Monday, October 29, 2007

The market is winning: Google's ranking is under strain

It's an interesting philosophy question: When an irresistible force meets an immovable object, what happens?

Google has invested billions in computing and commercial infrastructure in order to secure its immovable dominance in search, which is fundamentally based upon parsing the web graph of links better than anyone else. It has a gigantic $210 billion market cap that is also fundamentally based upon its continued ability to generate reality-defying free cash flow from that search dominance.

That free cash flow creates an irresistible force -- the market, in all its insane brilliance, attempting to reverse engineer Google's map of the Web and modify the link domain to suit the commercial desires of thousands, millions, of individual participants.

Unless Google had an endless bag of tricks up its sleeve, this was never going to end well. And indications are increasing that it's going to end badly. As Peter suggests, fighting paid links is like fighting terrorism. I've got an even more terrifying analogy than that: Fighting paid links is like fighting spam. At best the forces of order are at stalemate in the spam war, and there "we" benefit from the fact that there's no single obvious beneficiary from the current borked email system, and so every major Internet company is motivated to work together to stem the tide. In search, MSN and Yahoo and Facebook are perfectly happy to see Google go down... so defeat is even more likely than what we've seen on the spam front, not less.

I am seriously unhappy with the extent of Google's current dominance... but the prospect of a chaos without center may actually be worse.

This is going to be fascinating to watch. Google has incredible resources to bring to bear, and as I've said before, they haven't even really been tested yet. But... my money is on the market. I'm not short Google, but I'm not long, either. They're priced to perfection, and there are plenty of signs that all is not perfect out there.

Facebook: Options pricing is no hiring hurdle

The Wall Street Journal has a lurking-schadenfreude story which suggests that the imputed $15 billion valuation of Facebook after the Microsoft investment will cause them hiring problems. I think not.

Firstly, I'd be shocked if the clever guys at Facebook haven't structured this deal so that part of the money that Microsoft is paying doesn't accrete to an imputed valuation - instead, it's payment for ad rights, or something similar. If the accounting only puts half of the $240m against the stock purchase, then Facebook's value for options purposes is more like $7.5 billion.

Secondly, the rule of thumb in the Valley is to grant common options at a strike price of about 10% of the preferred, since these shares have less rights which can theoretically decrease their future value in many downside scenarios. That handily locks in a 10X gain in the all's-well-that-ends-well scenario.

Thirdly, most option-motivated employees are looking for a 10X gain in the enterprise value while they work there -- given the built-in 10X from the strike price rule of thumb, that nets out to about 100X in actual return on options for four years of sweat, toil, and tears. On the scale that FB is playing (high stakes, win or lose) that 100X translates to millions in any 'win' case. Can FB get to $75 billion in enterprise value over the next four years? It will be hard, but it's not impossible.

But there's another factor at work. Given the scale of what FB is trying to accomplish, they're looking to hire successful, smart people who have done it before. And it's in this context that the M$ money is a huge win. At this point in the evolution of the web, there's a very large class of professionals - engineers, product people, managers, BD and marketing people - who have mortgages, kids in school, and gold-plated resumes loaded with relevant experience for building great companies. People who, say, went from Netscape '95 to CommerceOne '98 to PayPal '01, just to make up one arc. Hiring those people is tough for an under-funded startup, because they're at a clearly different point on the options-vs-salary & benefits curve, favoring the latter. Google has accelerated its business by being able to pay top dollar for people at this stage of life -- the cream of the crop of HP, Sun, Oracle, and a dozen other companies -- and Facebook can now go toe to toe in compensation package for both upside-motivated and cashflow-driven superstars. As Google has demonstrated, when you join the young, hungry and brilliant with the wiser, experienced and brilliant, seismic events can happen.

The schadenfreude may yet play out, but not this time.

Monday, October 15, 2007

Want to talk local at Web 2.0? Email me

I will be at Web 2.0 for most of the conference. If you are going to be there and want to talk about local, email me myfirstname at If there's enough interest, I might try to put together a dinner on Thursday night. We could go to one of the 732 restaurants within half a mile of the conference.

Saturday, October 13, 2007

Internet Retail: Painfully slow still nets you $100K per hour

I'm buying some birthday presents for my nieces and nephews, so I did separate orders on Lands' End today - one package per kid. My first order, ID number xxx1398, went through at 12:31pm; my second order, xxx2206, went through at 12:51pm. That's one order every 1.4 seconds. If you assume an average order size of 40 bucks, mid-day on this particular Saturday, Lands' End has an Internet cashflow of about $100,000 per hour.

If their site didn't suck, they might be doing 5X that.

The reason I'm on the Lands' End site at all is that my sister (bless her) directed me there for appropriate gifties, and it's alluringly easy for a time-swamped entrepreneur to follow through on such highly targeted suggestions. I'm not there because the LE brand springs first to my mind; I'm not there because they show up highly in a Google search; and I'm certainly not there because the shopping experience is pleasant.

I would rate the functionality of their site as 'high' -- for the particular gift I got (monogrammed bath towels) they have a highly interactive AJAX configurator that shows the particular color, pattern, monogram, etc. But the responsiveness is HORRIBLE - so much so that I read several New York Times articles in another Firefox tab while waiting for basic tasks like the switch from the item page to the shopping cart to.... load. And it's no coincidence that it took 20 minutes between my orders -- that is the total start-to-finish time I went through on executing the second order. 20 minutes!! I could drive downtown and buy something in that time... and if I were buying for myself, I just might.

They've got to be losing customers in huge volume, who either abandon during the search process, or abandon during the ordering process, simply because they get frustrated with the lag. It genuinely feels like dial-up.

As GigaOm noted earlier today, Silicon Valley is all excited about Web 2.0 and consumer, advertising-focused businesses right now, but there's a fortune to be made by some smart new startups building modern, effective Internet software for enterprises like Land's End.

Update: Lands' End actually owns the #1 organic spot for 'embroidered towel' on Google, apparently thanks to smart use of GoogleBase catalog upload. That kind of prime distribution makes their execution even more frustrating -- all those people showing up, and each one having a crappy experience.

Monday, October 08, 2007

More on Google / IBM 'cloud computing' initiative: Tech journalism stinks

The New York Times has a near-verbatim repeat of the story. Mention of IBM's open-source tools is added:
"The centers will run an open-source version of Google’s data center software, and I.B.M. is contributing open-source tools to help students write Internet programs and data center management software."

I particularly enjoyed this Palmasino quote, given my observation last night:
Mr. Palmisano noted that cooperation between the two companies was easier because Google is mainly a consumer company, while I.B.M. concentrates on the corporate market. “We’re more complementary than anything else,” Mr. Palmisano said. “We don’t really collide in the marketplace.”

The Wall Street Journal actually adds a little journalistic merit to its piece, mentioning Sun, HP, and Microsoft as other players in the massive datacenter business; and laid out Google and IBM's open-source rationale as an anti-Microsoft differentiator:
"Frank Gens, an analyst with market-research concern IDC in Framingham, Mass., said the companies also are united by a rivalry with Microsoft, and "they'd like to influence the future of online business before Microsoft extends its influence." IBM and Google stressed that much of the infrastructure will be open-source programs that are freely available, rather than proprietary software programs such as those sold by Microsoft."

There are a number of commentaries visible on Techmeme, none of which goes any deeper than the source articles; Donna Bogatin questions the neat "consumer/business split" but that's all the analysis I see. And none of the commentaries or source articles mention Amazon, who's done more in this area with EC2 and S3 than anyone.

Hopefully someone will start asking some useful questions soon.

UPDATE: The Google press release has key details:
"For this project, the two companies have dedicated a large cluster of several hundred computers (a combination of Google machines and IBM BladeCenter and System x servers) that is planned to grow to more than 1,600 processors. Students will access the cluster via the Internet to test their parallel programming course projects. The servers will run open source software including the Linux operating system, XEN systems virtualization and Apache's Hadoop project, an open source implementation of Google's published computing infrastructure, specifically MapReduce and the Google File System (GFS)."

Key questions answered:
* No Google code open-sourced
* No advanced functionality (BigTable) -- just MapReduce/GFS as implemented in Hadoop.
* Yes, Kevin was wrong (sorry, Kevin :-)

This all fits quite nicely -- IBM gets a great new Open Source Java/Eclipse program to promote (Hadoop is all written in Java), and Google gets to promote its world-view without going through the hassle of open-sourcing any of its own code.

Sunday, October 07, 2007

Google, IBM to fund 'cloud computing' data centers for open research

So says this article in Clearly, IBM and Google have agreed to divide the Microsoft universe between them -- IBM wants to own B2B, and Google wants to own B2C. We'll see whether Microsoft acquiesces to this plan

I'll be very curious to see if Kevin Burton was right about the open-source release of GFS, MapReduce, and BigTable. If he was, I'm never playing poker with him :-)

* Is this just an open service, or open-source software?
* Is it separate storage and processing (like S3 and EC2) or linked processing and storage?
* Does it include advanced functionality (BigTable) or just lower-level components (MapReduce)

Looking forward to finding out more...

Monday, October 01, 2007

Nokia + Navteq: Local, meet mobile

Today's announcement that Nokia is buying Navteq at a $2 billion premium to its market cap is a very big deal. Winning the next generation of the Web is all about owning data, and Navteq is the clear leader in local data. As the second use case for local emerges -- mobile users with immediate contextual questions, as opposed to deskbound users planning for the future -- Nokia's mobile platform technologies and devices will play more and more central a role in user's access to information. Navteq's data underpins both of those use cases, both today and in the future.

When Google is building phones, Nokia is buying data service companies, and Apple is gleefully playing up and down the stack, it's time to recognize that the mobile/web, phone/computer convergence that we've been talking about for at least 10 years is finally arriving.