Overstock.com Inc (OSTK) 2006 Q3 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the third-quarter 2006 Overstock.com Inc. earnings conference call. At this time, all participants are in listen-only mode. We will be facilitating our question-and-answer session towards the end of today's conference. (OPERATOR INSTRUCTIONS). As a reminder, ladies and gentlemen, this conference is being recorded.

I would now like to turn the presentation over to your host for today's conference, Mr. David Chidester, Senior Vice President of Finance.

Thank you. Good morning, and welcome to Overstock.com's third-quarter 2006 conference call. Participating with me on the call today is Dr. Patrick Byrne, Chairman and CEO of Overstock.com, and Jason Lindsey, President and Chief Operating Officer.

Before I turn to my financial summary, please keep in mind that the following discussion and the responses to your questions reflect management's views as of today, November 6, 2006 only.

As you listen to the call, I encourage you to have our earnings release in front of you, since our financial results, detailed commentary and the CEO's letter to shareholders are included and will correspond with much of the discussion that follows.

As we share information today to help you better understand our business, it is important to keep in mind that we will make statements in the course of this conference call which state our intentions, hopes, beliefs, expectations or predictions of the future. These constitute forward-looking statements for the purposes of the Safe Harbor provisions under the Private Securities Litigation Reform within the meaning of Section 27-A of the Securities Act of 1933 and Section 21-E of the Securities Exchange Act of 1934. These forward-looking statements involve certain risks and uncertainties that could cause Overstock.com's actual results to differ materially from those projected in these forward-looking statements.

Overstock.com disclaims any intention or obligation to revise any forward-looking statements. Additional information concerning important factors that could cause our actual results to differ materially from those in the forward-looking statements is contained from time to time in documents that the Company files with the SEC, including but not limited to its most recent reports on Forms 10-K, 10-Q, 8-K and S-1.

With that, I will summarize our third-quarter financial results. Please note that all comparisons will be against our results from the same quarter last year, unless otherwise noted.

Total revenue was $159 million, a decrease of 6%. Gross margins were 14.5%, down 120 basis points from last year but up 10 basis points sequentially. Gross profit dollars declined to $23 million, down 14%. We continue to see considerable pressure on direct gross margins as we reduce and refine our inventory. Sales and marketing costs were 11.2% of sales, compared to 10.6% last year. Sequentially, marketing costs as a percent of sales increased 360 basis points as we increased both off-line and on-line spend in the third quarter.

Technology and G&A expenses combined were 18% of sales in the quarter, compared with 11% last year. Individually, G&A expense increased 27%. However, to get an apples-to apples comparison, excluding stock-based compensation, G&A was up 20% over last year. Technology expense doubled over last year, and just under half of that total expense relates to the depreciation, amortization and stock comp.

Total operating expenses increased 30% to $47 million, which equates to 29.5% of sales. This translated into a $24 million operating loss for the quarter or 15% of sales. Our net loss was $24.5 million, which equates to a loss of $1.19 per share.

Our cash balance at the end of the quarter was $39 million. We had no borrowings on our inventory line and $32 million of borrowing availability based on our quarter-end inventory balance.

Total inventory and prepaid inventory combined were $72 million, down almost $40 million from the $111 million of inventory and prepaid inventory that we had last year at this time.

Operating cash flows during the quarter were an inflow of just under $1 million and free cash flows were an outflow of $7 million. For the trailing 12 months, cash flows from operations were an outflow of $19 million and free cash flows were an outflow of $47 million.

I will also remind those on the call that updated quarterly financial and customer metrics will be posted to our Investor Relations web site following the conference call.

With that, I will turn the call over to Patrick.

Thank you, David. Jason, are you on the line?

Yes.

I'm going to be clicking through slides. Please go to Slide 2. There should be no delay. I do not have to read the Safe Harbor again.

I do have any e-mail from a Doug Griffin of Santa Fe, New Mexico shareholder suggesting that I say that while I have never really tried to make predictions, I've tried to just tell people what we knew, to say that because my predictions have not been very good recently -- I think that is true -- in the last year, year and a half -- or my expectations. We will not be providing guidance for the fourth quarter or beyond. I don't think I have ever actually provided official guidance, but I'm not even going to be providing unofficial suggestions.

What I will do is share my insights with you regarding the positive and negative factors I see which could impact our future results and leave it to you to estimate how things will unfold. That is precisely what we have tried to do in the past. Again, the letter you get this morning is pretty much a mildly sanitized letter of what I gave the Board not long ago.

So, again, no predictions but I am just going to try to be straight and call them as I see them, and let you know everything I know. Go to Slide 3, please.

These are growth rates, essentially since we went public. This is, of course, distressing. You see what has happened in the last year. We were staying up at least close to 170's, and then it has just been this horrible slide since then. I thought we had arrested it, and I thought the third quarter would at least hold its own and maybe edge better and then be better in the fourth. I'll get back to that in a second.

Internally, I can tell you we always scaled this business for $1 billion, so that was kind of our thought as to where we had to reach to at least make it a viable business. So this is distressing.

Next slide -- Slide 4 of 24. This is what has gone and the cause of it. These are our conversion rates. We are not showing you our absolute conversion rates. David, can I assume that these slides do not have a 20-second lag like they did in previous quarters? Is that what I was --?

Yes. They should not have a lag. Everyone is in control of their own slides, so, it is on their computer.

These are conversion rates and how they compare with the same weeks last year. You can really see what happened in the third quarter. It was not good this year in any case. It was not good. I think that if marketing is a three-legged stool, and off-line marketing and online marketing and everything we do internally to drive conversion -- I have to say that we really do not have that third leg. As much as we focused in our early years on figuring out the other two legs, that third leg has been our weak point, and it now has really come to bite us.

The third leg being internal?

Yes, the third leg being -- thank you, Jason. Jason speaks Patrick. Jason will interpret for the rest of this call.

So I have taken over internal marketing. I took over the last week of July there. I think there are clearly dramatic things we can be doing with internal marketing. Internal marketing again is everything related to conversion, customer retention, e-mail, site design, site testing and lay out and so forth. I have been on this since the 1st of -- really, the 1st of August.

I will tell you, it has been suggested to me, and I think that this is a fair suggestion, that I let the world know hey, the jihad is in other people's hands. It really is. It may be the case -- you know, I make these public appearances when asked. I have always been a bit of a -- you know, since the early days when we did not have a real marketing budget, I always just did PR, because it seemed like the easiest way to get our name out there. I get asked to do these radio shows now, an occasional appearance, and sometimes the invitations to the appearances are withdrawn and sometimes they are not. But it is the case that I am not really focused on the jihad. I'm not really crazy about doing interviews about it anymore. I think that it's a snowball and it has got its own momentum.

I am focused on this. I sit on the fourth floor. I run internal marketing. I've got a team of young folks reporting to me and we are figuring this out, and there is a lot we have gotten done in the last few months. But there is a long way to go. I think it is going to take us to March -- somewhere between March and June to really be crisp at this.

But this is our problem in a nutshell. If we were actually expecting to have by the beginning of third quarter some good increases in conversion over last year and we haven't -- in fact, obviously, it was a disaster through the third quarter. Well, I'll get into some of those things in more detail.

Slide 5, please. Some of the things we have done is I've just decided, although we finally did get Propeller live, and in some -- which is our own personalization engine -- and in some areas of the site, it actually got some nice lift. I have just decided it is too expensive and timely -- it takes too much time -- to do this ourselves, and I wanted to get something done by the fourth quarter. So we went out to three different companies and we had basically outsourced -- if you take personalization, you can break it up into a few different areas, and we did a lot of study and found what I think is the best in each of these areas.

There's basically two approaches to personalization. One is called Collaborative Filtering, one is called a Bayesian engine. ChoiceStream is really the last man standing in the whole personalization wars. They have a -- they're a great Massachusetts company.

But all we are getting live this year is a gift center. The gift center is done. It went through Q&A last week. It is finished. We're using it internally this week. We will probably start turning it on to some of our customers late in the week, and then you will start seeing it appear little places on the site in early next week, and then we will start -- just because A, we want to make sure it handles the stress of the traffic; and B, it gets smarter as people use it. But you're going to see us featuring this.

This is a terrific product. It is actually great. I have been using it the last week, and we are especially going to market to guys on the prospect of how would you like to do all of your Christmas shopping in 20 minutes? You come in, you tell us who you are shopping for, answer a few questions about him or her, your relationship to them, etc., and bam! It pops out a whole bunch of gift recommendations. They are good recommendations, and you can see that if you try it on yourself. So that is actually live internally now, and you will start seeing that at the end of the week, early next week.

Collaborative filtering is live. Aggregate Knowledge is a Kleiner Perkins company and I think it is relatively new, that has emerged. I cannot say enough good things about -- really about either of them. ChoiceStream has the most fantastic team to work with. The integration is very deep with ChoiceStream, but they were a fantastic team to work with.

Aggregate Knowledge is now doing recommendations on our product pages, and you will -- you see those -- if you go to our product pages, you will see, I think, what are good recommendations. They have been up a few weeks. They are providing a nice, measurable lift. The nice thing about Aggregate Knowledge is the integration is quite, quite easy.

Content management is a different subject altogether -- Kefta, the San Francisco company. Again, the IT team has had a -- which was on track with all of its projects this year -- they finish in early October, mid-October. I stepped in in early August or really late August and said I am throwing three more big logs into the wood chipper that we have got to get through this year. I will talk more about content management in the future. Kefta had very smart folks that I have been working with -- that we have been working with. We will tell you how it works out.

I will also mention that the ChoiceStream engine, which powers iTunes recommendations and Yahoo! gift center and a bunch of the big media sites, the ChoiceStream engine, we're going to in the first quarter roll out -- it is going to be doing product recommendations, and even doing some e-mail. Aggregate Knowledge is going to be -- by another 10 days from now, doing product recommendations not just on the product pages but on our other pages in our site, maybe as high as the homepage.

So we do have progress on personalization. These things are all either have turned on in the last two weeks or are turning on now. It will take at least a couple of weeks to get tuned in, but we actually see results on them now.

The only thing -- last thing I will say about this is this was a tremendous load on our IT department, who was just reaching the finish line on their projects for the year just so we could come through and put in three big implementations in the last few months.

Well, there is some good news. that is our brand recognition. We pumped an extra $6 million into marketing in the third quarter, and the main reason we did that is what we suspect is that there is more of a relationship than we previously knew between the brand recognition that you achieve in the third quarter and what happens later, what happens in the fourth. It seems that you tap the people all year who never come, but if you tap them enough and they come in in the Christmas season and start shopping, we have achieved what our -- three years ago, we set an internal goal that seemed fanciful of reaching the same brand recognition Amazon has. Well, our last study shows that we have done that, which I find kind of amazing. I will warn you, I am not sure I would accept these numbers at -- I'm not sure I believe these numbers. We use the same company and doing the exact -- we have used them for three years and asking the exact same question, so that there is no variation.

They did change their methods in this last survey, and they gave us a warning that they changed their methods and they computerized this and that. So I am not ready to swear on a stack of Bibles that these are the right numbers, because they did change their methods.

But if you go to Slide 7, this shows the difference between Amazon's brand recognition -- unprompted name recognition for Amazon and Overstock. So this would not be affected by a change in those methods. If anything, it would be more precise, because now it's -- they're [experienced] the same way. Here, they just ask people, where do you go online to get a good deal? I think the test is do they name you within the first three. Well, we have now essentially caught Amazon in how often we get named. So that was where a lot of that $6 million investment went. I am quite surprised that we actually caught them like this, and I do believe this chart maybe a little bit more than I believe the last chart. Just because the technique was consistent across both.

This is our net promoter score, which I have talked about in the previous couple conference calls. Again, according to the book, the average American company has an 8% net promoter score. Fred Reichheld identifies a superstar as companies in the, I think, the 35% to 70% range, something like that. Our net promoter score is back higher or as high as it was a year ago. You see that valley is where we collapsed from our internal problems. Now, Reichheld says in this book "The Ultimate Question" that this is the ultimate measure of how you are really going to do as a business.

So it seems odd to me, and I want to talk about this for a minute. If you had asked me at the beginning of the year, our plan was just focus on everything to do with fixing the customer experience. Even though it only dropped to 37%, I think that that was -- in the worst part of last year, that was partially a reflection of that we've already had a lot of -- that we had a reservoir of goodwill with our customers. We really did worse than that.

So Jason has spent this year fixing everything to do with the -- well, I would say Jason, just keep trying. Everything to do with the customer making sure they get their products in a timely way, managing the quality of the products, eliminating bad products, eliminating products that generate poor NPS. Stormy has run -- has took over customer care, and between them, this is to me a phenomenal achievement. This puts us back in the very highest reaches of American companies, in terms of our customer satisfaction. I know that sounds like -- I actually had trouble believing this statistic, that this had come back as much as it had, until last week -- the National Federation of Retailers and American Express released their survey.

There, they went out and asked -- I think it was 8000 or 18,000 people -- just some question like, what retailer in America gives you the best customer service? People named L.L. Bean, Nordstrom, Amazon and Overstock as the top four.

Slide 9.

Okay, sorry, thanks. Which to me is phenomenal. I grew up in New England, and I think of L.L. Bean -- I always thought of as just the guys who did it better than anybody did it. And we were not in the top 55 last year, evidently. So I am not quite sure how this happened, other than Jason and Stormy and the rest of the team -- the logistics and everything is working, the IT is working.

But Jason, do you want to --? I am going to go back to Slide 8 for a second. Do you have any -- do you want to comment on Slide 8 or 9?

Well we spent a lot of time over the last several conference calls talking about how our focus was making sure -- and I think we even walked through the scoring system that we developed internally for customer aggravation points, and how we had driven that down from when we first measured it from 4 million or 5 million aggravation points to now, the Overstock aggravation points are a couple thousand. So -- and the only way you do that is just being fanatical literally about every single order and making sure all 17 steps, from when they click by until they have it, are executed and they do not ever fall outside of our service windows, and the process just flows continually. And literally there would be 20 of us in a room looking at all the order flow over the past several days. We had a way of mechanistically watching it and making sure it all went through, and any time there was anything that ever hung up, it was not necessarily just about the order, although that was important. What was -- mattered, were there any cracks that things could fall into, and we just -- oh, for six months feel like we have plugged basically every one of them. You know, things break and power gets turned off and we are always making sure things flow through. But as far as just the system working, it works now, and when people order something they generally get it on time just like they were supposed to.

So I think we have largely, at least for now, solved the business problem of making sure that all these new systems can talk to each other, and that our order flow works.

If I had to say something I have learned between now -- and one more comment on Slide 8. You see a little dip there. That is actually when we went live was right now on the 1st of August and there was a mild dip, but that is probably the smallest dip of any implementation we've ever done.

What I would have thought at the beginning of the year was that if we could achieve this -- and this was our goal, to have this done by the end of the third quarter, which at the beginning of the year seemed impossible, but by the end of the third quarter to have this back where it had been before our troubles -- I would have thought okay, that meant sales were going to pick up at the same time that we were going to have a sluggish first or second quarter, but if we got this back and our customer satisfaction was so great, we would have sales pick up. And it didn't, and I do not know how to explain that.

There is a possibility that what it means is you have to get this to pick up and then you have to get through a Christmas season, and then everybody comes back and discovers and tells their friends and stuff, and then it picks up. Or maybe not. Maybe there is some factor I am not mentioning.

Any comment on that point, Jason?

No.

No, okay. I know you have an opinion, but -- okay, Slide I mentioned. To me, this is -- having developed a brand recognition that has essentially caught Amazon unprompted, having developed a net promoter score that is a superstar and a customer care -- people name us, volunteer us, in terms of having exquisite customer service, up there with Amazon, Nordstrom, L.L. Bean is something I would not have thought possible. But to me, it is what, to my mind, drives the intrinsic value of the business, and it was the thing we had to solve first. There's clearly a big problem left we have to solve in marketing -- that third leg of the stool. But this was an important focus for us, and I'm glad to see it worked.

Inventory turns -- Slide 10. This just shows we were getting the hang of things by 2004, kind of 2005 we -- I think we probably got a little bit -- well, we started going blind in 2005, as our old system, our old business intelligence systems, the homegrown ones were dying, and instead of keeping them up, we switched to these new [honking] systems that were not really ready yet. They are ready. Jason is running merchandising, doing a fabulous job, and you see us climbing back out of the ditch.

I think that there is one more -- as David pointed out, we are basically supporting the same level of sales we did last year, with 40% less inventory in prepaids. I think there is one more stair step, maybe not as big a stair step, but there is one more stair step in what we should be able to get out of our inventory.

Jason, since you are the guy who is doing it, do you have any comments?

No, I agree with everything you just said. I mean, we are basically -- if you look at Q3 to Q3, we are almost twice as efficient with our inventory now as we were then, and I do think there is still room for improvement.

You'll see us at the end of the fourth quarter, I think we should be about where are going to get. In other words, I don't -- well, there's always room for improvement, but what we have as our internal goal is something that I still think is achievable by the end of the fourth. Then I do not know if there's any dramatic improvement to come after that.

Slide 11. This is what is really going on with margins. Our partner margins are holding their own, and in the last year they got hurt, the same as everything else, by our internal problems. They have come back and they are holding their own and getting a little bit better, and that slight improvement is coming largely from -- we still handle returns and we do all of the customer service and we do the credit cards, and those are going to the gross margin and we are getting better at each of those.

The direct margin, what we call the core business, is diving, and it is, I think, going to -- well, I do not want to make predictions, but Jason, what do you want to say about the core?

Well, I think for another quarter, these lines going in opposite directions are probably going to continue as we get rid of the slowest of our moving inventory. And as we don't replace it, hopefully we do not have to go through that again. Then I think you'll see in Q1 it begin to come back quickly, and I think it will start to climb and hopefully catch our partner margins. I do not know if they can catch it in 2007, but I hope it could.

Boy, I think it should. I am going to put some meat on that, because we think that we can dramatically reduce inventory from here. But to move that amount of inventory, of core inventory, we are giving great deals, better than -- well, we are giving great deals and clearing a whole bunch of stuff out so we will end with extremely fresh inventory and a much smaller amount than you have ever seen us run with as a fraction of sales. But that is -- to do that, to clean out every nook and cranny of the warehouse is going to require no clearance prices on it. So that is why the margins are kind of hurt.

But we are really committed to getting -- the same enterprise data warehouse that we installed last year that sort of was part of the problem last year is actually working very well for us now. Because what we are able to get now is profitability on a per-SKU basis in a way -- with a granularity we never did. We can attribute returns cost, returns handling costs and how many times does a customer service agent have to pick up the phone to handle a problem with this kind of SKU and so on and so forth. So now that we get that with such great granularity, of course we find, as you always do, that there is far -- that the costs are not spread evenly. The costs are spread in certain SKUs -- are concentrated in certain SKUs, so as we -- and we realize we should be getting out of those and probably going to -- and going to partner, probably.

Jason has been managing that beautifully, and we are not putting new capital into this categories. We are replacing them with partner products. That is going to have effects throughout the Company. To have a warehouse instead of a warehouse handling 70,000 SKUs to have it handling the 10,000 SKUs that do the greater -- 10,000 or 15,000 SKUs that do the bulk of the profit is going to simplify so many different processes in the Company.

Our customer service will be cheaper, the warehouse handling will be cheaper. It is just going to be good all around, setting aside that we are not going to have capital tied up in products that we see now, are not really the right ones for us to be in.

Slide 12, please. Gross profit less marketing expense, contribution margin. This is again what I said last year I believe can be a 10% -- what I said last quarter I believe can be a 10% in the first quarter. I am not retracting that. I am not making any more statements about the future, but I am not retracting that.

Jason, do you have anything to say about this?

No. I agree. I think it is obviously made up of two numbers. Gross margins -- so gross margins need to go up and marketing expense, and marketing expense needs to go down. I think if they both go in the correct direction, which I anticipate and hope they will in '07, that 10% is not out of the question.

First quarter? Okay. We are going to -- well, okay, next slide -- 13. This is tech expense, and what you see is the dark blue is the expense not included in the depreciation, and the blue -- the light blue is the depreciation, and there is a bit of stock option expense. So what you see is really that it is flattening. It is not completely flattening, and it probably should be flat, but it is flattening. We made our big stair step up, and we have got to live with that blue bar for next year. Then in the following year, we start seeing a bunch of that flush away. We have $35 million, $40 million of depreciation this year that should start -- in not next year, but the year after -- start drying up on us. We really have -- we have built such capacity in our systems now, I really do not think we should be doing very much new CapEx for a long time in technology.

Slide 14 -- these are our G&A. You actually see our G&A -- again, it is the dark blue is the cash and then there is a bit of depreciation. That is actually flat or coming down.

Slide 15 -- this is just fixed asset turns and measure of efficiency. You see -- of course, I always like to see how Amazon is doing something. We actually had, until our -- I mean, Amazon is a scale business that is going to do $10 million this year. Even though they are in scale, we actually ran at the higher -- this is just saying we ran lean. We ran very lean until last year. Then last year, we went out and built all this -- I think we started last year. I should check this number, but I think we started -- I mean, you can check for yourself on our balance sheets. I think we started last year with fixed assets of about $11 million, and -- on which we were trying to do $800 million to $1 billion. That is what just snapped on us, and last year we realized we had to go out and spend a whole bunch of money.

We did it, unfortunately, just when -- of course, if we had known the growth was going to stop, we probably could have eaten through on the systems we had. But if we had eaten through on the systems we had, we would probably still be growing.

The next slide is just the same comparison with Amazon when they were about our size.

Slide 17 -- Q&A. We ask people to send in questions. There are some that have come in by e-mail. Some came in, I guess, last week. Kevin [Moon] put these together.

Jason, do you have anything else to say on the formal presentation before we go to --?

No.

Okay, Slide 18. Customer service logistics -- how are your customer satisfaction scores trending?

We covered state of warehouse operations and capacity. I think it's -- I cannot believe how well our warehouse is now running. They made some changes this year. Steve Tryon is managing it beautifully.

Jason, do you want to add to that? I mean I just --

Yes, it is running really well. In fact, we did a couple of tests where we basically gave everybody a long weekend off, made everybody take four days off and let the orders just pile up on us, and then made everybody work for -- they still got the same amount of hours in a week, but we just stress test the system, including holding all the orders in abeyance here and so we could push them all through all of our systems all at once.

Last year, we pushed through more orders than we ever did at our peak season last year, and last year, even with slightly less orders than we pushed through during this test, our systems were red-lining, and in some cases falling behind. This time through that stress test, our systems were purring, and I do not think we ever got over 10% CPU usage on our IT systems. As far as the warehouse, Steve Tryon, who runs the warehouse, said at the end of the day, they all looked at each other and said, was that 20,000 lines -- or whatever the number was. It just did not feel like it. They feel like they can do several times that.

So I think our logistics systems of what they can pick at the warehouse and the training that they have gone through and the training that they have at different shifts, as opposed to just having to throw more temporary labor at it, we have never been in the situation we are currently logistics.

Then as far as our IT systems, obviously that they did really well through that test. So I could not feel better about capacity right now.

Yes. I am actually going to throw some numbers on that. Last year, with our new ERP systems, which were not yet dialed in, not tuned in, we maxed at about 6000 orders processing capacity an hour just on our IT systems.

Unfortunately, we had an hour last year and the peak week that we actually had that many orders or maybe even more. We now have in sort of theory -- well, so we spent six months this year dialing it in, getting the right switches set in the database. There is like about 2000 switches you have got to get just right. We got them and we now have -- in simulation, we have a capacity of 100,000 line an hour, or 16 times what we did a year ago. And we actually -- when we did this -- when simulated Christmas week, by what Jason describes, holding orders for a Friday, Saturday, Sunday or Thursday afternoon, and then releasing them all at once, we started to process 56,000 orders in an hour. So it has capacity for at least eight to maybe 15, 16 times our current size.

Our warehouse is has been changed, simplified. An old friend of mine, James Joyce, came in and helped us this year, and a logistics specialist. He is really -- he and Steve Tryon have gotten the warehouse just humming.

Our variable costs -- and we have people picking, on average, I think much more than twice as much as the average picker was able to pick in an hour last year. Remember, we are not picking a million identically-sized CDs. You're picking all kinds of very different-sized products, and closeout has a lot of problems in receiving. They have to repackage things and all kinds of stuff. So it is a difficult problem, but we actually have -- without -- and it is something you really cannot automate, but just through changes within the warehouse, they have -- we used to pick an average of about 35 lines per person. I think we have got many people picking up over 120 now. I am sure the average is more than double the 35.

Slide 19 out of 24. Changes you made to the merchandising strategy. Jason, how much do you want to talk about that?

Well, we have talked quite a bit about it already. The reality is we have made a lot of changes in the merchandising strategy. Some are personnel changes. Also we have kind organized all our reporting differently, so that every buyer -- every store that you see on our website has an income statement now, and so we can see the profitability by store, which we have never seen before.

Including capital tied up, and we charge them for the capital.

Yes, and we are allocating more capital to those who have a higher return on capital and taking capital away from those who have a lower return on capital. So I think it is going really well. Unfortunately, you see margins continue to tick down. That does not sound like it is going very well, but I think you have to get rid of everything in your warehouse that you do not want and that you do not want to reorder to get to a more efficient state. So margins that have come down -- our core margins, as you saw several slides ago -- have come down dramatically, and I think we still have another quarter of that, and I think next year you'll see that trend completely reverse itself.

I think that in fairness, we actually had -- when we were much smaller, we had the same basic idea going. I think we probably outgrew our capacity to look at things in that kind of way, and for a couple of years. And then when we got the [EDW] tuned in and we can now look at it by chimney and by capital and by all these other costs and getting them allocated, it is just very eye-opening that the things that you maybe thought were profitable were not so profitable.

And it feels like the same thing all over again.

Yes, it feels like something -- you mean in the sense of something we did (multiple speakers) years ago?

Yes, I mean, as efficient as we got at it three or four years ago, it feels like obviously we're not even close to that efficient again, but it feels like we're solving the same problem we thought we had solved three or four years ago, and just lost visibility into how to maintain it.

On about ten times the size. Yes, that is fair.

Contribution margin, 10% next year. Well obviously, I will just tell you how I get there. I think in terms of gross margins at 17% and marketing at 7%. That can only happen if we move through the last of this inventory that we now identify as not really being as profitable as we thought it was.

So that is how I think we get there, and yes, that is based on focusing its -- to tell you the truth, this 20% generating 80% in our case it's even more skewed than the [straight O] principle.

But Jason, do you think -- do you have any comments to number two?

No, I get there -- it's funny, I get there by saying we spent 5% or 6% on marketing and having 15% or 16% gross margins, but either one of them would get you there.

Yes. New compensation incentives for buyers -- what do you want to say about that?

Well, I think the new P&L by buyer is allowing us to see profitability by buyer in a way we never have. I think as we organize our compensation incentives around those, it makes the buyers act like business owners, and I think it is working really well.

I think next year, we're going to be even more -- have a -- I think, a good idea about how to tie their compensation. Their compensation as a group was tied to results before, their bonus as a group. With sort of a tip of the hat to individual performance, but it couldn't be any more granular than the data. It is now to the point where in the new year, we could have a very specific compensation plan tied to every buyer, every tab, basically. I do not even want to call them buyers, because some of them are things other than buyers. But specific compensation plan tied to each tab.

Page 20 -- analytics internal marketing, conversion rates, progress on initiatives. The problem of the third quarter is that conversion rates were down 25% or so from last year, as that slide showed earlier. Put differently, if they were not down, I would have expected to have been growing 15% or so by the third quarter, and we did not, and it is all the conversion. Everything else, the expenses internally, the buying, -- everything seems to be managed. We have screwed back down. But the conversion rate, no.

I am not saying it is solved now, but we are buried in the right analytics to solve that. And we built a team, we have built a solid team of a lot of smart people in a small number of very experienced experts in different areas and these different teams that make up the whole conversion system all report to me directly. I am running that. It has been four or five years -- or five years since I have been involved in that part of the business. There is a lot to learn, but there is -- I mean, things have changed in five years. But the technology is better.

On the new Web site, there have been some design changes. You are seeing improved on site conversion rates. Well, as that slide I showed you near the beginning demonstrates, yes, we were -- we certainly have, since the middle of the second quarter -- middle of the third quarter, we're starting to see some things come in. We are definitely redesigning aspects of our site. There's a lot of testing, A.B. testing, and multivariate testing going on these days. Of different corners of the site, we are finding lifts -- a lot of small lifts, but we are really at the beginning of this process. I think that we really dragged behind the world on this, frankly. You'll see some more changes before the end of the year. You'll see more changes over the next two weeks are going through Q&A and some fairly significant changes that simplify the checkout process and things like that.

Slide 21 -- sales and marketing. Marketing plan for Q4 -- I don't think we will be able to get down to the 7% this quarter, because, there are some -- clearly other than the brand seems to have soared, we did not get our money's worth for the extra $6 million we put into it in the third quarter. But we did other than just got the great brand recognition. Another way of saying that is we should be able to cut, I think, at this point a lot of marketing. I can be honest.

There is just a piece of marketing that I am realizing now we completely missed the boat on. I do not want to reveal what it is for competitive reasons, but there's definitely a big chunk of marketing that we have completely missed the boat on for years. And we are definitely making improvements in it. We see on Friday -- we saw some big results. I mean, not in a way that would move the needle overall for the Company, but we definitely see results based on what we have done in the last three months. But there's a long way to go. As I say, I think it will take us at least until March to June to even get that reasonably up to speed.

What that means for Q4 is, though, that we are going to be shrinking our budget down from the 11%, 11.5% or whatever. We're not going to be able to get it down to 7%, just because there are some things into which we are locked.

As I said in my letter, some things are changing the online spend game rather dramatically at the portals and such. The opportunities are not there. Actually, a big part of what happened in -- well, some significant part of what happened in Q3 is last year, we had some portal deals that died at the beginning of July this year. The portals -- the MSN's, the Yahoo!'s and AOL's -- have changed their strategy, and some of the things that were available in the past are just not even available. They were some of our best campaigns.

Has the campaign begun yet in earnest?

Yes. This weekend, we started our Christmas spot on TV. How do you keep -- how do you spend --? I said how I intended to keep sales and marketing in 7%. I expect to get it towards there, but we're not going to be able to get all the way there, because there are some contracts that are already fixed that we cannot get out of.

How do you think about new customer growth, repeat customer growth?

I would rather answer that at the end of the year. I think we will get a lot of new customers starting in the fourth quarter. Other than that, I do not really want to quantify this.

How do you view the overall marketing landscape as we move into the Q4 holiday season, graphic ad inventory and so forth? Well, what has happened, there is a lot of -- I thought that we had reached the high water mark a year, or year and a half ago. That is the money that was flooding in from off-line advertising had reached its apogee, and the prices were not going to keep going up.

Prices -- I guess people are just still realizing it is more effective than some -- than a lot of the off-line advertising. So, for example, as I say, the portals have changed their strategy. Search key words are areas that we did have little niches before, have profitability. Other people have come in and found them, and the arbitrage has gone away. So that game -- I do not think there is a lot to be gotten out of that, in that world anymore.

Jason, do you have any comments on sales and marketing?

No.

Okay, Slide 22. Revenue growth through Q4 in 2007. I'm not going to talk about 2007. Even Q4. All I can say is I'm a little surprised.

Jason, why don't you take this first?

I do not know that we have anything to say about that. I mean, we're obviously very disappointed in how revenue growth is trending. The good thing -- the only good news I see about all that is just that we are diving into it into an area where we really have been derelict in looking at. We're just drowning in low-hanging fruit. It is not like there's nothing there to be done. It is quite the opposite. It is there is so much there to be done, you are overwhelmed by it. Every single night, when the IT -- when we roll new code to fix things, it is just how many things can we push through QA in a night? There's just so many things that we think and that we have tested that do move the needle one tiny blip at a time, but move the needle, that is just sorting through them and trying to prioritize and test all of these things and get the biggest winners to the top and get them pushed through. So I do think there is a lot of upside in conversion. That is the real positive I see.

Right. We have really been -- we are discovering how much we missed here. I know that other people focused on this first, and maybe they -- I am sure that was smarter. They focused on this and got this drilled in.

Issues of site design, personalization and checkout and ease-of-use and such is what we are focusing on now, and there's -- well, as well, like Jason says, I am surprised that there is such low-hanging fruit that has escaped our notice before.

Slowing growth changed the economics of the business?

Yes, it has. As we always -- I am not even sure I have talked about this publicly, but we always thought of sort of the $1 billion mark, charging to the $1 billion mark and then that was the scale. Being stuck at the $800 million mark presents a problem.

Slide 23 -- Gross profit. You said that direct gross profit margin should be higher than partners once you clear out some of the lower. Yes.

When do you expect that gap to narrow? Jason.

Q1.

Q1. What degree are improved return rates going to help improve gross margins versus better product mix?

I think right now, the reason the product mix affects our margins so much is because there is such a huge difference between the gross margins of each. As you look forward to next year, when we think that the gap will be closed dramatically, the change in mix will not be that big of a deal. As far as our return rates, I do not think there's a whole bunch -- I do not think there is a lot left on the table to be gained there. I think there is a lot left to do --

You mean -- are you talking about product return rates or customer return rates? When you say you do no think there's much to be gained in return --?

I'm talking about customer return rates. Well, what is the difference?

Well, return rates in terms of how often people return something that they order?

No, I am talking about products coming back to us, whether they are the product -- I mean, the products coming back to us from customers. I am not sure if that is what this question is asking.

I think I'm with you on that. Customer return rates -- we're not really disclosing anymore, but they are higher than they were before, but that is partially just when you slow -- when you're growing 100%, it is impossible to have 70% of it be customer return -- return customers. Now that we are slowing, and now that we have dramatically slowed, yes, the return rate is quite a bit higher. The number is quite a bit higher.

Slide 24 -- profitability. Kevin had to slip this in on us. Jason, when are we going to become profitable?

I do not know the answer to that. I wish I did. I think that in our history, we always thought whenever growth slowed, we would become dramatically profitable. That was because we had such a small expense structure, and I think just about any time during our history, you could have slowed growth and we would have become wildly profitable on our tiny expense structure. I think last year, we flipped the switch and changed our expense structure dramatically, as you can see in all these slides that have shown how expenses stair stepped up, and that simultaneously happened with our decrease in growth rates, which put us in the bind that we are in.

I think that our growth needs to come back some, needs to come up some, and our expenses need to come down some, and when they do, there is a very valuable business buried here in the middle. With the brand recognition of Amazon, it is an extremely high traffic web site, and we have a huge section of our traffic that is extremely profitable and generates a gigantic piece of our gross margin. Unfortunately, attached to it, we have a wildly unprofitable business attached to it that causes the whole thing to lose money. That is why you see us -- or at least you see me -- talking only about simplifying our business and getting rid of all of these things that cause expenses throughout our entire business operations that is unprofitable.

So as we get out of that, what you see us doing, I think, really quickly this year, you will see the business become much more profitable. Exactly when the entire business will turn profitable -- I do not have an answer.

The way to get out of that -- those unprofitable businesses, in our case is, you just stop putting capital into those tabs. We actually have it at the SKU and the sub-cat level. A lot of this is being done at the sub-category level. But some businesses, you can get out of by just -- some businesses, you just close. In our case, the way to get out of a business is you just stop buying that product.

And then remove the expenses associated with all that additional product you do not have anymore.

Right. And then, you -- but you have this period of three to six months where you're just going to be flushing through that product and getting it out of your warehouse. Even simplifying -- think how that helps the warehouse, that you do not have all these different types of products that you can reduce them. We built this warehouse mezzanine I was talking about six or nine months ago, and that has dramatically helped us. But that only works on because we are focusing on the very high-velocity products that we have found buried underneath in all the data. So I am with you, Jason.

Yes, and as you get out of those, you then realize well, there's a whole bunch of other expenses associated with that, with holding that inventory and dusting it and lighting it and counting it and advertising it and buying it and so on and so forth that you really can move -- -- they can leave your system along with the inventory.

That's the last of our slides. Let me see. I do so have some mail-in, and see if there's any -- I believe people e-mailed Kevin these questions over the last few days. Some came in this morning that did not make the slide deck.

Have you seen partners defecting (indiscernible) commission rates, eBay and eBay Express?

No, I don't. It is a big world, and no, we have not seen any defection or anything like that to eBay.

Do you have any insight into whether your partners are experiencing gross margin contraction? No. You're seeing in your direct business.

Our contraction in the direct business, again, isn't from buying, like it is getting harder to buy. It is just that we are saying hey, here's $10 million to $20 million in inventory that we do not need, and we're going to chop the price and make sure it all moves out.

None of these questions -- there is a lot of questions about new customers that are women, from Jim Krueger, repeat customers that are women. We do not really want to get into -- maybe at the end of the year, we will go back and give sort of an overview of what has happened with new and repeat and so forth.

Do I still consider Club O to be the best loyalty program on the net?

I do not know enough about others, other than Amazon. I think it is well-designed, and I think it works very well for us. Our Club O program works beautifully for us. So let's get -- maybe at the end of the year, we will go back and have some charts about repeat revenue and repeat customers.

Arnie Olson sent me something saying please tell the world that you are not fighting a jihad every day.

I am not fighting a jihad every day. I do not know what to do if a big radio station says hey, why don't you come on and talk to 50,000 people about it? It seems like a nice -- it seems like a cheap way to reach 50,000 people and also get them to hear about Overstock. I am not sure I would not do those anymore, but no, I do not like flying all over the country and meeting the lawyers and doing that. I am really not doing that anymore. I am here at Overstock and focused on this.

Let's see. Going to my letter real quick -- I talked about -- Furnace Brook -- this was a major victory. Jonathan Johnson, our head of legal affairs, this is sort of an infamous patent [control], and this got drop-kicked out of court. That is a big deal.

And then, Page 4, what does the future look like, or NPS score? Just while I was waiting to come on this morning, I turned on the TV and there they are on "Good Morning America" talking about us. They are talking about how if you want to buy high-end apparel, go to Overstock. They did not even mention anyone else. They just talked about -- you go to Overstock. They showed our site. So something has happened, in the sense that we have become this household name, and we have great associations in terms of customer satisfaction and customer service. We just have to find a way to capitalize on that.

Okay, well that was a long talk. I did not mean to use the whole hour. On the other hand, it is probably a pretty efficient use of time, because we got through a lot of detail. Why don't we go to questions? I will stay on and answer questions for -- Jason?

Sure.

Okay. I'm getting handed a list here. Nat Schindler from Piper. Nat Schindler?

Just a little question about the conversion rates. One of the things you have been discussing a lot about is on the marketing side, but is it possibly more a merchandising issue? From my personal experience, I go and look at the site and look for shoes, and end up finding five or six styles I like and none of them in my size. Without investing in the inventory, how can you have conversions or drive conversion rate?

Well, as we are choosing which inventory to invest in, that is one of the issues that is being addressed. If there are shoes that we know we can buy that are going to turn quickly, then we will continue to buy them. But maybe holding onto a pair of shoes that is going to sell just in case and holding onto it for a year and a half might not be a good business.

In addition, the way to get at that problem, I know how hard our site has been to shop for, for apparel, for example, because you have had to go in and you find a pair shoes you like and you have to drill down and you find out what size they are in, if they have them in your size. Well, that is one of the things that changed about three weeks ago.

We now have guided navigation in our search, and so you can come in and you can search by -- you go into to the apparel, you go into shoes, and you say I am a size 7. It shows, okay, we have got 5,000 shoes or 2,000 shoes -- types of shoes, and then you say I am a size 7 and it comes back and says, okay, we have got 300 size 7's -- 300 models. Then you say I want brown loafers, and you can click your way through. It is called guided navigation. It is not new. It has been around a couple years, but it has really just gotten mainstream in the last year or two, and we have just got it turned on in our site -- oh, probably in mid-September.

So it is -- to your specific question, it may be easier to find shoes that you like because of that feature now, and so we are solving, with technology, your problem. But also, Jason is going to be solving your product problem by if there's enough people like you, that want a certain type of shoes, that is going to show up in our numbers and that is where we will be investing capital. We just will not be investing it anymore doing -- I mean, we were being way too liberal with what we were willing to test and what we were willing to buy, without the fundamental analysis that would justify it.

Jason, since this is your world, do you want to --?

Well, we hear that a lot, and the answer to the problem is hopefully guided navigation. So we hear over and over, people search and search and search, finally find something that they like, and then it may or may not be in your size, when in fact at the very beginning, you did not have to search and search and search. You just said, show me all your size 9 shoes because I am a size 9, and then you looked through them and picked the ones you liked. It seems to be a much better shopping experience and a much more efficient use of capital.

So the truth is where people go and what people want to find, we do need to have inventory there. It turns out, if you look across all our categories, it is not just one category where we make all our money. The 20% or less than 20% that does most of our gross margin, believe it or not, really does have a nice smattering across all of our stores, and then places and subcategories where we don't want to be in, we can back-fill it with partner. It is just a much better business model, so --

Yes, it is kind of funny. In a way, you can think of us as eliminating 75% of our business. But it is not like we are getting rid of 75% of the tabs across the top. Within each tab, each tab is a store. Within each store, there's departments, categories, sub-categories and then SKUs. It is more like we are eliminating maybe 75% of the SKUs that we are buying. So, in answer to your question, I just went to apparel, clicked men's apparel. I do not know what size you are, but for example, you click 11 and 12 -- there's 186 shoes. So it just should make it much easier to shop for shoes.

By the way, our search design, I hate. If you go to our search page, I hate. By the way, there has been a tremendous change in the last few years. Four years ago, 21% or 22% of the people who came to our site searched. Now, sort of 38% to 40% on any given day. Call that the Google effect. People have gotten used to searching, and searchers convert much higher than non-searchers.

So you are going to see us rolling tomorrow night a brand-new design for our search results. I think this is very confusing, the one that you see if you go to our site now. The new one rolls tomorrow night, and it is much closer to, say, Backcountry Store in how it lays products out. But if you use our drill-down search, I'm hoping you can find --

Actually, I did do that, and I'd say take a look at it, because of the five shoes in the size 13 that came back in this last time I went, three of them were not in my size, were not in the 13 or above.

But on a related note, you make a lot of comparisons of the business to Amazon, which should be described as kind of a broad retailer. They have -- if you want to buy a mixer, as an example, they have 45 different colors of that mixer in stock and will ship it to you. You are more of a Costco. I wonder -- the marketing seems to be pushing -- message seems to be more similar to what I would expect for Amazon than for Costco, where I'm a -- look for a deal, I would come out to look to Overstock just to find something that was a good price. But that is a browsing shopper, versus a go to Google, search, click on Kenneth Cole shoes, click and come to site and not find it. So I am wondering whether or not the marketing message is different from what the business is.

Because we are marketing it as too broad a -- we are marketing it too close to Amazon to say that we are -- well, that is a point, although -- do I have you right, that you are saying that we are marketing ourselves too much to be like Amazon, rather than come here for just the best price on a much narrower range?

Yes. I mean, like browse and find great products cheap, versus come here and buy whatever it is you're looking for, because that isn't what -- you're more like Costco. I go to Costco all the time to look for stuff that just happens to be cheap, but then I am not particularly shopping for it at the time.

Well, that is a good point, and that is how we started off. We used to think of ourselves as a treasure hunt, and we saw, like I say, 80% of people browse, and they had the same conversion rate as searching. That has changed. In the last few years, we're down to 60% browse.

We used to focus on how do you enhance the treasure hunt atmosphere of the site, so we have these things like "just about to sell out" or "just sold out" or things like that, to reinforce the treasure hunt. But I am not sure that the world sees -- consumers' behavior is now indicating that they do not really come and browse. At least 40% come and search, and they are the ones who really buy. But -- well, you make a point. Your point does not fall on deaf ears.

And we had somebody -- could I see that list again, please? Frank Gristina, Avondale.

Obviously, the direct margin -- that is really what we are waiting to turn, to prove this model works. I was just curious, maybe Jason, if you could put some parameters around how much inventory you flushed in the quarter, maybe how much is left? Or if you cannot do that, give us an idea of when you liquidate liquidation inventory, what kind of gross margins are you getting on that stuff that you just want out of the warehouse? So if you could put any detail on that, it might help us to figure out where -- I guess what I am trying to get at is how do we know you can get to 17? Have you done some study that tells you if you are not selling the stuff with 0% gross margin or negative gross margin, you get 17%?

Yes.

Actually -- do you want to answer that question? Or is that all -- you want answer that question more fully, don't you?

Well, yes, we did run a test, and we know we can get to 17%. We know we can get north of 17% when we do not have the things we do not want on the site not there.

What kind of margins, when you are liquidating, are you getting, if you are able to give that to us?

Yes, well, unfortunately, they are all over the board. There is $10 million to $20 million worth of stuff in our warehouse now that we do not ever want to have in our warehouse again, and we're clearing it out nicely. Some of the stuff is negative, and some of the stuff is fairly profitable. We just do not want to be in that category anymore for a lot of --

We have never priced it correctly, to be honest. I mean, some of the stuff had 30% margins and had not moved, and that was just a bad pricing decision by somebody.

Then last question. In terms of the conversion, it is possible that it is what the previous caller was mentioning, that the marketing is misaligned. But is it also possible that competition has changed year over year, and that you are doomed to deal with low conversion because you're dealing in competitive spaces? Has it changed that much, or do you really think it is just something on your end that you can fix with some hard work and hopefully not more IT investment?

Well, go to slide four, if you would then, please, if you can still pull up the slide deck. If you go to slide four, this does not look to me like some sea change out there in the world. This is just we were doing something badly. It finally caught up and bit us badly, and we are fixing it and it is going to take another six, seven months to get it good.

Things are changing in the world, but they have not changed that much. They have not changed enough -- when we drop 0.6 or 0.7, and you are only converting 2.5, 2.6, 2.8 -- you're losing a quarter of your business right there. But there are changes in the world. There are changes in especially what is going on in online marketing. There's a lot of changes. But there's an area that -- I do not want to give it away, but there is just an area we have been in kindergarten. In a couple of areas we really did, I think, figure things out early. But there's an area we are discovering that we are absolutely in kindergarten about. So we have got the right -- we have got some smart people in, and we have got some experts in to help us develop that area.

Just one other thing about product selection. I mean, more than half the world still does browse, even though it has shifted dramatically towards search, and somewhere, product selection somewhere between one product and 10 million products, or whatever the infinite number of products is, is the perfect amount to have. Not only did we test about removing certain items from our site for a while and seeing what our gross margins could be, we also test the amount of products on our site, and we found that we have too many products for the browser. So our selection has continued to go up over the last couple of years, and in fact it has gone up so much that is to the point of clutter to a lot of our customers.

So yes, when we do not have the right size shoe, you are unhappy. But you're also unhappy if you just have too much junk to sift through. So there is a trade-off there, and we do think by simplifying the business, there are wins to be gained in just minimizing clutter.

In fact, that was actually pointed out to me a year or two ago by some guys like I think Scott Devitt, who is going to be up next, and maybe you, Frank, and [Krevlin] in New York that you guys have gone from treasure hunt to being you're just too broad. Jason has been pounding on this all year. We did some tests, and it is true, when we took off products -- we took some sub-categories and we just stripped out all the products we didn't want to have there, and to see what happened, and actually, conversion went up.

We realized that we were -- first of all, our search has not been good. Mostly, it is our search -- our search algorithms is good, a company called [Adecca]. The engine is good, but our layout, I think, has been terrible and it has been very awkward to search. Again, a lot of that changes tomorrow night.

But Jason has been saying, look at the clutter. That is what is so distracting. So we have tested it, and we were even testing things like how many products do you see per page? 24, 60, 100, or whatever -- and trying to find the optimal settings. But I think that in the curve Jason just described, we were too far down that curve. When you've got 13 pages of men's brown loafers or something, then that is a problem. People do not click all the way in, and so forth.

So there is some optimum point, and I think we were past it, besides which, it is not like we are not going to have the products. In some areas, we're going to just replace all our core products and partner products. It turns out there are -- in all these little niches, there are partners who are more efficient, probably better buyers in their little niche, and more efficient. Some of them, it's just because it's a guy working out of his garage or something, and he is very, very efficient at one little niche in his handling and everything else. So a lot of the stuff that we say we're getting out of, it isn't like you're going to see necessarily less of a selection, although we test different areas -- some places you will. We intend to replace a lot of it with partners. We have no shortage of partners who want to get up on our site.

Operator, I think there is one more question, but you wanted to make an announcement.

(OPERATOR INSTRUCTIONS).

I have a note that Scott Devitt is on the phone.

We had, I guess SKU counts are down year every year, conversion rate is down and inventory is down, and then conversely, marketing spending -- both the customer acquisition costs and in just general percentage terms, I think it is the highest rate since first quarter of 2003. So I am wondering if there is a correlation between the reduced SKU count, the higher marketing spend, and as you continue to reduce the SKUs, if you think marketing spending in percentage terms is going to remain higher than this 7% number, driven by the less selection on the site.

I am not sure that we really have a significantly reduced SKU count yet, or at least in the third quarter. It is coming down now. Jason, do you --?

Yes, I do not know what information you're seeing it coming down, but --

It has not really come down yet.

How about then just focus on the absolute inventory and conversion rates, and how that maybe is impacting the marketing spend?

I really do not think -- I could see why you would say jeez, if selection is getting worse and worse, they are having to push harder and harder on the marketing throttle, and they are still not able to get conversion up, I think it is really two very different mechanisms. I really do not think -- I am not arguing with you. You could be right. But I do not see that as what is driving the conversion problem. The conversion problem is being driven by -- it is not across the board. It is actually up in the -- it is even up in the third quarter, in most segments of our business.

There's two segments of our traffic that account for a lot of our traffic, which is where it has completely fallen off the cliff. I mean, it fell from 2.8 to 1.5. That is 40% of our traffic.

So I do not -- if it really were being driven by SKU count and stuff, I think you would've seen a more broad-based decline. What we have is two segments that have gone unnoticed, unmanaged, unanalyzed that we are now just focused on fixing and getting it back. Jason, what do you --?

Agreed. I agree.

Do think that a reduced SKU count is what's -- anything to do with this conversion problem?

I do not. I do not. I think that the first set of SKUs that you're going to see come off of our site -- and hopefully, it continues to -- our SKU count does come down, but it is mostly just stuff that is not really selling at all anyway. So it is just removing clutter. I do not think it is a problem with SKU count.

But that is a good hypothesis. I mean, I am not --

Well, if I could just follow up in turns of the marketing spend. Maybe sequentially speaking, what component of the increase was driven by incremental off-line spend versus online spend?

On a percentage basis, much higher on the incremental off-line. Much higher. I think that we should be spending 7%. Jason would even argue, I think, a couple points less than that. We should not be spending more than 7%. We cannot get down to that this quarter, but we can by the first quarter. We just can't get down to it, because we are in contracts that run out this year. So we cannot -- but we can get part of the way there, but we cannot get all the way there this quarter. Then after that, I do think 7% should be a constitutional principle, [as a cap]. Jason, any other comment or answer to Scott?

No.

[Jim Krueger] has come on. Let's make this the last question. It is 10:20.

Patrick, please talk about employee retention and changes in management.

That is a good question. Well, I will mention -- I am not sure if I have mentioned this publicly before, but Tad will be leaving at the end of this year. Tad got married to a lovely woman in New York and they're moving back to New York. Tad is the last of our -- we have lost some of our depth on our bench at the senior level. [Cami] went back to Stanford. Holly went back to working on (indiscernible) at a brokerage. Jason, what do you -- my thoughts are colored emotions here. What do you have to say?

Well, I am encouraged with the team that we have. If you look at just the different sections of the business and who is running them, I am quite encouraged. I think everything to do with our operations is improved and it is as good as it has ever been. Our whole merchandising team -- I think we are undergoing a lot of change there, and it is all about simplifying the business, which I think is a really good change.

Another big area where we have failed is our conversion rates and the whole study that has been going on of how do you make your website convert as high as it can. I think Patrick's completely dove into that issue, and it sounds like he is just drowning in low-hanging fruit. So if you look at the different sections of the business and who was running them and how they are doing, I think I am quite pleased with where we are at.

Yes, I would say, Jim, you have hit an interesting point. We actually have a -- on the one hand, we are -- it is not just us, the names that you know and that I talk about. I would say that the level below that was probably -- we were too much running around doing too many things ourselves. We have really built up a solid junior executive and management level, I think, in most areas of the business, where -- we are working on that in this internal marketing area, which is the big one. We have some very smart people and we're going to see how they work out. They are generally new.

One of our problems is we get these -- we have lost some very sharp kids, folks that come in here -- one of our problems is we pay -- we have managed to attract people -- good, college graduate, very sharp kids -- paying well under a market rates. But the problem is they come here for a year or two, and they get offered -- you know, they are working for us for $40,000 or $45,000 and they get offered $180,000 by some Silicon Valley firm. So we have had some loss there, and we have had a lot of loss in the accounting department to a company that is really -- a local company that has come in and just hired a bunch of our people and our controllers. I think Jason and Dave Chidester built a really able team there. So we lost a lot of players in the finance department.

I do have to say that in marketing, we have lost -- all these executives can walk out and triple or quadruple their salaries. So that has been something we have given some thought to. I would say countervailing that, we are finding -- actually, Jason, you had some problems with that, too. It is amazing. There are people who are leaving here, who have left here for quadruple the salary or more. A lot of people, actually, when I think of it. But we are getting sort of -- they are getting replaced in many cases by actually some more experienced people, or people from different areas or different industries, and I think we're actually doing -- we are doing okay.

Jason, do you want to talk about employee retention?

Well, we have lost a lot of people over the last year and a half, although I think we have a pretty good team in place to take us to the next level. I do not know what else to say.

I am not talking about the executive team, so much as the next level or two. I see some people and very -- I mean, Cami came in, for example -- Cami and Holly came in and just stunned us. They were so smart and so able. Well, now they are gone, but we must have a dozen folks like them, in the sense of being very smart, mid-20's, who have come in may be without a lot of experience but with real eagerness. So that is where we are.

So, once again, our mantra all along has been we're just going to try to tell you -- you folks know everything we know. We are telling you we are as straight as we can be as about what is going on. We have had ditches before we have climbed out of. We do not think the fourth quarter -- the fourth quarter is still going to be part of a ditch, but it is going to be part of a ditch -- but some things are happening. It is not like we can go from the kind of quarter we just had to a resoundingly successful fourth quarter. So it is going to have been a lousy year, but I hope that the fourth quarter is when we put behind us the mistakes we made in 2005 and are really ready to go on.

Jason?

I hope we end this year -- I hope we exit 2006 with a more simple, crisp business that you will see throughout our financial -- and that you will see the results in the financials.

Dave Chidester?

No, I agree. I think all the things that Patrick is doing to fix conversion, we did have to step back and fix the entire shopping experience first, and that is what we did this year.

Yes, but -- go ahead, I'm sorry.

I think we fixed order management, we fixed the customer experience, and now it is time to fix everything related to conversions.

Yes, and that means fixing the site. To tell you the truth, I think it is going to take us until March to June to really get very good, but I know we can have a lot of stuff. Well, we came in -- I came in in August, identified a bunch of things to test and change and implement, and we have worked our IT team furiously to get these things built and installed and redesigned and changed. Literally two nights a week now, we do code rolls where you see -- if you watch our site, you see these changes. You will see another -- about four rolled last week, four changes, another few rolled tomorrow night, and you will start seeing -- if you go to our product pages, you should be seeing good recommendations. Those recommendations are going to be expanding through the rest of our site in another 10 days. And you will see a gift center up, and you will see a lot of changes actually over the next 10 days. We want to finish by November 17.

But it is not like I can say it will all be fixed by then. We will still have a long -- we're just going to lock down for the Christmas wave, and then we have another three months at least of solid work in front us in the first quarter to really get it good.

Well, to all of my shareholders, all our shareholders, I say I am eating the same cooking you are. I have not sold a share. I'm not going to sell shares. I'm just going to ride this out. We have been in ditches before and climbed out of them. Happy holidays to all.

Thank you for your participation, ladies and gentlemen. Have a wonderful day.