isabel wang

The User Generated Sheep Controversy

I've always thought The Sheep Market is way cool. UCLA art student Aaron Koblin collected 10,000 hand drawn sheep via Amazon's Mechanical Turk. He paid $0.02 per sheep; now he's selling sheep stamps for $1 each. How clever is that?!

When I told my mother about the project, she said it's such a racket. She'd be mad if her sheep were on the market. Koblin should have disclosed what he planned to do with the sheep. And sheep artists should be entitled to royalties. I emailed her Jeff Howe's Wired article with a note that she's so unhip. Crowdsourcing is where it's at. Smart business is all about convincing large numbers of people to perform trivial tasks, then finding ways to monetize the results of their collective labor.

As it turns out, many of the sheep artists shared her point of view.  Both the AWS blog and O'Reilly Radar posted copies of Koblin's thesis, in which he wrote:

The response was relatively hostile...  The first responses in a ‘Turker’ discussion thread titled “They’re selling our sheep!!!” included, “Does anyone remember signing over the rights to the drawings?” and “Someone should contact them and see how much they'd charge you to buy back the rights to one of your own sheep.”

Also interestingly:

 * Approximate collection rate - 11 sheep/hour
* Time spent drawing (average/sheep) - 105 seconds
* Average Wage - $.69/hour

I assumed nobody would take proprietary interest after 1.75 minutes of doodling, which apparently isn't true. (I'm debating whether to tell my mom. She's the queen of I-told-you-so's.) In which case, are MySpace and YouTube users upset about the ad revenue that their content helps generate? Did the iStockphoto contributors in Wired's example feel cheated, because the National Health Museum paid only $1/photo for their work after seriously negotiating with professional photographers who charged 100x more? And what about Mechanical Turk workers who took part in other projects? As a point of reference, I've just made $0.03 looking up an URL on my phone browser and typing in a code. I have no idea what this is for, nor do I care.

Here's my hypothesis (which could well be wrong again): producers of user-generated content are less concerned about monetary compensation than being acknowledged for their creative work. Any YouTube videos you upload are associated with your account. You receive recognition each time they're played. Likewise, every iStockphoto you sell shows that someone really liked your work. In contrast, my $0.03 HIT ("human intelligence task") involved no self-expression.

If the sheep artists had been offered the ability to sign their work, and given access to stats on how many times their sheep have been viewed/purchased, would they have reacted with less hostility? If so, maybe there's hope yet for a crowdsourced tech support platform.

I've been wondering whether it'd make sense for a web hosting company to connect certain support ticket categories (Linux questions, for instance, or cPanel error messages) to something  similar to the Mechanical Turk? Anyone could sign up to provide responses. Participants would receive points for satisfactory answers. Just as high post count is a status symbol on forums such as Web Hosting Talk, might a high score on WebHostingAnswers.com become a worthwhile goal among sysadmins?

The WebHostingAnwers.com domain name is taken, by the way. As is HostingAnswers.com. And Amazon already has a general purpose service called NowNow, which gets its answers from the Mechanical Turk. As the user base for S3/EC2 grows, I wouldn't be surprised if an Internet-infrastructure-specific version of NowNow gets integrated with AWS.

Love and Money

A couple of months ago, Denise Howell live-blogged an Attention Trust event. She quotes Christopher Carfi's comment that "people generate income from the people who pay them attention.  Those who earn less do so because they garner less attention.  The potential of doing well without attention becomes harder and harder."

Earlier this week, Doc Searls makes a similar point in a blog post about Flickr: "Love and money. That's the best business mix ever made. Love is why most businesses start. Lack of it is why most businesses end. Love makes money. Hard-bottom-liners don't like to talk about that, but t's true. Few go into business saying, all I want is to please my stockholders."

Guy Kawasaki goes one step farther. Citing a recent study in which participants who were prompted to think about money (by being shown photos of different currency, etc) became 50% less helpful to fellow subjects in the experiment, he concludes  that: "if a company brings money into an evangelistic relationship with its customers, it could create barriers and instead of incentives. After all, evangelism is the process of selling dreams, and selling dreams doesn’t necessarily require monetization."

It was all making sense in a feel good kind of way. Love is profitable, love is free!

And then I read this Techdirt story on FreshDirect, which has been in business for 4 years but is still not profitable on $240 million in annual sales. Carlos' take was: "plenty of people think online grocery shopping and delivery is a good idea -- including customers that rave about FreshDirect. Raving fans, however, don't necessarily mean the business model is sound. Just ask fans of Kozmo's home delivery service."

Business 2.0, too, points out that Zappos, which was founded in 1999, became profitable only this year, on sales of $600 million. BUT - its more-optimistic-than-Techdirt article points out that 65%  of Zappos' shoppers are repeat customers. Also, the company holds 20% of a growing market that's currently worth $3 billion.

Via O'Reilly Radar, I came across this BusinessWeek story on Ryanair: "Despite a 42% hike in fuel prices, Ryanair's profits for the six months ended Sept. 30 soared 39%, to $422 million, on sales of $1.6 billion". CEO Michael O'Leary hopes to continue improving the company's margins - by offering lots of free flights. Last year he gate way 25% of his seats; he plans to double this figure within 5 years, and make all seats free in the near-ish future.

Nick van den Brul, an aviation analyst, calls Ryanair "Wal-Mart with Wings". The company charges extra for every possible amenity: water, checked luggage, wheelchairs... It plasters the insides of its planes with ads. Its flight attendants sell scratch-card games, perfume and digital cameras. Upon arrival, passengers are presented with offers for ground transportation tickets. Its website, of course, is used to showcase vacation packages, travel insurance and hotel rooms  from partners. Ryanair earned $332 million in commissions on such products.

In other words, while the ability to capture attention doesn't automatically guarantee profitability, it does open up a wide range of opportunities. So maybe the folks at Techdirt aren't giving FreshDirect enough benefit of a doubt; it's not such a bad business, it just hasn't found the right formula for monetization. As Tim O'Reilly puts it, it needs to figure out what's the razor, and what's the blade.

Wow - Storage Used to Cost $200 per MB?

From "The Perfect Thing", Steven Levy's Wired article on the iPod (which was originally code named P-68):

Consider the situation when Apple unveiled the original Macintosh in 1984. The computer badly needed a hard drive, but they were so expensive that including it in the package would have almost doubled the $2,500 price. When one finally appeared as a third-party peripheral almost a year later, it was half the size of a shoe box and cost around $2,000. It held 10 megabytes, which seemed like an enormous amount of storage at the time. Ha! The entire capacity of that 1985 disk drive is insufficient to store the MP3 file of Neil Young's "Down by the River."

It's Wiki Week: Jotspot Acquired by Google Just as SocialText integrates with SharePoint

Dion at Ajaxian writes that "the future isn't with emailing Word attachments". Nor will it stop at  online document sharing. Following the release of Google Docs & Spreadsheets earlier this month, Google announced today that it's acquired JotSpot , a wiki company launched in 2004 by Excite.com co-founders Joe Kraus and Graham Spencer. JotSpot's website suggests that its wiki be integrated with Google Docs, Google Apps for Your Domain and Google Groups to form one big productivity and collaboration suite.

Joe Kraus' "we've joined Google" post appeared on Google's official blog less than 24 hours after SocialText's press release on SocialPoint went out. SocialPoint = SocialText's wiki running on SharePoint. CEO Ross Mayfield says the integration will help sell more seats of both products.

ZDNet's Dan Farber says  Microsoft is making nice, calling SocialPoint a best-of-breed Enterprise 2.0 application. And David Berlind writes that

To match Google, [Microsoft] may have to figure out how to acquire or bring into the Microsoft-fold something that's culturally incompatible with Microsoft's primary mode of operation.  "Acquiring" open source and taking it "closed" is very difficult to do. Moving forward, my sense is that Google, Yahoo and others will force Microsoft to not only consider the acquisition of open source-oriented companies and as a result, but also to leave them as open-source oriented properties. In other words, expect Microsoft to be more of an open source company down the road.

If you read this side by side with point #2 on Ross' SocialPoint post, doesn't it sound like there might be some potential for, um, closer collaboration? Especially now that Ross has "changed his attitude towards the convicted monopolist" and written about the mutual profit motive?

I believe that our Open Source solutions cooperating with Microsoft may grow the Open Source community as a whole.  Most people won't see this in Microsoft's best interest.  I do.  Some people won't see this in Open Source's best interest.  I do.

By the way, back in May, TechCrunch speculated on a possible deal between Yahoo! and JotSpot. Now that Google's done that deal, might Atlassian be an option? Susan Scrupski writes that Confluence has over 100,000 users and possibly $15 million in annual revenue. On the other hand, maybe Atlassian's customer base (which includes half of the Fortune 100) is too enterprise-oriented for Yahoo! And as Don from Ajaxian says, JIRA would be a better match with code.google.com.

Snails are Faster Than Broadband!

From the October 16 issue of the New Yorker:

Yossi Vardi, who is sixty-four, helped found ICQ...  He presented meticulously researched, technically correct but completely ridiculous charts and graphs. He compared various data-transfer systems: ISDN, ADSL, Wi-Fly (that is, pigeons). Then he showed a slide of a snail hitched to a tiny chariot with DVDs for wheels. If each disk contains 4.7 gigabytes of data, and if the snail (chasing a scrap of lettuce) travels at 0.000023 metres per second, the snail-system performance rate is over thirty-seven megabits per second. That blows ADSL out of the water. (There are flaws, however. As Vardi noted, "In some regions, most notably France, culinary habits may pose a denial-of-service problem.")

The Importance of Being Like Amazon

Towards the end of Tier 1 Research's hosting conference (Andy and the Dans are really awesome speakers, but I wish David Snead had been there to run the panels. David is the king of moderating trade secrets out of panelists), I had a brief conversation with the CEO of a prominent hosting company. I'll call him X.

X wants to increase signups by 50%, and he's willing to give his marketing team whatever budget they need to make it happen. He thinks many of his customers are developers. Of web apps, you know? As for what kinds of apps those are, or how they're used (within the developer's company? By some third party to whom the developer sells the app? The general public?), he's not quite sure.

Shortly thereafter I ran into a former client who thinks most of his low end customers are 'small businesses'. I've been telling him that there isn't necessarily any correlation between a company's headcount/revenue and its hosting requirements, but he doesn't believe me. I want him to create a 'venture capitalist' role within his sales organization, so that someone mines his customer database for promising opportunities. At the very least he'll end up with some info on who these small businesses are, and what kinds of sites and apps they have.

(According to John Lee's HostingCon slides, that's how Hostway discovered Fox News among its $20/month customers and grew the relationship into a 100+ server account. Which gives Hostway a case study to market to other media companies with.)

In the latest issue of Business 2.0 there's an article on Amazon.com. The company promoted The Stolen Child, an unknown novel that it's turned into a bestseller. And now it's optioned the film rights. Business 2.0 says Amazon can tell from its traffic and sales data what people will pay to see. It can probably selectively pitch the movie to just the customers who would most likely be interested, and extrapolate high potential marketing media by studying this group. That'd be way cheaper than advertising on prime time TV and in major newspapers. 

I think every web hosting company ought to work on imitating the close tabs Amazon keeps on its customers. It'd only be fair, given Amazon's entrance into the web hosting market.

PS - Speaking of knowing your customers, Business 2.0 also has an article on BlueLithium:

BlueLithium serves up 8 billion ad impressions a month to 100 million users. Each of those ads drops a cookie on your browser, and when you show up on another site that serves BlueLithium ads, or one of its advertisers' websites, it adds that history of clicks to its database. Using this 'clickstream' data, it determines within 10 milliseconds which ad to serve up.

"The more we see you, the more we know about you," the company's founder says. He tracks not only what ads you click on, but also the behavior of other people with similar clickstreams.

And last but not least, Applied Location is working on a satellite-based system that can pinpoint every GPS-enabled car's location to within 1.5 meters. The company proposes to use this technology for utility-based tolls, parking fees and insurance.

Knowledge is efficiency.

Someone Ought to Build Some Data Centers in China

During the past 16 months, investors have poured US $8.2 billion into "Web 2.0 companies" in China. (Business Next, the magazine in which I came across this stat, seems to define Web 2.0 as any web-based business.) Over 70 VCs have an additional $5.8 billion earmarked for the Chinese Internet market; they hope that at least $2 billion will be invested before the end of 2006.

Currently there are 122 million Chinese Internet users. Within 3-5 years, China is expected to have a larger web-enabled population than the US. Even now, 13 million regularly access the web via mobile devices, and 30 million are frequent online shoppers.

According to Business Next, the 4 top Internet business models in China are content, social networking, sale of virtual goods, and seach technologies for social media. Some success stories it cites include QQshow, a virtual social network whose members bought US $81 million of accessories for their avatars last quarter, and 51.com, the "Chinese MySpace" that spends US $250,000/month on new servers to support its rapid growth. (They are hosted in Zhejiang province in China, to which engineers from its Shanghai headquarters frequently travel. This seems to indicate a lack of good remote management tools at its data center - and possibly inadequate hosting facilities closer to home?)

Ironically, despite the massive Internet infrastructure requirements that $14 billion in new venture capital funding will generate and support, Tier 1 Research's July Data Center Supply report shows that the amount of web hosting space in China *decreased* by 2.43% between June 2005 and June 2006. Someone ought to go out there and build some data centers!

Waiting for the World to End

I spent all afternoon at Eslite, a Taipei book shop that's Amazon-ized itself into a 7-story department store. I ate Korean food, tried on shoes and bought the latest issue of Business Next (imagine Wired crossed with Business 2.0 and written in Chinese).

My Chinese is rusty: so far I've only struggled through three articles:

1. Ying-hsun Wang (we are not related), a Dan Jiang University professor, argues that Google has a strong brand because its products are widely used - not the other way around. He's sure that your initial exposure to Google wasn't through any kind of advertising or promotional campaign. In fact, he doubts you remember coming across any ad for Google, ever.

2. Writer and researcher Wenchen Pang thinks that Web 2.0 ain't no bubble. BusinessWeek's recent cover story on Digg was instantly blogged by 350+ readers; this shows that Web 2.0-style interactivity has become a way of life, which is a more important fact that any particular site's success or failure.

3. My favorite is Shanghai-based reporter John Wang's (I am not related to him either) recap of an acquaintance's career. The unnamed young exec used to be a fearless entrepreneurial type who leapt from one early-stage startup to another. He had a keen eye for cool new technologies and wasted no time worrying about job security. But through an acquisition he ended up with a comfortable position at an large public company, where his sense of adventure began to slip away. Now he's nothing but an anxious middle man who's in constant fear of getting axed in the next round of layoffs.

According to Wang, corner-office routines are far more dangerous than basement brainstorming sessions. The more someone takes stability for granted, the less equipped he is to handle new risks and opportunities.

His story feels almost like a parable for the web hosting market. Ten years ago, the industry was new and exciting; now it's all grown up. And everyone is clinging sooo tightly onto proven business models that innovations are few and far between. There's no end of depressing headlines on the Web Hosting Industry Review: Intermedia more than doubles storage space at no additional cost! AIT slashes prices by 40%! Unfortunately, the more preoccupied current providers are with maintaining the status quo, the less equipped the industry will be to deliver that next big thing end users are looking for.

You Can Learn More from an Entrepreneur

I just started reading MarketBusters. The book summarizes a three-year study by Rita Gunther McGrath (from Columbia Business School) and Ian MacMillan (from Wharton) on strategic moves that create at least a 2% change in market share, or a 5% annual sales growth. (These may seem like to0-modest goals, but many of their research subjects aren't technology companies.) The two concepts that I've learned about so far are consumption chains and attribute maps.

A consumption chain is the process through which a customer becomes aware of his need, chooses a solution, and makes use of his purchase. Amazon, for instance, became a marketbuster by reconstructiong just about every link in the book buying experience: Its auto-recommendations generate demand. Its one-click payment system is less of a hassle than standing in line. Its free shipping service is cheaper than driving to the mall. And last but not least, customers can even end their consumption cycle by selling used books on Amazon's Marketplace.

Similarly, back in 2000, RackShack (now EV1Servers) created a new market by replacing links in the dedicated servers consumption chain. Instead of custom built equipment, it offered instant activation on standardized units. And instead of long-term contracts, its service was available on a month-to-month basis. Most importantly, its pricing was about 20-30% of what competitors charged.

More recently, RackSpace began offering a different hosting experience via its Mosso division. Instead of configuring, updating and securing their own servers, customers now have access to similar processing power, storage space and bandwidth through a vendor-managed cluster.

An attribute map sorts a product's features into several categories. Some positive features are non-negotiables; every competitor offers them. Others are differentiators. But best of all are exciters, which are so overwhelmingly appealing that their presence triggers purchases. And some negative attributes are tolerable while others dissatisfy. But worst of all are enragers, which lead not only to discontinued orders but bad PR.

Unfortunately, today's exciters too often become tomorrow's non-negotiables. For instance, the month-to-month server lease that RackShack pioneered is now industry standard. And yesterday's tolerables might soon be perceived as enragers. I'm thinking the hardware failure risk associated with running websites on individual servers might fall under this category. Clustered hosting environments haven't yet garnered enough public awareness to become an exciter, but they'll become non-negotiables before you know it!

The moral of the story, I think, is that nothing in business can be taken for granted. Tom Evslin makes a similar point in his post today. His advice is to

1. UNDERSTAND what the important ingredients of your success are;
2. Keep QUESTIONING whether the conditions that made these ingredients important still exits;
3. DON’T assume past results predict future results;
4. Make a new plan BEFORE the old plan fails.

And that pretty much sums up the 59 pages I've read.

Open Infrastructure: Part 2

I read some eye-opening stuff this weekend.

1. Have you ever heard of CoralCDN? I hadn't. But Digg, SlashDot and Fark.com use it to serve links. And about 300 other sites have "Coralized" their software, image, and audio/video downloads. Coral is a P2P content distribution network run by New York University.

Both end users and site owners can take advantage of the free service by appending ".nyud.net:8080" to any URL. Once you access any Coralized content, your computer automatically becomes a distribution node. Coral's distributed sloppy hash table (DSHT) indexing system creates self-organizing clusters of nodes that communicate with each other to optimize download speed and avoid hot spots in the infrastructure.

And that's not all. Like other large websites and content distribution networks, CoralDNS sends requests to the "closest" node based on the client's public IP. But this approach does not take into account the distance between clients and their DNS resolvers. Earlier this year Coral launched the Illuminati network measurement project to quantify this distance, as well as explore the extent of NAT and proxy usage. The results could help refine CoralDNS' algorithm.

2. Through CoralDNS' website, I found my way to an MIT paper on the Cooperative File System (CFS). It's a P2P read-only storage system that runs on Linux, OpenBSD, and FreeBSD. CFS' decentralized architecture uses a distributed hash table for block storage. Blocks are cached and replicated to optimize speed and redundancy. With 4,096 servers, retrieving a block of data involved contacting only seven servers. Performance was unimpaired even when as many as half the servers failed.

3. The website for Berkeley's somewhat related OceanStore project was last updated in 2002. Even then, researchers were working on a prototype for a high availability storage utility than runs atop an infrastructure of untrusted servers.

Any computer can join the infrastructure, contributing storage or providing local user access in exchange for economic compensation. Users need only subscribe to a single OceanStore service provider, although they may consume storage and bandwidth from many different providers. The providers automatically buy and sell capacity and coverage among themselves, transparently to the users. The utility model thus combines the resources from federated systems to provide a quality of service higher than that achievable by any single company.

4. Last but not least, there's Pastry, whose project members include Microsoft Research, Purdue and Washington University. It's an application-indepedent overlay network for routing requests and locating data among P2P nodes. It includes mechanisms that support application-specific object replication, caching, and fault recovery.

None of these ideas have yet been commercialized, but don't you think it's only a matter of time?

Yesterday, in response to a New York Times article on voice over WiFi, Tom Evslin predicated that

WiFi support in mobile phones will shift the balance of power from the big wireless operators to the cellphone hardware and software makers. Phones will be purchased independently of calling plans just as computers are purchased independent of Internet connectivity arrangements.  Coupons for access may be included with phones instead of phones being included with calling plans.  Why? Because voice calling will be too cheap to meter and hardware will still cost something.

Might the relationship between site owners and web hosting providers undergo the same transformation as a result of P2P technologies?