A friend of mine just sent me this article cum press release for a new IBM web platform: http://www.pcworld.com/businesscenter/article/152771/ibm_crafts_web_30_collaboration_tools.html. He commented that it felt to him like a lot of todo over nothing. Here is my response:
Scalability is the question that plagues everything in business. Do the operations scale? Do the systems scale? Does the organizational structure scale? Better phrased: What does it take to scale this business?
Scale is often cited as a major determining factor in the output of various business models. Consulting, law, and accounting are examples of businesses that scale as a factor of people. As in, the company's production is directly governed by the number of employees. Since the only physical products of these industries are documents, it's fair to say that output is entirely contingent on people. Therefore, to produce more you must hire more people, which is difficult and expensive.
Contrast that with Google, the epitomy of hosted services. The company essentially provides hosted software, which is monetized through ads. It employs roughly 10,000 people, mostly developers, and serves hundreds of millions of users. That's a massive scale, although they have their own problems based on maintaining sufficient revenue per user.
Here's one way to think about scale:
Say you have a web company, XYZ.com. XYZ.com earns $1 per user per month and has 100,000 regular users. That means they're making $1.2 million per year. That's enough to support a staff of 15 at an average salary of $80,000 (assuming no other costs), or really a staff of 12 at that salary with $240,000 in annual web hosting costs.
What's the incremental cost of scaling to support 1,000,000 users? You'll have to add a couple more servers at your hosting facility, say doubling the cost to $480,000. Will you need to add more employees?
Possibly, but you won't really need more programmers, and if you run a service like Google's, you won't need more customer support staff because there's almost no tech support. Therefore, your hosted solution virtually eliminates your scaling costs. Incredible, isn't it?
Here is an article I wrote in November, 2005. It discusses a few opportunities that I envision in the future. I have actually written a lot more on the topic, but I'm holding back from posting it all right now because there are opportunities for distribution through other media channels.
Since 2003, there has been a quiet revival of internet-based services, commonly referred to as Web 2.0. The change in the services from the original, Web 1.0, services of the Dot-Com era is a focus on the long-tail economic model, the use of entirely web-based platforms. There are many advantages to this model because it captures a broad customer base and allows for constant software updates without affecting users. The focus of these platforms is tailoring the service to the customer – even advertising, such as Google’s AdSense, is customized.
Web 2.0 is also used to describe the emergence of new social networking websites, such as Friendster.com and Facebook.com, that allow people to maintain profiles with their contact information and establish a link network with their friends and acquaintances who also have profiles on the site. This model has now been extending to allow users to include pictures and descriptions on their sites.
This paradigm also encapsulates a newfound trust in the user. For example, Wikipedia, an online encyclopedia, receives contributions from any user with the expectation that enough users will view each entry that an informed consensus will emerge. This belief is called the principle of critical mass, which asserts that when a certain, large threshold of users is surpassed that a consensus will emerge on a topic. However, this principle also assumes that the emergent consensus is accurate.
The Long Tail economic model describes the large number of products that appeal to only a few customers each. This model proposes that companies can position themselves to take advantage of the economies of scale provided by the internet in order to reduce the cost of carrying products that are bought by only a few customers. When one integrates across a large number of customers all buying a product that very few people like, it can equal or even exceed the revenues gained from a narrow product line that a lot of people like.
On one side of the economic graph is a small base of investors, each with lots of capital to invest; on the other, a large number of users, each with a little capital to invest. The current regulatory structure and investment marketplace restricts investment in early stage start-up companies to only the small base of accredited investors, those with over five million dollars in net worth or over one million dollars in income for at least the next three years, who are believed to be financially solvent enough to handle the high risk of investing in a start-up company.
However, a few companies are beginning to challenge this paradigm. Zopa.com is a website that conducts peer-to-peer banking –an online marketplace for issuing and securing microloans that diversify a user’s risk over a number of loan issuings. This model avoids the securities regulatory structure, which requires many expensive filings with regulatory agencies in order to secure a large number of small investors, because it uses a loan structure that requires the repayment of the microloans rather than exchanging money for equity.
In the future, it is possible that small investors will be able to invest in the start-up market using an online investment marketplace that is the microventurecapital equivalent of Zopa.com. The risk to the investor can be reduced by only allowing small investments – in the thousands of dollars – in individual start-up companies, forcing the investor to diversify his or her portfolio across a number of start-up companies.
The drawbacks to such an approach are dealing with the strict regulatory environment in the U.S. and other Western securities markets. An interesting, but high-risk, alternative to entering this model in a Western market would be to take advantage of the under-regulated securities environment in a country such as China or Russia to debut this economic model, which would hopefully force a change in the regulatory structure of Western nations in order to keep pace.
Under this model, users contribute to projects in exchange for a flat fee or revenue-sharing. High-powered, knowledgeable users have an incentive to contribute because they would now have the opportunity to earn compensation for their efforts. A simple example of an application of this model would be the use of advertising revenue to pay Wikipedia users for their contributions. Amazon.com could also pay users for writing product reviews. In such models, there is a concern that users could attempt to abuse the system for their gain. However, basing their pay off of user ratings, comments, or page visits would limit the abuse of this system.
In addition to broad-based consumer sites such as Wikipedia or Amazon.com, another application of this model is outsourcing development to a large user base. For example, many software projects are ripe for a distributed development model. Most software is divided into a large number of small source files, each of which could be written by a different user who is paid individually for their contribution. Such a system could capitalize on users’ specific skills and significantly reduce the development time for projects by spreading the man-hours across thousands of users. In such a model, quality control can be achieved through unit tests developed by the staff of the company outsourcing the development.
Collective Design allows users to post ideas and set them up the same way as in the consulting model. However, instead of receiving pay for their work, users help one another with implementing their ideas for new technologies. To combat the issue of intellectual property ownership, users would sign a contract releasing their work to the owner of the idea. Collective Design depends on the development of an “eBay” style community of users who are interested in contributing to one another’s ideas.
This model is great for entrepreneurship, as long as the website can develop a sufficient number of users to see enough ideas to completion that users continue to return to the site and the site can maintain a profit.Data mining refers to using search algorithms to infer information from a data set or find a certain, hidden piece of data within a large set. Such algorithms are receiving a lot of attention from groups such as the CIA, Department of Homeland Security, and the Defense Department, as well as large corporations with large data sets stored in knowledge management systems. However, web vendors have equally large data sets, but do not employ such mining techniques to allow their customers to infer information based on the information already present in the data set.
There is a proliferation of many types of websites, such as social network sites, that in order to keep in touch with everyone, one must become a member of multiple sites. However, there is little evidence of consolidation in this industry. A service that is much needed is a single website that permits people to collect their diverse array of websites into a single interface. This website will access the data from all of the websites the user enters, then present it to the user in a single, simple interface where they can interact with the data without needing to know from which source the data comes.
Current Web 2.0 platforms rely on data ownership as their proprietary edge and barrier to entry in the market. However, in order to interact with and attract users, these platforms provide multiple interfaces with which to interact with the data they hold. Most of these sites provide inadequate data mining and search techniques to their users. The data contained in many of these sites is extremely underutilized. Using advanced data mining techniques on data contained on other websites, such as social network sites, a company can gain niche markets such as match-making by providing better solutions than any of the competitors.
This market is a very tenuous hold, but is ripe for leveraged-buyout exit strategies where the companies are merged with the companies on whose data their techniques are used. However, it is unlikely that the data-owners will implement these techniques without pressure from external competition.
The principle of critical mass is critically flawed. Since the principle assumes that the emergent consensus is accurate, it is vulnerable to common misconceptions. Many of these misconceptions can be harmful fads, such as the Atkins Diet. With the propagation of information globally, misconceptions are more harmful than ever. Bad data on Wikipedia can affect a multitude of decisions and attitudes. Running data mining algorithms on knowledge databases will highlight inconsistencies between related pages and data sets, which can then be flagged for resolution by site administrators or super users. These data miners act, in effect, as checks and balances on these data sets (think republic versus democracy), which will help prevent inaccuracies and misconceptions from propagating in these data sets.