FSOSS 2008
The Economics of Open Source
Disclaimer
This paper was written in lieu of an FSOSS paper due to my absence from FSOSS this year. Due to an illness, I was unable to attend. Since taking in some talks at FSOSS was a requirement of this paper, not attending makes it difficult to meet this requirement. After some discussion it was decided that a paper about the economics of open source would be an interesting topic to write about.
Without further adieu, here is that paper...
Executive Summary
In the past, the open-source economic model has developed a reputation of have weak fundamentals. The reality is quite the opposite, however. The open-source model creates distributed costs and risks, employs communities of developers, can dramatically broaden the market for a product, and can reduce development time.
By utilizing open licensing methods that promote an anti-lock-in mentality, you create a user base that feels better about their purchases. This, in turn, creates a community of people who want to contribute to your product; increasing the conversion rate of free-loaders.
It is surprising to some that giving away software can result in profits. Companies like Red-hat and HP have made billions of dollars in revenue using multi-tiered support within the open-source economic model. Other companies like Mozilla create revenue by providing customization services; basically, delivering a version of their software to a company with that company's desired customizations.
More and more people are starting to understand more clearly what the open-source model entails. They are seeing it for the potential of value driven communities and generating revenue through alternate channels. Thus, creating a more open and free environment for your software, its users, and the employees that create it.
Economics Defined
Firstly, it is important to understand that economics is not about money. Economics is about the analysis and description of the production, distribution and consumption of goods and services. This is often forgotten in the marketplace. Too often, companies view money as the object of value, when it is the products and services themselves that create value.
Open Source: The Early Days
In the early days, the fathers of Open Source and it's facilitators did not fully understand their own economics. When Eric Raymond wrote The Cathedral and The Bazaar, there was not much company interest in Open Source projects. As a result, many people believed that Open Source had very weak economic fundamentals. This view is largely unchanged in the mainstream today. Eric Raymond explains that the benefits to Open Source developers are largely intangible; developers relied on their “day-jobs” for their monetary needs. With large, traditionally closed-source companies like IBM and Microsoft dipping their toes into the open-source pool, and open-source success stories like Red Hat, Eric Raymond's assertions are really untrue in today's world. Even though monetary returns are not often as direct as closed-source, working in open-source these days has strong monetary returns.
Types of Economies
Software development economics breaks down into retail, in-house or contract, collaboration without open-sourcing, and open-source models. The primary driving force behind this is the need of software for businesses. This need, in turn, drives a need for software developers.
The Retail Model
In the retail model, the manufacturer carries the cost of the development of software. The manufacturer typically aims to recover these costs, and create profit, almost entirely from the consumer. In this model, the manufacturer must carry the burden of any cost of development until development of the product is complete. Due to the high costs of this model, very little of a manufacturer's income gets reinvested into the development of products. On average, companies that use this model yield a 10% efficiency in returns. In 2004, only 16.8% of Microsoft's revenue went into research and development of products. This low level of reinvestment typically leads to a mismatch between what the customer requires and what a particular software delivers. This also creates a tight budget for product innovation. Due to the high costs of projects under the retail model, only about 50% of projects are deemed successful.
The In-House/Contract Model
The in-house/contract model is based on customers hiring someone or some company to create a software given certain specifications. In this model, the customer controls the end product. If a product does not meet the requirements of the customer, the developers do not get paid until it does. The flip side to this is that the customer usually bares the cost and risk of development of their product. This model is seen to be about 70% efficient, on average. The greatest risks to companies that use this model is that of finding new customers and keeping expert employees. Due to these risks and the strict demands of customers, only 50% of projects are seen as successful.
The Collaboration without Open-source Model
This model is based on the consortium; differing companies collaborate to create non-differentiating software. Development in a consortium costs much more than with open-source projects. This is primarily due to the far fewer members available for cost and risk distribution, greater overhead, and more structure than that of open-source projects. This model is often referred to as a gated-community. A consortium of companies will be formed. All companies in the consortium will be given access to the source of any software developed from within. No companies external to the consortium will be granted access to the source. This model typically leads to failures and subsequent adoption into open-source projects.
The Open-source Model
The Open-source Model involves many people, companies and agencies coming together to work on software development projects. Typically, the first iteration of a project will be created using the In-house model. The software will be opened up to a larger community of developers as soon as the software achieves a useful foundation.
One advantage to this model is that costs are incremental, just like the development of the software. Costs and risks are spread across various iterations instead of a single product release. Another benefit is that the software is developed by it's users. The end result reflects the feature set the users want. The users themselves become part of the project. They feel a personal attachment to the software and a connection to the community.
Another advantage to the open-source model is open-source marketing. Open-source marketing and distribution channels use the same fundamentals of open-source development. This increases the breadth of markets software can reach and helps to level the playing field when dealing with your competition; both large and small.
The biggest challenge to open-source projects, and the primary reason these projects become failures, is customer and community attraction. The success and lifespan of an open-source project depends heavily on the community. About half of all new open-source projects fail, but the distribution of costs means that the open-source model is extremely efficient.
Contributors
The economics of open-source is largely supported by it's contributors. These contributors come from all walks of life and all sectors of many industries.
The primary contributor is volunteers. Volunteers make up the largest base of developers to open-source projects. Their monetary contribution is their personal time that they spend contributing to projects. A community of a thousand volunteers can bring multiple skill sets and the possibility of a 24-hour development day (in the case of international communities).
Another major contributor to open-source is the academic community. A large portion of open-source software and the BSD open-source license came out of the BSD project at Berkeley University. There are also many open-source communities that are working on research. Open-source software supports researchers in collaboration; it allows them to share costs and lessons learned from experiments. In addition to this, many students are volunteers in open-source projects and educational institutions are starting to provide open-source related curriculum.
Economic Impact
Many businesses rely on the business of other companies to support them. Software allows them to conduct this business more efficiently. Microsoft is probably the most visible example of this. Their software has become so popular in the enterprise environment because it has increased business inter-operational efficiency greatly over the years. As businesses seek to improve efficiency, they look to companies like Microsoft to create the software they need.
The same can be said about open-source companies. Most of the world's network services (web, email, e-business, etc) run on Apache or other open-source Unix/Linux based softwares. It is not a far stretch of the imagination to think that open-source's economic impact is in the billions of dollars. Open-source software allows businesses to have better software quality and control at a far reduced cost over close-source alternatives. This savings is typically reinvested into R&D initiatives resulting in faster innovation and efficiency improvements.
Free-loaders
Unfortunately, a large problem with the economics of open-source is the presence of free-loaders. That is, people who take advantage of a product or service without giving back. Every user of open-source software typically starts out as a free-loader. They download the software for free, to fill some personal want or need. They often do not consider contributing to the project until there is something they want to be able to do with the product, but can't. From an economic perspective, free-loaders are an accepted norm.
Assuming your product has one million freeloaders. If one percent of those free-loaders become contributors, that's 1000 people adding value to your product for free. This is invaluable to a company, considering that hiring 1000 new developers could cost a company tens of millions of dollars.
Linux: The Open-source
In the last couple years, open-source software has been gaining share of the market and share of market coverage by the media. A search of news.google.com for articles that mention “linux” from the year 2000 returns 26,900 articles. Compare that to 27,099 articles just in the last month and 43,300 articles so far this year. This is largely due to large companies seeing Linux and open-source software as a viable for business and the open-source model as a viable business model.
In 2002, IBM made $1.5B alone from Linux initiatives. Comparatively, HP made $2.0B from Linux initiatives over the same period. In addition to this, Judy Chavis, HP's Linux Director is quoted as saying, “Every single business unit, whether it be hardware or software, has a Linux roadmap and Linux deliverables”. This shows that there is definite business interest in the open-source economic model and the open-source technologies that are borne from it; but why?
Why Open-source?
There are many reasons to adopt the open-source model. Some of them have already been mentioned. The following are some of the most important and popular aspects of the open-source model.
Development Time
Due to the community aspect of the open-source model, projects are opened up potentially thousands of developers. This creates a much more efficient development process, more varied skill sets and differentiating perspectives to projects. This is something that is much harder to achieve in the other economic models without dramatically increasing cost and development time.
Licensing
No commercial software is without licensing. Most proprietary licensing is designed to protect a business investment into a product. Companies like Microsoft and Apple take this to an extreme by not letting you share the software whatsoever. This protects the millions and billions of dollars they have spent in research and development of their software.
Historically, open-source licenses like the BSD (Berkeley System Distribution) License and the GPL (GNU Public License) have been seen anti-commercial. This is quite a skewed vision of what these licenses stand for. These licenses are designed to be anti-lock-in. This means that people should not be punished for the software and devices that they purchase. They should be free to explore all the possibilities their purchase should open up. Take, for example, the iPod and iTunes.
While the iPod is almost definitely the most popular MP3 player, at least in North America, it is also one of the most locked devices. By purchasing an iPod and activating it, you agree to adhere to Apple's strict guidelines of using iTunes and limiting yourself to Apple's “approved” media formats. There are many MP3 players that do not lock you into a particular file format and/or a particular software. Sansa, Creative, and Samsung are just a small example of companies offering such devices.
Making Money
Since the open-source economic model often means giving your software away, the question of making money is often asked. The primary method of making money using the open-source economic model is to provide paid-support. Most Linux distributions have two levels of support. The first level is free, supported by the community. The second is a pay-per-use or subscription based support service. Most companies adopt the second method because it is direct support coming from the development company. This represents a good stream of income for the companies developing the software.
Another way to make money in open-source is to provide extension services. A company that buys open-source software may not have the time or resources to learn how to extend and to implement customizations to the software. Hiring an outside firm to create extensions or customized versions of the product generates revenue and adds value to the product itself. For example, from time to time, Songbird, an open-source media player, develops partnerships with other companies to provide a customized version of their software for the partners' clients. Mozilla has done this in the past with Firefox as well.
Conclusion
In years past, open-source was viewed as having weak economic fundamentals. Companies, businesses, consumers and the media are starting to wake up to how skewed a view that is. IBM, HP and Red Hat are just some of the examples of companies that have been quite successful with the open-source economic model.
The power of open-source is something companies, consumers, and the media is just starting to understand in recent years. This power comes from the communities that form around open-source projects. These communities allow software companies to decrease development times, reduce costs, and create software of value to its users. This results in an economic impact in the billions of dollars annually.
The open-source economic model truly is a viable model and seems to become more and more relevant.
References
1. Zymaris, Con. “How to make money from Open source”. Strategy – Business Management – Builder AU. 2002. <http://www.builderau.com.au/strategy/businessmanagement/soa/How-to-make-money-from-Open-source/0,339028271,339191343,00.htm>
2. Perens, Bruce. “The Emerging Economics of Open Source”. Cyber Security Policy Research Institute, George Washington University. 2005. <http://perens.com/Articles/Economic.html>
3. Green, Eric Lee. “Economics of Open Source Software”. 2002. <http://badtux.org/home/eric/editorial/economics.php>
4. Lerner, Josh. “The Simple Economics of Open Source”. Harvard Business School. 2000. <http://www.people.hbs.edu/jlerner/simple.pdf>
5.Raymond, Eric. “The Cathedral and The Bazaar”. Thyrsus Enterprises. 2002. <http://www.catb.org/~esr/writings/cathedral-bazaar/cathedral-bazaar/>