Tuesday, June 21, 2005

Speaking Of Which…

Patrick Byrne is one clever fellow.

The voluble, erudite CEO of Overstock.com has a way of presenting himself at conferences that appears reasonable, informed, and responsive to questions—even when he is saying nonsense.

I saw Doctor Byrne two weeks ago at the Bear Stearns technology conference—the summer kick-off for tech investors at the Grand Hyatt in sweltering New York City. I don’t know why, but New York City weather is always especially miserable during Bear Stearns; nevertheless, the conference always brings together a worthwhile variety of interesting companies with plenty of opportunity to for Q&A.

After all, nobody wants to go outside the building in that kind of heat.

Thus it was that I found myself in the Schubert Room of the Grand Hyatt, listening to the CEO of a public company—namely Patrick Byrne—repeat a false rumor about eBay and Yahoo rather than answer a question.

The question he didn’t answer was about what it might mean to Overstock’s faltering auction effort that Yahoo had waived listing and transaction fees on its own auction site.

“I’m surprised” at Yahoo’s moves, Byrne told the small room of investors, before cleverly shifting everybody’s attention away from the actual question—that is, how Yahoo’s moves might affect Overstock—by saying the following:

“The scuttlebutt in the industry is that there was a private understanding” between eBay and Yahoo, “where eBay would leave Yahoo alone in Japan and Yahoo would leave eBay alone in the U.S.”

Having grabbed everybody’s attention with a fake rumor that had no bearing on the actual question, Byrne repeated the rumor that there was some kind of illicit under-the-table deal between Yahoo and eBay, and concluded the non-sequitur by saying,
“The rumor is that goes back five years or so.”

Then it was on to the next question, and yet another non-sequitur—this time intimating that Overstock is an acquisition target, even though the next question wasn’t about Overstock itself being acquired (it was, as I heard it, about whether further eBay-for-Shopping.com-type acquisitions are likely in the industry):

“The number of phone calls I get where it sounds like that’s what’s on people’s minds [acquisition of Overstock] is starting to pick up…” Byrne said. “But I’m not interested in selling at anywhere near this price.”

All in all, during both his Ballroom D presentation and his Schubert Room Q&A session, Byrne made a money-losing hodge-podge of backwater technology (“We got to where we are on a shoe-string”) supporting me-too jewelry/auction/travel offerings sound vaguely sexy, alluring, and ripe for a takeover—while deflecting one of the better questions by repeating a false rumor.

Meanwhile, he casually dropped a bomb that nobody in the Schubert Room seemed to grasp—probably because they were still fantasizing about the takeover inquiries Byrne intimated he was starting to get.

The bomb being Overstock’s cost of acquiring new customers—which is, to continue the metaphor, exploding.

Overstock’s cost of attracting new customers (called CPA), hit $21 in the first quarter of 2005: asked about this, Byrne—who has always focused Wall Street’s Finest on Overstock’s low CPA—talked about the rising costs of attracting customers and said:

“I don’t want to chase it [CPA] over the mid-$20’s.”

This is the same man who 24 months ago bragged, “we basically have the lowest customer acquisition cost on the internet,” when the CPA was $8.69 per customer.

This is the same man who explained away an $18.30 per customer CPA in October 2004 by saying “remember, we’re coming into the fourth quarter…I suspect that you see it drop.” The same man who called that $18.30 CPA “the high end of respectable.”

Now he says he doesn’t want to go over $25?

As usual, Byrne had a ready comeback to the notion that maybe $25 a customer is not economic:

“We’re getting a lot more lifetime value in a customer.”

Right. Okay, Patrick. Whatever. I didn't want to disturb the fantasy by reminding Byrne that large catalogue retailers view $2 a customer as a reasonable CPA.

Oh—one thing more.

Shortly after the Overstock Q&A broke up, I had the chance to ask somebody about Patrick’s “rumor.” That somebody was Rajiv Dutta.

Rajiv is the CFO of eBay, and he happened to be conducting a Q&A about an hour after Doctor Byrne. So I asked Rajiv about Doctor Byrne’s rumored illicit agreement between two of the most successful companies in the world.

After smiling with incredulity as I read him Byrne’s absurd remarks, Rajiv became sober and firm: “There is no agreement, tacit or otherwise,” he said.

His face hardened: “this is an extremely competitive market,” he went on, getting very sober and very firm. Leaving no doubt that eBay and Yahoo were engaged in a high-stakes internet version of Mortal Kombat, Rajiv added, “If we see them doing something well, we will look at it.”

I left the Grand Hyatt for the hazy, hot and humid streets of New York, debating which guy to believe: Patrick Michael Byrne of the tripling customer acquisition costs and the “decrapitating” technology and the $7 million diamond “steal of a lifetime” and Project Ocean and Project Rocket and Project Whatever…or the CFO of one of the most successful Internet companies ever.

It wasn’t much of a debate.

Jeff Matthews
I Am Not Making This Up

The content contained in this blog represents the opinions of Mr. Matthews. Mr. Matthews also acts as an advisor and clients advised by Mr. Matthews may hold either long or short positions in securities of various companies discussed in the blog based upon Mr. Matthews' recommendations.


DaleW said...

I didn't want to disturb the fantasy by reminding Byrne that large catalogue retailers view $2 a customer as a reasonable CPA.

At a 2% response rate and 100% conversion of those responses and $0.45 per mailed piece, it costs $50 per gross add.

At a 2% response rate and 100% conversion for a middle market drive-time radio ad at $125 CPM, CPA is $20. These are obviously high conversion rates and response rates. How do we get to the $2 CPA for catalog retailers? Not that Dr. Byrne isn't on Mars. Dude's a freak. Just wondering how we get to the $2 number.

Jeff Matthews said...

"dalew" raises an excellent question and has his facts entirely straight regarding catalogue retailer customer acquisition costs: the cost to acquire a retail customer via catalogue mailings is far higher than the $2.00 figure I quoted as a "reasonable CPA" for a catalogue retailer.

The $2.00 CPA that I quoted is not the cost of acquiring a catalogue customer by mailing catalogues: it is the INTERNET acquisition cost of a new customer for certain well-managed catalogue retailers.

I used the $2 internet CPA because it is an apples-to-apples comparison with Overstock's CPA: it costs Overstock.com north of $20 to get a customer via online marketing; it costs well-run catalogue retailers $2 to get a customer via online marketing.

Thanks to "dalew" for highlighting my unclear data.

Unlike "dalew," however, "mw blogger" has his facts entirely wrong when he suggests I have confused the term "CPA" and refers to it as "cost per action"--whatever that actually might mean.

"CPA" is in fact the term Patrick Byrne uses to define the cost of acquiring a customer for Overstock.com. There is no such thing as "cost per action" that I am aware of, even in the zany world of Patrick Byrne.

And Overstock's CPA--its "customer acquisition cost" as defined by Doctor Patrick Michael Byrne--is now $21 and climbing.

I have not followed Amazon closely enough lately to know their CPA, and welcome anyone with that knowledge to remind us how much it costs Amazon to get a new customer, vs. Overstock.com.

The Sandman said...

Patrick Byrne is certainly one very clever fellow.

Apropos another conference (the Piper Jaffray Internet conference) I was surprised when reading the WSJ's orginal story on Google's online payment system that Mr. Byrne had a hand in it, and wondered about it, but then thought that it was the Journal and they must have got it right (doesn't figure I know but anyway).

Your latest post has me re-visiting the story. The last 3 paragraphs are intriguing: (the italics and emphasis are mine)....

"Rumors about a new Google payment service escalated following a panel discussion at a Piper Jaffray Internet conference on Thursday. At the conference, Scot Wingo, chief executive of ChannelAdvisor, a Morrisville, N.C., e-commerce consulting firm, said he believed the payment service would be launched soon.

In an interview, Mr. Wingo said he based his statement on questions from retailers with which his company works. Mr. Wingo said the retailers have asked him whether ChannelAdvisor would support the service, which some believe goes by the code name Google Wallet.

During the discussion, Patrick Byrne, president of online retailer Overstock.com Inc., recalls saying, "Yes, this Google Wallet sounds like it might be great. But is all this public yet?" In an interview, Mr. Byrne says he hasn't had any "substantive discussions" with Google about a payment service."

Interesting how that last paragraph is constructed (in more ways than one but I leave that for now).

I take it to mean that Mr. Wingo based his observations (on the new Google service) on the discussions he had with retailers he works with, and that is was during these discussions with (these) retailers that one of them turns out to be none other than Mr. Byrne. The very same Mr. Byrne mentioned many times in this blog as being someone who " has a way of presenting himself at conferences that appears reasonable, informed, and responsive to questions—even when he is saying nonsense."

Not sure that Google cares one way or another that the Wall Street Journal has been seen to run an exclusive on them, but think the Journal should take more care basing it's stories on anything Mr. Byrne has had to say, espicially third hand.

Talk about being conceited: "hasn't had any "substantive discussions" with Google...." I mean, why would they even approach him in the first place?

Orginal sub only WSJ article on JUne 20 Google Plans Online-Payment Service

DaleW said...

From what I have seen, cost per action is a results-based fee for something. Contrast this with impressions or click-throughs. CPA is a pricing term like CPM whereas CPGA (cost per gross add) is not really a pricing term but is a ratio of two things. CPA (cost per action) is a real term while CPA (cost per add) is shorthand for "cost per gross add." "Cost per action" has muscled in one all of this as a new term.

OSTK's CPGA is going up because it's maturing. Pure denominator effect. Well, that and paid search prices are up as well. Perhaps their reputation is also poor (likely) and they are burning through trial users with diminishing returns. I think the site is poorly done and I believe based on Jeff's work the thing will eventually just wither away unless Byrne does something spectacular to put it on a different trajectory.

Kevin H. Stecyk said...

Jeff, I thoroughly enjoy reading your blog. You're a terrific writer with a keen sense of observation and a wry sense of humor. I think everyone who reads your blog will come away a better and smarter investor.

Thank you for sharing your insights!

tiddman said...

I am also not sure if Patrick Byrne is a loon or not. But in the meantime until we decide (those of us who are undecided that is), let's stick to unbiased facts and figures.

Regarding customer acquisition costs, Overstock broke this out by quarter in their last 10-K, and the same figures are available on page 8 of their investor presentation:


Byrne was boasting about their CPA of $8.69 in Q2 2003, and he should have, since it was the lowest they've ever reported. The average for the last 9 quarters is $14.51.

They reported $21.14 in the last quarter, though backing out spending on auctions, it was $19.54. It isn't that the dollars spent on auction marketing don't count, but obviously auctions are a new business, and the cost to acquire customers for that business are going to be higher than the rest of their business.

That's an increase of less than 8% from the previous quarter, which I don't think is something to jump up and down about.

Byrne mentioned two marketing "experiments" that he did last quarter which apparently did not turn out well, and the dollars spent on those experiments goes into this CPA figure, so one could deem that number to be skewed.

Oh and about CPA, Jeff there is in fact a term called "cost per action", just Google that phrase and you'll get bazillions of hits. However I don't think that is what Byrne is talking about, he is talking about cost per acquiring a customer.

That said, I'm waiting for an explanation for how a "well run catalog retailer" can acquire customers for $2. Like the previous poster here, I don't think there are any well run catalog retailers these days with that new fangled thing called the internet, and $2 per customer sounds like it is measuring a different thing entirely.

gmghk said...

On Yahoo vs Ebay in Japan: strange way to go about things if they had an under-the-table deal.

1st: Yahoo only owns 33% of Yahoo Japan, isn't the largest shareholder, and doesn't control it. (http://docs.yahoo.co.jp/info/investor/en/holder/status.html)

2nd: Ebay has already had a try at the Japan market, but left in 2002, pushed out by Yahoo Japan, as it happens. (http://www.ecommerce-guide.com/news/news/article.php/981061 or http://www.thebostonchannel.com/money/1697666/detail.html).

Jeff Matthews said...

I was wrong: "Cost Per Action" does exist, even outside the zany world of Patrick Byrne.

According to "Website Workshop":

"In Web advertising, CPA ads are ads that require the user to perform an action before you receive a comission. This action is typically one of the following:

1. Filling out a form (cost per lead). You are typically paid a set amount for each user that completes the action.

2. Making a purchase. You are typically paid a percentage commission on each item sold.
CPA ad programs are also known as affiliate programs."

My apologies to "mwblogger," who confused "Cost Per Action" as defined above with Overstock.com's customer acquisition cost, which Patrick Byrne abbreviates as "CPA."

The two meanings are entirely different, and I should have googled before responding.

"mwblogger" mistakenly, however, lumps together all of eBay's sales and marketing costs to come up with an inflated customer acquisition for eBay. Companies spend money on sales activities and they spend money on marketing activities, but the two are distinct spending items.

What Overstock is talking about as customer acquisition costs are strictly its marketing costs.

Apples-to-apples, one needs to know how much eBay spent in online marketing to attract a new customer. Any informed opinion on that topic is welcome.

As for 3 year old customer acquisition data for catalogue companies--that's irrelevant because it's not online marketing cost per new customer. Google and Overture were mere pups in the online marketing game three years ago.

Speaking of Google, "the sandman" is absolutely correct in pointing out Doctor Byrne's role in bringing "Google Wallet" to the Wall Street Journal's attention.

I did not mention it in my "Speaking of Which" post on Byrne's appearance at the Bear conference, but Byrne did--in the interest of evading questions about the poor showing of his own company--hint at Google's plans for something big.

"They're working on something really big," he told us knowingly. Nobody followed up on it because it was totally irrelevant to whatever the questions were--but it did, I suppose, provide him with his own feeling of self-worth by intimating that he was privy to Google's secret projects.

After all, his own secret projects haven't come to much.

Remember "Project Ocean"? The auction site involved in "a long war of attrition with eBay"?

According to today's Wall Street Journal front page story regarding eBay's problems...it appears to be Amazon, not Overstock, winning that war.

DaleW said...

It means Jeff is 1) back to shorting OSTK (or frustrated that their stock hasn't collapsed yet) or 2) he's supporting friends that are doing (1).

I'd say the move from $70 to $40 is pretty satisfying.

What is wrong with shorting, anyway? Only an amateur or a very dim professional would fail to believe there are lots of ways to make money in the markets. As long as Jeff's or anyone's analysis (long or short) is reasonable and well thought-out, who cares about the direction of their bet?

DaleW said...

As for 3 year old customer acquisition data for catalogue companies--that's irrelevant because it's not online marketing cost per new customer. Google and Overture were mere pups in the online marketing game three years ago.

This is why Google trades at $287. If the CPGA of a customer via paid search is $2 or $8 or whatever and it costs $30 to acquire a customer through traditional advertising or marketing, all else being equal the dollars are going to flow to the lower cost channel until that difference is arbitraged away. I don't think they're apples and oranges at all.

Now, if the customer is worth 10x as much "offline," then a $20 or $80 CPGA is justifiable relative to a $2 to $8 online CPGA.

DaleW said...

Just as a benchmark, by the way, subprime credit card companies claim CPGAs of around $80-90. Maybe $60 when things really accelerate off a cyclical bottom. Super-prime card companies claim around $250-350. MBNA is probably around $250 -- which is why they get killed when they re-price customers. The lifetime of the customer drops off the cliff, making that investment worthless. Margins go to hell and they hurt their reputation with customers. All of which indicates hair-trigger re-pricing of good customers, just because MBNA wanted to play the yield curve, is about the most braind-dead tactical choice the company could make.

bsilly said...

For all the naysayers, do you think the online closeout industry isn't viable in the lon run? If you do think it's a viable industry, then who is gonna be the top dog or dogs? Ostk dominates right now.
Thanks for any responses.


mfairview said...

"What is wrong with shorting, anyway? "

Absolutely nothing. I've been making money on the long and short of PM stocks lately. I was merely pointing out Jeff's alterior motives behind his Byrne tirade. Looks like I'm not alone here.

Its_strange said...

When does spin become a lie ? Well i hate to say it but when the stock price is the game you gotta figure they ALL are lying...Sad , ain't it ?

Jeff Matthews said...

I take no issue with "dalew" and his analysis of the cost/benefits of online customer acquisition costs vs traditional catalogue customer acquisition costs.

I was merely using the $2 CPA cited by catalogue retailers who use internet search to generate customers in comparison with the $21 it costs Overstock.com to generate a customer via online marketing.

To say that Overstock is doing okay getting customers for $21 a pop using online marketing simply because catalogue retailers pay $20 or $30 a pop through traditional offline means is not apples-to-apples.

What matters is how much does it cost Amazon and eBay and LL Bean to get customers via online marketing, compared to Overstock's $21-and-climbing bogey.

As for "mw blogger," he is getting increasingly hostile and less value-added with each post, and any further posts along the Infantile Yahoo Message Board lines will be deleted.

Regarding Overstock.com's $21 cost of customer acquisition, the bulk of their money is in fact spent online, despite the TV ad campaign.

I'll quote Patrick, from his appearance at JP Morgan recently, regarding this topic [keep in mind, when Byrne uses the term "CPA" he is talking about his customer acquisition cost]:

“The cost of acquiring customers is definitely going up—it’s going up dramatically online. I kick myself, oh, we didn’t have the capital, but 4 or 5 years ago we were just the only guys who were completely focused online, so we were paying $75,000 for a space that 3 years ago we would occasionally pay $120,000 for and feel ripped off, that last year was $300,000 and they now want $700,000 for it. By the way, this is the portal negotiating season, so the portals are—what’s happened is the branding money, the GM’s the Coke, Visa, they’re coming in with money, especially the portals… all that money is floating into the top 50 sites, and the prices, especially for the big, key placements are getting out of control. … So you see our CPA drifting up there. On the other hand there are areas where our CPA is drifting down. If you take out auctions and a little travel that we do and see what we spending online to push shopping, that’s really stayed around the $19 level…”

The above quote was one month before the Bear conference, where he talked about his acquisition costs heading towards $25.

As for "bsilly"'s question regarding whether online closeout is viable or not, the answer is absolutely yes.

eBay, obviously, is highly successful in the closeout field using the auction model. And check out "liquidation.com" for another successful inventory liquidation model.

bsilly said...

Liquidation.com is B to B, whereas ostk is primarily B to C. Also I don't see ebay as a direct competitor.

It looks to me as though ostk has an insurmountable lead and will be the top player in the industry. Given the fact that scale advantages tend to widen the moat over time I would expect ostk to be the main player for many years into the future. That's my take anyway.


DaleW said...

Liquidation.com is B to B, whereas ostk is primarily B to C. Also I don't see ebay as a direct competitor.

They're not? How? Sears has an unbelievably successful, low-cost liquidation model going on eBay:


Overstock.com's auctions are terrible, they have a very high proportion of very junky risk-based merchandise, the diamond thing is atrocious, and by Jeff's work, their systems are terrible. Plus Byrne is obsessed with the shorts and sounds like he has emotional problems. Not a great combo.

Tom said...

"What matters is how much does it cost Amazon and eBay and LL Bean to get customers via online marketing, compared to Overstock's $21-and-climbing bogey."

Fair enough. You're suggesting that the current numbers for those three companies are very low, in the range of $2. Using specific, current data for eBay, I've logically argued that your suggestion has to be very wrong.

There's simply no way that eBay can be acquiring customers online at an average cost near $2. They spent $271M in sales and marketing last quarter and only added 11.6M customers. If your thesis was true, they must be spending less that 10% of their sales and marketing dollars acquiring customers online. Why in the world would they do that, quarter after quarter, when they could pump more dollars into acquiring customers online so cheaply?

The only answer must be that your thesis is very wrong, and that it does in fact cost eBay *way* more than $2 to acquire customers online.

DaleW said...

I agree with MW here. No one is acquiring customers for this amount these days. Maybe way back in 1997/1998, where the natural growth in customers made marketing investments appear unusually productive. No large company is generating customers for this amount of money on a large scale. Perhaps for specialty retailers who have hit upon a nice thing with paid search, but not national/international scale companies. As we hear from Byrne, the arbitrage is underway as well. Coke, GM, P&G, etc are going to drive up CPGA on the internet.

Jeff Matthews said...

"mwblogger": read this entirely, please. Don't start writing until you have read it entirely. Don't respond by re-printing old posts of yours. Read it, go back and get better data, and then respond. Otherwise, post your nonsense on Yahoo because it will not appear here.

"Sales and Marketing" is a P&L entry that encompasses more than marketing costs, and way more than online advertising costs.

For starters, the word "Sales" should give a clue that this category covers ALL costs of selling your products, above and beyond "marketing" to attract new customers.

Here's how eBay defines it: "Sales and marketing expenses consist primarily of advertising, tradeshow and other promotional costs, employee compensation for our category development and marketing staff and certain trust and safety programs."