Finally, The Internet As Intended.

Ultimately, Web 2.0 is an industry buzzword. Like any buzzword, it’s usefulness can be debated because they are easily mis-used and misunderstood. For me, it is best to think of “Web 2.0” as a handy moniker to describe several characteristics about two things: Companies, and Software (though ironically, “Web 2.0” itself is about redefining software to a degree).

In that it describes both companies and software, we can look at Web 2.0’s defining characteristics as “conceptual” and “technical”. The conceptual characteristics describe a company’s approach or way of thinking about their product, and technical characteristics describe the shared programming architecture principles of the product itself.

It should be noted that these characteristics are not a checklist. Successful companies are exhibiting all or some of these characteristics. O’Reilly notes that deep adherence in one may be more advantageous than shallow adherence to all (O’Reilly, 5)

Conceptual Characteristics

• Service the Long Tail
• Users as a Source of Value, The Network Effect
• Software as a Service, Not a Package

Technical Characteristics

• Delivered over the web (but not necessarily through a web browser)
• Rich Application Interfaces
• Syndication and Micro Content
• Simple, Light Weight
• Open Data/Information API’s

Conceptual Characteristics

Service The Long Tail

The Long Tail, as noun, was coined by Chris Anderson, in a 2004 Wired magazine article . Technically, the Long Tail, is a type of “power curve” (figure 1, below) describing a statistical distribution characterized by a dense clustering of a population which “tails off”. Interestingly, this curve describes many distributions commonly occurring on the web.

Figure 1: The Long Tail

For example, if you were to graph iTunes song sales on the X axis and song titles on the Y – you would end up with a Long Tail curve. Plot the number of Google searches against search terms, and you get the Long Tail. Web site page views by web sites, the Long Tail. Amazon book sales by book titles, again, the Long Tail. Basically, when you plot popularity against inventory, you end up here. What does this mean?

One can use the curve to describe markets in general – with market value on the X axis and markets on the Y. Traditionally, companies like to identify one or two markets on the “head” of the tail – the tall red spike on figure 1 – and play there. Historically, it has proven difficult for a company to profitably service “the long tail”. Why? Servicing multiple markets takes more capital, and a more diverse inventory, among other things. However, the internet has significantly impacted a company’s ability to profitably service the Long Tail. And the Long Tail (the yellow on Figure 1) collectively represents a market far larger than the cluster of markets on the tail’s head. As Joe Kraus, founder of Excite and present CEO of JotSpot, puts it “Its no longer about [servicing] twelve markets of millions, its about [servicing] a million markets of twelve” .

Where are we seeing this work? The most impressive example is Google’s AdWords program. The ad serving companies pre Web 2.0 were never able to extract value out of the long tail. Companies like Double Click based their business model on creating a network of the web’s largest/most popular destinations (the head end of the tail). They wooed them to join their “advertising network”, and then sold advertisement space to advertisers looking for exposure on those major destination sites. Double Click had a network of 100’s of the largest web sites. Compare that with Google Adwords. Google recognized that while there were only a handful of very popular web sites out there, there where millions of web sites populating the long tail. They created a simple way, and very importantly, a free way, for sites occupying the long tail (those with low page-views, serving narrow content markets) to become advertisers. Any site purveyor, using Google’s adwords program can place advertisements on his site, and generate revenue from them, for free. Google handles all of the ad-serving, and serves up advertisements that are contextually relevant to the site on which they are served. As a result, Google can offer ad space to advertisers on web sites that cover virtually any topic. Google’s ad revenue is projected to top $9 Billion in 2006.

Users as a Source of Value: The Network Effect, Collaborative Value, Collective Intelligence

The network effect is a simple one. Some things enjoy exponential increase in utility when there are many of them. Fax machines are the perfect example. A single fax machine is a useless item but a network of millions of fax machines has tremendous value.

Web 2.0 companies have recognized that user interaction, in and of itself, represents value to their services. Your service’s users are a network. And you can leverage that network to significantly strengthen your service. In fact, for some of the most successful companies on the web, the network essentialy is their service (think eBay). You can harness the network’s collective intelligence (Amazon.com book reviews), you can harness the networks behavior as data (links in as a determining factor of relevancy for Google search results), you can harness the network to perform work (wikipedia.com, moveon.org’s Bush in 30 Seconds campaign )

Successful Web 2.0 companies and applications are baking the power of the network right into their offerings.

Let’s use another comparative example to illustrate the point more clearly. Snapfish.com is a photo sharing and printing service that lets subscribers upload and store their digital photos, share them with selected individuals, and order prints. Flickr.com is a service, that has the same core services. However, by default, photos on Flickr.com are shared with the world (the network) unless restricted. On Snapfish, the opposite is true – a photo can only be seen by those invited to see it by the owner. Further, Flickr.com allows users to describe each uploaded photo with “tags”, and allows all users to browse and discover photos by tag – not just view photos they have been “invited” to see by friends. These seemingly simply distinctions make a large difference. Flickr is leveraging its network of users directly within its application itself. As a result it is the fastest growing photo-sharing service on the web with ~775,000 users with 30% monthly growth, a database of ~18.5 Million photos, of wich 80% are “tagged” and publicly available . Flickr.com was bought by Yahoo in March of this year. Snapfish, despite having been around for years prior to Flickr has much slower growth. Snapfish was recently acquired by HP.

The Web as a Platform and Software as a Service

Today, most software is distributed as a “package” and operates on a desktop computer with limited reliance upon the web for most of its functioning. Web 2.0 companies have recognized the web’s capacity to deliver “desktop-like” performance without ever having to install a piece of traditional software. The architecture of Web 2.0 applications relies upon centrally managed software, pushed to the edges of the network, and delivered as a service. This thin-client architecture is not new, but the ubiquity of the web as a delivery medium, and its low cost of distribution (essentially zero) is making the “web as a platform” incredibly attractive for software companies of the future.

Beyond simply leveraging the web as a distribution channel for software to a thin-client, Web 2.0 companies are, as stated above, baking the network effect of the network right into their applications. So “web as platform”, means more than just delivering a service to a thin client – it means enabling that service to take advantage of the collaborative value of its users to create a positive feedback loop – the capability inherit in the platform. By that I mean that Web 2.0 applications are not simply a collection of tools handed to a user (like Word or PowerPoint, for example) with the user performing tasks with that tool in isolation. Web 2.0 applications typically are tools where users perform tasks that take advantage of a central datastore, and their interaction with the tool itself is often a valuable source of information in the datastore. Flickr, for example, uses an algorithm that combines several user contributed data points (number of times a photo has been viewed or commented on for example) to determine a photos “interestingness” – and makes these “interestingness” rankings available to photo browsers.

Delivering software as a service also disrupts the traditional methods of software upgrade cycles. Software as a service means continual improvement, no installs means no patching, and no technical support beyond the support of the end users. Granted, it presents its own operational challenges – scalability, and service maintenance in particular. But the point is that a paradigm shift has been realized, and based on early returns, presents a very formidable challenge to the incumbent models.

Technical Characteristics

What are the technologies frequently used to support the conceptual characteristics that make a company or software “Web 2.0”? The interesting thing about Web 2.0 is that technologically, very little is new - most have been around since at least 1998 These technologies include:

• Asynchronous Javascript (or Flash) and XML to create rich user interfaces, typically referred to as AJAX or AFLAX.

  Programmers are leveraging javascript’s and flash’s ability to make calls back to a server without the user reloading a web page to radically change the user experience of web applications. Today, most web applications require that the entire “page” be refreshed anytime the application makes a request for information to the server. This means users of web applications spend a lot of their time waiting for the server to respond after every click. AJAX and AFLAX enable a browser-based applications to behave much like standard desktop applications, making it easier to leverage the web as a platform, and deliver fuller-featured software as a web service.

   

  

   

• REST and SOAP driven open data/information API’s

  Web 2.0 applications are giving programmers access to their information datastores and basic functions through simple programming interfaces (API’s) based on service oriented architecture principles. A program’s content and functions can be exposed to other programs running on the web by “wrapping” them with these API’s.

  There are really two common API standards – SOAP (simple object access protocol) and REST (representational state transfer). SOAP has a more formal set of standards for handling many programming interface requirements, and is often used for connecting larger business systems. REST, which is essentially just making XML structured content available via documented HTTP requests. Amazon.com, for example uses the SOAP architecture to integrate with partners such as Toys r Us, but makes all of its product catalogue, user reviews, and customer wish list data available with simple XML and HTTP – enabling any site to very easily become a book reseller by embedding amazon content directly into their site. Amazon reports that ~80% of its API use is REST, not SOAP based.

• RSS and other light-weight “micro-content” formats for information syndication

  The example above, of Amazon making its product available for integration by 3rd parties is a perfect example of “micro-content” and syndication. Amazon has broken its content into small, well structured chunks, and using open API standards mentioned above, enables the simple syndication of their content to virtually any 3rd party.

  This openness of information sharing is viewed as extremely dangerous for many content organizations because it strips the content owner of control over the information’s presentation. Still, organizations like Amazon, Google, Flickr, Yahoo, and even the BBC are moving quickly to make much of their information available in small, easily distributable chunks.

• Creating “Mashups” or “Remixes” that combine 3rd party data and information in innovative ways using the technologies described above.

  When you combine lightweight programming models, freely available 3rd party content, and the new ability to make a web page behave almost identically to a desktop application, interesting things start to happen. The result is a surge of applications that combine data and services available through open API’s and “mashes” them together to create some amazingly innovative uses of information that the original distributors of the information never even envisioned.

  For example, housingmaps.com provides a map-based interfaced for finding homes that are for sale or rent in 37 different cities in the United States. But HousingMaps.com doesn’t own any of the listing data; and HousingMaps.com doesn’t own any of the GIS mapping tools responsible for its interactive interface. HousingMaps.com pulls list data from CraigsList.com using their open API, and “mashes” it up with Google’s available mapping API to create the application.