The quintessential key to success of software platforms has always been their ability to build large, well-functioning ecosystems of third-party producers who build on top of the software platform. Without them, no one would have any need for the platforms. The relevant management skills are:
Knowing the community and being able to identify and attract the third-party producers who can build the most valuable innovations.
Knowing how to manage the ecosystem: Deciding the right balance between quantity and quality; knowing when to compete and when to collaborate with partners; being able to manage the conflicts of interest inherent to such compete-collaborate relationships; knowing how to monetize the value created by the ecosystem (which member of the ecosystem is needed most—developers or end users—and how the pricing scheme must be designed to get both sides on board).
Knowing how to architect the physical design and scope of the platforms: what to do in-house vs. what to rely on others to do; which features to offer and which to forego. Knowing how to spot competitive threats early, both within the industry, but most important and difficult, in adjacent industries, where other platforms might be poised to strike (e.g., smart phones moving into the PDA industry); and conversely, spotting expansion opportunities in adjacent industries (e.g., NTT DoCoMo expanding into payment systems).
Not being sure of the practicality of the above quotation, given its descriptive rather than prescriptive nature, we'll have to read Hagiu's co-authored and recently published Invisible Engines: How Software Platforms Drive Innovation and Transform Industries. The value of Hagiu's approach seems to come from carefully considering the tradeoffs at several junction points along the pathway to building a platform.
Until reading the book, here's one more Hagiu-insight on platforms: a multi-sided platform trumps even a best in class product--a multi-sided platform being a platform supporting a multi-sided business. Hagiu restricts this observation to multi-sided markets, which in this case may well account for network effects. The next theoretical quest(ion) becomes: When (read after how many supported sides) does a platform enter diminishing returns?