11 Jan 2010  Research & Ideas

Mixing Open Source and Proprietary Software Strategies

Open source and proprietary software development used to be competing strategies. Now software firms are experimenting with strategies that mix the two models. Researcher Gaston Llanes discusses recent research into these "mixed source" strategies. Key concepts include:

  • Software companies are taking a "best of both worlds" approach by creating products that use a combination of OS and proprietary software code.
  • The researchers wanted to get a clearer sense of when a profit-maximizing firm should adopt a mixed-source business model and what that model might look like under different circumstances.
  • Results indicate recurring patterns and strategies that managers can take into consideration when setting strategy.

 

Adopting a new business model can be a strategic, game-changing play in any industry. But knowing when and how to try something new can be tricky, particularly in the constantly evolving software industry.

The open source (OS) movement is one model—it's going strong after nearly 30 years and still has its die-hard supporters. Meanwhile, other firms try to maximize profits by keeping a tight, proprietary hold on all intellectual property. Increasingly, however, software companies are taking a "best of both worlds" approach by creating products that use a combination of OS and proprietary software code.

In their working paper "Mixed Source," HBS associate professor Ramon Casadesus-Masanell and postdoctoral fellow Gaston Llanes consider scenarios in which theoretical software firms compete through different business models under a variety of conditions.

"Microsoft is being very strategic, approaching it product by product."

Their findings highlight the complexities managers face in navigating an increasingly competitive industry. But if there's no easy recipe for software firms, even in the seemingly ideal balance of mixed source, these results indicate recurring patterns and strategies that managers can take into consideration when setting the course for their own firms. A Q&A with Llanes follows.

Julia Hanna: How did you come to be interested in this particular area of research?

Gaston Llanes: We started to look at this topic because so many commercial firms were devoting resources to developing open source products. That sort of behavior is initially puzzling to economists because the firm is participating in the development of something that is going to be given away for free. Once you begin studying these organizations, however, you notice that one way they do profit is by selling complementary goods and services. For example, IBM sells consulting services and proprietary software that are complementary to the OS software it develops, and Sun sells complementary hardware such as servers.

Another combination that interested us is mixed source software, where firms develop OS software while selling some part that is closed, or proprietary. That model sounds like it should work well for all companies, since it seems to represent the best of both worlds—innovation and value capture—but it's more nuanced than that, as our paper shows. We wanted to get a clearer sense of when a profit-maximizing firm should adopt a mixed source business model and what that model might look like under different circumstances.

Q: Can you first explain the technical details behind mixed software?

A: Sure. In industry speak, software is composed of two modules—the core, or base, and the extensions. There's an asymmetry between these two because the base can be used without the extensions, but to use the extensions you need the base. So, there are a few overall business models that currently exist in the software industry: proprietary, OS, or a mixed model, opening either the extensions or the base. In the software industry, the open extensions model is sometimes referred to as "open edge."

Q: Walk us through one of the market scenarios in your paper.

A: In the first case, we examine a monopoly-type situation, where the firm is the only one with a product in that market segment. We find that when the value of the complementary good and user innovation is low (meaning that the technical possibilities for users to improve the software are limited), the firm will choose a mixed source model. In this instance the firm doesn't open its core, because it has something to lose by creating a direct OS competitor. So the firm opens the extensions, because it can benefit from user innovation, but customers can't use the OS software without purchasing the core product. It's a relatively inexpensive way to become open—you don't lose much. Stata, a software product used in econometrics, is a good example of this.

"Our research plainly shows that value creation doesn't necessarily lead to value capture."

In cases where values are intermediate, we find that the firm prefers to open the core product only. And when the value of the complementary good and user innovation are high in a monopoly situation, the firm will open both modules and use a pure OS model, because the quality improvement that occurs when modules are opened outweighs any potential competition.

Q: But it's not always a good idea to adopt an OS model.

A: Right, it really depends on your product and the competitive landscape. For example, if you have an OS competitor like Apache or the Free Software Foundation, you may not want to be so open. This is an interesting point to make, because usually people think if you face competition from OS you are going to respond by becoming more open, always. But that's not the case.

In fact, one of the examples in our paper shows that a firm is more likely to use a proprietary business model when it faces competition from an outside OS project, particularly when the outside OS project has a base module. In that case, a combination of the outside base module and the proprietary firm's open extensions could result in a stronger free competitor. So the firm is more likely to open substitute modules, rather than complementary ones, to the outside OS project. We do find that the firm will prefer to adopt the modules developed externally when they're of higher value than the firm's own modules. In that case, increased openness obviously results in greater value creation and value capture.

Q: It seems that open source software, particularly when it's part of the "mixed" model, is becoming more prevalent.

A: Yes, even Microsoft has jumped on the bandwagon to some extent. It's partnered with Novell to put some of Microsoft's technologies on Linux and other open platforms. The Mono project consists of porting the .NET framework onto Linux, and the Moonlight project is about providing an offer of Silverlight, a Web-based digital video technology, for Linux. And in July 2009, Microsoft agreed to contribute some of its technology to Linux under a licensing agreement that allows developers outside Microsoft to modify the code. They are being very strategic, approaching it product by product.

Q: Your paper also considers several different scenarios between two competitive, for-profit firms. Can you talk about those findings?

A: The method we use is a two-period game where in the first period (the strategy period) business models are chosen, and in the second period (the tactics period) firms interact by making tactical pricing choices as allowed by their models. Many different things can happen in this instance, so it's much more complex.

We study two firms, H and L, where we assume that firm H has modules of higher quality than firm L. We find that when the external quality difference between two firms is low, one firm competes through a proprietary business model and the other opens one module, generally the extensions. As the quality difference grows, cannibalization concerns lessen, and both competitors elect to compete through the same mixed source business model.

We also find that both firms may prefer to compete through the proprietary business model when H is the first mover, although that never happens when L is the leader—a low quality firm is more interested in competing through a mixed source business model than a high quality firm. But when the low quality firm is first to act in the market, and the higher quality firm reacts, the lower quality firm may "leapfrog" the higher quality firm. It's a good recommendation for the low quality firms and a cautionary note for high quality firms. If a low quality firm is quick enough to act in the market, it may surpass its higher quality competitors.

Another interesting outcome is that when user innovation between the two products is low—regardless of which firm moves first—one firm decides to use a mixed source model, while the other is proprietary. So a firm differentiates itself instead through its business model. But when the innovation and quality differential are higher in both firms, the firms tend to look more alike, choosing the same business model.

Q: Can you give an example of a firm adapting its business model?

A: Yes, this case actually motivated me to start working in this area. IBM was facing competition from JBoss, a growing new firm with an open source business model. In response, IBM in 2005 bought a small firm called Gluecode that sold products in the same market segment as JBoss. IBM then opened the Gluecode product and adopted a mixed source strategy as a response to the competition it faced from JBoss. It's a good example that shows firms can adapt their business models in response to competition from other firms.

Q: What factors should managers take into account when it comes to using, or not using, a mixed source business model?

A: They should take into account who their competitors are and what the nature of the competition is. They should also weigh the importance of user innovation for their market and the value of their complementary good. In the case of some software products, the complementary product is very important. For instance, a server operating system is very complex and will probably require support, or at least training, if you run into problems. But if you are talking about a desktop operating system that is easier to use, the complementary good is not as valuable—it's more difficult to use an open source model in that case because then you don't have much to sell in the way of services.

If anything, our research plainly shows that value creation doesn't necessarily lead to value capture. Instead, for-profit firms should choose a business model to capture as much value as possible, taking into account the likely strategic and tactical reactions of other firms. It's impossible to give a unique recipe for all software firms—it really depends on the industry's configuration and the placement of your product with respect to competitors.

Q: Your paper concludes by citing other industries where new technologies, regulatory changes, and customer demands have driven the innovation of new business models. Can you talk about that a bit?

A: Yes, even if your company is operating in an established market, some organizations have shown that it's possible to open up space in a crowded industry by thinking of a new way of doing business. Ryanair, IKEA, Cirque du Soleil, and Betfair (online betting) have all grown quickly through innovation in their business models.

Q: So are mixed source business models seen as the next big innovation strategy in the software industry?

A: This is being discussed quite a lot right now. In the past, a pure OS model was the big thing, but there were many firms that then found they had a hard time being profitable. For example, some observers say that Sun Microsystems ended up being acquired by Oracle because it was too open.

There is definitely a lot of thinking out there about what is the right degree of openness. The conflict between more "fundamentalist" developers who believe all code should be open and those who don't see it that way still exists. In the long run, however, it goes without saying that you can only subsist if you have enough revenue to sustain the development of the product.

About the author

Julia Hanna is associate editor of the HBS Alumni Bulletin.