Two-sided competition
While the concept of earning revenue from two separate customer communities (developers and end customers) and leaving the hard (and risky) work of creating content with mass appeal to ot!
hers sounds attractive, there is a hidden cost. A two-sided platform company must not only attract customers to its products, it also must compete for attracting the best developers to its platform.
Is competing for developers any different from competing for customers? In some ways, the same dynamics apply. In his new book, "Free", Chris Anderson (of "Long Tail" fame) describes some subtle nuances in different subsidization approaches.
Mobile operators frequently provide free handsets to customers who sign contracts. Similarly, software platform vendors that run sophisticated two-sided models frequently give away their development tools. Any company seeking to run an app store strategy must be prepared to give away at least its development tools at no cost. It must also support its developers with training, documentation, a call center - in short, all the infrastructure required to support any product. However, no direct revenue is generated by this investment. So this approach only makes sense because of the enormous potential return offered by reaping even a fraction of the revenues produced the "virtuous cycle".
As noted above, giving away something for free is not alien to telecom's sales vocabulary. However, it is usually alien to the business model. Of course, "free" handsets are not really free. An operator offering a "free" handset deal knows exactly how and when the cost of the handset will be recouped.
Free development tools are different. Developers never pay anything back to the vendor that provides them. Instead, the payback comes when the applications the developer creates lure new subscribers to the operator's network (or entice existing subscribers to spend more money with the operator) - a cross-subsidization effect. There is no guarantee that this will happen, nor is there even an obvious single metric that the operator can monitor to determine whether or not it has happened. How, then, can an operator know whether or not a two-sided business is succeeding or failing?
Monitoring performance
Ultimately, the success of the two-sided model comes down to whether the investment in attracting and maintaining the developer community is paid back by the increase in revenue brought about by an increase in subscriber numbers, and/or revenues per subscriber due to the applications. This is complex to determine, and simply measuring individual business performance metrics is unlikely to provide a definitive answer. More specifically, such metrics alone do not provide any direction regarding what to change if one or more values are not within their desired range.
Instead, to successfully monitor and steer a two-sided business (or any complex business model with indirect effects, feedback loops and similar non-linear structures) a more sophisticated approach is required, based on building a model of the system being monitored.
Under this approach, the behavior of the model can be explored under various conditions and used to develop an understanding of the basic characteristics of the real-world system.