There is a lot of excellent literature on software platform economics.  (Some of this literature is listed on the resources section of my website.)  Yesterday, I ran across a new academic study that adds to this body of work.

We generally accept that software platforms tend to have “winner-take-all” outcomes. That is, once adopted, software platforms are often very hard to displace.   “Network-effect”-driven economics provide real advantages to a first movers (e.g., Windows OS).  This study, however, tried to simulate real-world conditions in which several platforms might form at the same, with one being superior in terms of its ability to find matches for the two sides of platform (e.g., what eHarmony claims to do for singles). The study aimed to see if participants would just gravitate toward the network that was growing the fastest or whether the superior matching technology would win out–even if there were financial advantages to being on the largest platform and switching costs to moving to the superior one.

The authors concluded:

“We found, however, that despite allowing for switching costs, over time users ultimately all choose the superior platform. This suggests that the winner will be decided by the platform that can provide the most value to both of its user types. This result persists even when we allow an inferior platform to have a head start—to gain dominant market share and lock in customers who want to avoid switching costs. Perhaps, then, there is hope yet for Google+, assuming it can deliver the kind of step change in user experience that Facebook delivered in comparison with MySpace.

How does a platform business achieve a better value proposition than its rivals? The primary factors are increasing the quality of matches between complementary users and charging the most competitive fees.”

The study seems to apply mainly to platforms of the “match-making” variety—e.g., eharmony, ebay, social networks, Angie’s List, Elance etc.  Most B2B platforms have such a match-making element, but typically also have other “stickier” elements that make switching costs much higher for their participants.  These switching costs are part of what makes Industry Cloud and other B2B platforms such high return businesses (and stocks).  The study highlights that platforms need to balance the value they provide with the prices they charge to participants, lest a new, superior entrant arrive.

The study also emphasizes that new platforms are more likely to arise, the more heterogeneous the needs of the users are.  In other words, the more segmented the market, the more platforms can be supported.  So ChristianMingle.com, JDate, and blacksingles.com can compete against match.com and eharmony.  (By the way, those first three sites are owned by the same company, Spark Networks!)

It seems to me that the rise of Industry Cloud platforms is, in part, an attempt to meet the heterogeneous needs of different industries, versus the generic platforms of SAP/Ariba, GXS, Tradeshift, Xign, etc.

The article sites the gaming console market as an unusual one in which there are three platforms that live in an uneasy equilibrium: Xbox, Nintendo, and Sony Playstation.  I’m not a gamer, but it seems to me that all three of these platforms could now be under siege by the games that are played on tablets and phones. Xbox is responding to this threat by trying to become even more of a platform—that is, the platform for your entire home entertainment system.  Microsoft practically invented the platform strategy in software, so this strategy comes as no surprise!

The article, the challenges facing the gaming industry, and the recent sale of the GXS remind me that platforms can be dominant and sticky, but they can never stop innovating, lest another platform challenge them by adding more value and/or lowering price.  The stakes in all of these industries are enormous, so a challenger faces a daunting task, but great rewards if they succeed.  Remember that Facebook was once the challenger!

Platforms are often indeed “winner take all”—but only for a while.