Ross Mayfield's Weblog

Help Wanted

We're hiring a Senior Internet Applications Developer.

BSEC: Tim O'Relly

Impressionistic transcript of Tim O'Reilly's talk at Buying and Selling eContent:

Similar to open source, Web 2.0 is about collaboration.  Companies that are information businesses, software as a service, the internet as a platform and are harnessing collective intelligence.  Yahoo as the ultimate content aggregator. Google's PageRank studies links, not just documents. eBay has 800k people making their living online. Amazon has more than ten million user reviews. 

Only a small percentage of users will go to the trouble of adding value to your applciation, therefore set inclusive defaults for aggregating user data as a side-effect of their use of the application.  Cornucopia of the Commons -- people building a database as a byproduct of using it, where the default is to share.  Flickr drives people crazy who work with formal taxonomies, but look at what you can do with it, the user innovation.  Shows Alexa comparison of Shutterfly vs. Flickr and Wired.com vs. digg.  Postgenomic.com.

Asynmmetric competition: competing with people with no business model or a different business model.  Encarta was a $100M business that destroyed one 10x larger and now comes in Wikipedia and Google.  Is Harry Potter a book, or an imaginative entertainment.  World of Warcraft is competition. Craigslist has 19 employees and is disrupting those with thousands.  The PVRblog makes more money on ads at $4-5 per click than I could pay him to publish a book.

The perpetual beta, when devices and programs are connected to the internet, applications are no longer software artifacts, they are ongoing services, therefore do not package up new releases into moolithic releaseas, but instead add them on a regular basis as part of the normal user experience. engage your userrs as real-time testers, and instrument the service so that you know how people are using it.

The PC is no longer the only access device for internet applciations, and applications that are limited to a single devcice are less valuable than those that are connected.  We have a lot to learn from iTunes, a three tier app from handheld to cloud to pc as control station. (personally, I think iTunes could learn a little from the Cornucopia).

Data is the next "Intel inside" where applications are increasingly data-driven.  Therefore, owning a unique, hard to recreate, data set gives you leverage.  Navteq as the data inside for Mapping, Digital Globe for satellite data, Gracenote for music applications, NSI for the registry monopoly (a lock point for value, despite Esther's efforts with ICANN).  1 Billion songs in a proprietary file format.

A platform beats an application.  Two types: one ring to rule them all or small pieces loosely joined. I'd like to see the latter survive for Web 2.0.  Housingmaps.com as an example that leverages data in mashup, shows the value of other platforms. 

The CD Burn-o-matic made Napster possible, huge difference with the eBook market.  The InternetBookmobile solves this problem.  Shows Safari, if books are used for reference, searching a library of books is what people need.  As you are reading, there is a See Also of related content.  SafariU lets you roll your own textbook.  RoughCuts, a wiki-based front end for collaborative authoring with subscription access to see it being developed.  AuthorsCut, EditorsCut maybe even ReadersCut to come.

In the content business, we have to learn to let go.  A lot of what we have done successfully is to open our content, and get benefits from that.  The Long Tail of Safari use has spikes near the end, where people have found new uses en mass.  This shifts the demand curve.  Books that produce 6% of sales produce 23% of Safari usage.  In the top 10k computer books, 27% of views from books delivering 2% of unit sales, 47% of views from books delivering 9% of unit sales.  Search really does drive discovery.  This is why the Google Library project matters.  The number of books commercially exploited by publishers is 4% of those in print, 20% are public domain and 75% are orphaned works.

Make (described by Steven Levy as Martha Stewart for geeks).  90k paid circulation in 12 months, newstand sell-through of 60% (30% is industry average).  How did we get noticed?  We leveraged the consumer internet.  Find a parade, get in front of it.  Buzz Marketing demands a movement; a story.  Communities make better stories than magazines.  We celebrate characters, not products.  Follow the conversations, blogged like no tomorrow (3k posts in the first month), let readers submit ideas, Flickr phots, Google Video.  Digital distirbution is key to ubiquity, a free offering.  Diversified media: site, blog podcast, video, rss, IM  bot, SecondLife, in-person events, Maker Faire and local Meetups.

"I'm an inventor.  I became interested in long term trends because an invention has to make sense in the world in which it is finished, not the world in which it is started." -- Ray Kurzweil

Buy Side Publishing

What are the alternatives to DRM with a viable business model for content producers in an increasingly decentralized market?  A new media landscape is unfolding, where revenue leakage does not stem from theft, but competition from participatory models and co-opetition from aggregation models.

Participatory models can be exemplified by Craigslist, blogs and Wikipedia.  Users directly generate the content.   Craigslist serves as a datapoint of disruption ($10M in revenue while disrupting $60M ). Blogspace is generating it's own media as conversation  Wikipedia is a constructive community that collaborates to create content.  Production on Craigslist is driven by both market and social signals, blogs and Wikipedia largely by social signals.  The first monetizes a subset of potential posting fees, the latter is all about ME, as in Monetize Elsewhere or repetitional benefits.  All benefit from superior production economies, if not arbitrage conditions between social incentives and markets or firms.

Aggregation models can be exemplified by Edgeio, simplyhired and Google.  Edgeio leverages contributions at the edge of the network into aggregated listings that are complete and complicit.  SimplyHired is a vertical search engine for jobs that scours the web for summaries and links.  Google search provides fair use summaries in result.  All monetize attention wrought through aggregation, largely through advertising market-targeted to the externalities of attention.  All realize production and search economies.

While the long bet may be that Wikipedia eats Google between these models, the game at play is between traditional media and participatory and aggregated media.  The winners will balance freedom and profit-motive with social contracts that beget trust.

Quite frankly, traditional media cannot compete against participatory models.

Simon Waldman thinks out loud:

Fact one: aggregation is a fact of online publishing life - now, and even more so in the future.

Fact two: even if it’s just headlines and links back to content owners sites - if an aggregator is making money from this, they are effectively making money from our content.

Fact three: there is currently no structure for content owners and aggregators to have a useful conversation about rights/ licensing etc, beyond the mutual assertion of ‘you need us more than we need you’.

Fact four: ultimately, individual aggregators dealing with individual rights holders/ content creators is unworkable for both parties.

His solution is a centralized rights agency, the solution that worked for record labels and radio stations.  Unfortunately, that solution was invented in an analog era, and applying it in a digital era will lead to digital rights management.

Information wants to be free is an expression first recorded as pronounced by Stewart Brand at the first Hackers' Conference in 1984, in the following context:

"On the one hand information wants to be expensive, because it's so valuable. The right information in the right place just changes your life. On the other hand, information wants to be free, because the cost of getting it out is getting lower and lower all the time. So you have these two fighting against each other."

In 1990, Richard Stallman put a normative spin on Brand's slogan:

"I believe that all generally useful information should be free. By 'free' I am not referring to price, but rather to the freedom to copy the information and to adapt it to one's own uses."

Creative Commons offers a baseline option, a set of standardized contracts for licensing content that range from all rights reserved to public domain.  CC offers a new business model, by baking in attribution.  Some businesses recognize that reserving all right still results in reuse and remix, and even in some cases of fair use, credit and increasingly valued attention does not come their way.  These externalities are largely positive, and when attribution is leveraged, the attention can derive profit.

Alternative to DRM

Others have argued effectively that DRM destroys value.

I came up with the model for sell side advertising out of a desire to better incentives for quality advertising content.  An open ad is created, the ad has commercial terms baked in but has unrestricted use, publishers choose which ads to place, even copying the ad off of other publisher properties, and are duly compensated.  I believe the same model can apply to content.

The core construct that is needed is standardized commercial contract that is watermarked to only restrict use for compensation.

Watermarked content with these terms can be administered by a rights clearinghouse that has two duties:

• register and track watermarks and distribute royalties
• protect the core license that enables paid reuse

In this otherwise decentralized model, a publisher could locate content on another publisher's property, look up the terms, acquire and republish with the only restriction being payment.

Immutable requirements

This can only work with three immutable terms:

* A presumption of innocence for users and publishers
* The content includes an identifier, such as a watermark
* The content may not technically restrict reuse

The Central Tenant

Web 2.0 and even Enterprise 2.0 rests upon a central tenant.  Nothing ado with the alphabet soup from Ajax to RSS.  It is simply, sharing control creates value.  This common pattern can be found across social software and it's practices of use.  Unfortunately, as Steven Weber first pointed out, while we have buy vs. build frameworks for the valuation of property, we do not have a framework for informing the decision to open.

When you open property for others to innovate upon, I believe the following should be valued:

• you grant real options that create, in turn, real options for innovation
• you reduce search and distribution costs for qualified buyers
• you gain commons-based peer production based on social signals

The costs of opening are relatively nominal, but the risks are not:

• you grant arbitrage opportunities to the market
• you open yourself to scenarios of inappropriate use according to some moral values, but letting go of control can reduce liability

What's wrong with this model

I'm sure I'll learn more after posting this, but my fear is even with baking in requirements, it could trend towards DRM as well.

Disclosures, even for implicit cause: Socialtext is the first wiki company.  I'm on the board of advisors of dabble, which is at the heart of remix models.

Manage Knowledgement (MK)

Manage Knowledgement is a way of describing KM that's backwards but works.  With KM, users were supposed to fill out forms as a side activity to extract their tacit knowledge.  Then some form of artificial intelligence would extract value.  Turns out, users resisted and the algorythms didn't match reality.  With MK, through blogs and wikis, the principle activity is sharing, driven by social incentives.  Contribution is simple and unstructured, isn't a side activty and there is permission to participate.  Intelligence is provided by participants, both through the act of sharing and simply leaving behind breadcrumbs of attention.

Buying and Selling eContent 2006

I'm at BSEC in Phoenix for the second year in a row.  A conference where content as property is a comfortable notion, but where attendees are embracing disruption. Esther Dyson gave a keynote this morning that I missed because the hotel room clock was set two hours behind (spring forward, not back!).  Right now there are some industry incumbents doing blogging 101.  Content distributor NewsTex is licensing blog comment and distributing as "commentary" to customers such as Financial Services BigCos and through LexisNexis.  Rafat Ali is here, waiting to make his move.  It's easy to write off some of the incumbents for not knowing what they are talking about, but we have a lot to learn from them.

Rumor has it that Marc Canter set up an unconference across the street, but has extended the idea beyond having no panelists to having no attendees.

GM Does Open Source Ads

A long time ago we had some theories about where advertising was going.  Jeff Jarvis had a notion of open source ads.  Well, it's all happening.

GM launched Chevy Apprentice:

Contestants are given a variety of images to work with and are given the ability to splice together the visual elements over which they can display their own advertising copy. A contest of this sort doesn't come without risks. As we expected, people who are opposed to SUVs for a variety of reasons quickly discovered that they were also welcome to participate.

Early on we made the decision that if we were to hold this contest, in which we invite anyone to create an ad, in an open forum, that we would be summarily destroyed in the blogosphere if we censored the ads based on their viewpoint. So, we adopted a position of openness and transparency, and decided that we would welcome the debate. (As an aside, we have been truly disappointed by the number of submissions we had to filter out because of their vulgar content.) I won't bore you with the details, but the overwhelming majority of the 22,000 submissions thus far have been earnest attempts at creating positive advertisements.

Was the risk worth it?  Yes.  The worst case scenario is people remixed the brand in the way they already related to it.  A set of already known truths, and if they weren't known, the brand manager wasn't doing their job.  Besides, such remix happens with or without the participation of the vendor.  At least this way they are part of the conversation.

Charlene Li:

While some people point to this campaign as an example of the failure of viral marketing and social computing, I think it points to a great success. Our definition of social computing is when technology results in power shifting from institutions (like Chevy) to communities (like customers). By losing that control over the brand experience, Chevy actually brought more people into it -- witness the debate over the campaign itself. The environmental and SUV fuel economy debate has always existed outside of the Chevy experience, but by bringing it into chevyapprentice.com, Chevy has harnessed it into a promotional benefit.

The point of this remix isn't the outcome, although I'll wager there will be some great creative, but how the process fulfills the ultimate advertising metric -- engagement.  The constructive intelligence phase of the experiment, next is collective intelligence. 

DCamp

Socialtext is opening it's doors to host DCamp, the first ad-hoc event focused on design & user experience, May 12th & 13th.  The folks at BayCHI are initiating this one, here's Rashmi Sinha's announcement:

D is for designers and D is for developers. We hope that this event will attract both designers, developers and anyone else who cares about the user experience. We hope that we will address issues of mutual concern together under the same roof and help build connections between the various communities that care about User Experience.

Go to the wiki, you know what to do.

Software 2006 Wrap-up

I had the accidental honor of being the last speaker at Software 2006.  So as I often do, I veered off of my showcase company presentation to point out how Socialtext is a conclusion of a theme. 

Not only was Ray Lane and other first day speakers decidedly grokking enterprise social software, but the last panel of the second day had Toby Redshaw, the CIO of Motorola, revealing wikis and blogs as perhaps his most successful project.  I wish someone took notes on the stats he threw out, but he viewed it as a bottom up initiative he sheilded from management until it was too late to pull the plug.  They did it in house using open source, had incredible adoption rates (he noted the HCI has two metrics: adoption and turning novices into experts), called it a KM initative and it filled the gap between silos.

It was impressive that MR selected showcase companies that were exemplary of the themes on the main stage.

• BDMetrics -- CRM for tradeshows and events (helping people spend money to get people to spend money)
• blinkx -- Created a new form of contextual browsing and their new video search
• ClairMail -- Mobile applications platform, doing deployments as large as 20M users
• Jotspot -- Focused on SMBs with a strategy of rolling out applications through search direct marketing
• Socialtext -- launched Miki!

Vint Cerf on Innovation Journalism, Blogs and Mobile Devices

David Nordfors provided an introduction to Innovation Journalism.  I liked his simple definition of the business of journalism: attracting attention that you can sell.

Vint Cerf from Google begins with his motto: "Power corrupts and Powerpoint corrupts absolutely."

There are a billion users on the internet, and while that seems like a big number we have five more to go.  We need to take into account how technology is changing the economics of journalism.  On radio and TV your run out of time, with print you run out of space, on the internet you run out of neither.  What you do run out of is the attention span of the users, a finite resource.  In China the statistics are dramatically different, there are more Chinese 300M online than there are US Citizens, they spend more time, 10 hours a week on average for Americans, in China it's 15 hours.  This illustrates how people who have been starved for information act when they have access, and despite the laws in place. 

Early browsers had one interesting feature, View Source, which made all of us students of HTML, and later XML, people learned by copying what other people had done.  Some people in the Intellectual Property community have issues with this.  This has now lead to a substantial amount of people blogging their thoughts.  They are rapidly evolving from text to audio and video.  iPod has become a great tool for listing to audio logs.  As bandwidth gets cheaper, we see iPods turning into vPods and audio blogging turning to video blogging.  Our attention can be drawn to these things with tools like RSS.  The problem is there is a lot of information to choose from.

A interviewer once challenged him with the issue of too much information.  Did we complain about Gutenberg?  We learned how to deal with large quantities of information.  We don't read every book, listen to every radio program, watch every television show.  We get advice about what we should watch.  Sometimes we turn to trusted editors to get advice, sometimes our friends, sometimes we even think about it ourselves.  We will use all those tools.

The blogging world is online and machinable (read, indexed, searched by programs) which allows feedback you wouldn't otherwise be able to get.  The feedback loop on eBay for transactions could be similarly used for information on blogs.  In print media, the bylines have email addresses, where you could potentially interact with the writer.

Telcos talk about realtime streaming video aspirations, but there is a shift to downloading instead of doing something in real time.  The change in people's views on how to deal with these different media -- an important shift, a file transfer is quite forgiving, nobody is watching except a computer.  You can download at a low data rate in the background. Tivo is an example of this, PVRs, iPods.  The ability to work faster or slower than real-time, instead of being confined to real-time is a dramatic change for the distribution of these mediums. In Stockholm you can buy 100Gbps for 100 Europes, download a 1hr movie in 16 seconds.

Technological change changes business models.  Publisher is trying to get people's attention to sell to advertisers as the basic business model.  In the online world, advertising is the core of the business, odd when the service side is helping people find information, but you can make that attention sellable to advertisers, with clear distinction between commercial and searched for content.  You can go beyond reading an ad to doing a transaction on the spot.  The notion of classified advertising in yellow pages becomes a lot more alive.  Moreover, because the net is global you can aggregate a market that otherwise would not exist, such as with eBay.  But you also need local information, and online ads that are localized are highly valuable. Geographically indexing information is a powerful business model.

The power of mastheads like Boston Globe and NY Times is that the more trusted you are the more you stand out, authority and trust is important in this chaotic environment.  People who'se blogs are widely read will have value, but high quality editing is important. 

A Blackberry this morning was my sole access to the internet.  When I needed to make a call to a Representative Baucher in Washington, I looked up the number and clicked on it.  They are far more powerful than we realize and don't understand how dependent other parts of the world are on mobile devices.  Understanding that the journalistic experience needs to be delivered through the mobile device is important (Got to slip Vint a Miki).

The mobile is one example, a small display, but incredibly useful.  In an internet enabled world, there is no reason that a projector could not be online and downloading images, maybe using the Blackberry as a control device. Surrounded by networked equipment that is reachable anywhere, devices harnessed on a temporary basis to do something for you and then released.  I am predicting that during this decade, we will see more systems interacting with other systems like this and in journalism it will become part of our toolkit for sharing and feedback.

Miki

We launched Miki today -- the Mobile Wiki.