing our analysis initially earlier this year, making decisions on what we were going to do in terms of changing the return policy, we had IT drill it down to the item. Now, think about it -- even though they are computers and they are fast, it literally was a time cruncher and a massive eater of computer time, because literally we took every returned item and I tried to identify when it was returned -- I’m sorry, when it was originally bought. When we first started that, because o the time crunch, we literally had to look at the item and then look at all the previous sales from that member, identified member, virtually all cases -- most cases you had the identification of that member -- and then lag that. We originally started with data that was mid-05 to mid-06 data, because we could look back further there. Given the time crunch, we looked back over the course of the year and then extrapolated the last 10% or 12% of the return dollars, because most of it is in that first year. With that indicated, needless to say, at Q2 end was a big change in what we had historically booked. That’s what we booked and we booked to those numbers. Unfortunately, in accounting you have to book to data. Needless to say, these are estimate but we booked what we felt was the correct estimate. Now, what changed during the next 10 to 12 weeks? During the next 10 to 12 weeks, we completed that study -- not only completing it in terms of going back beyond a year but completing it in terms of looking at all data from the middle of ’06 to the end of Q306. Well, two things happened; the tail on that last 15% or so which we extrapolated over the course of a year and beyond looking backwards, had a longer tail on it than we anticipated, than we estimated. Secondly, lag data, even for the current 12 months looking backwards, so what was returned today looking back over the last 12 months, people in the last five quarters compared to the previous five quarters, so Q205 to ’06 versus Q306 to Q307, are taking longer to return things. Why, I can’t tell you but those things together added up to a big chunk. Now, why didn’t we know that? When you have a big chunk in front of you and you are looking at it and it is based on literally millions of items returned that are only a year-and-a-half old, and it was so bigger than the previous estimate, you figure that’s got to be correct. But more importantly, that original data, the first churn of that major study which we completed literally after Q2 end a few days before Q2 reporting, under accounting rules we had no choice, correctly so, to book according to what our best estimate was. We did so. During the next 10 to 12 weeks, we completed that. That’s what it is. Now, one of the questions I was asked is does this have any impact into the return policy change? It has a very miniscule impact on that simply because of the that that as of Q3 end, there are some purchases that have been made over the last 60 days that in the future won’t be returned because they will finally anniversary beyond their 90-day from when they were purchased. But that is a very small amount and really any impact from that you’ll start to see on an increasing basis going forward.