When your product extends a platform’s functionality, one of the main risks you face is that the platform could embed your product’s key features within the platform – what is sometimes called subsumption risk. This happened to a lot of startups in the 90s that built products for the Windows platform.
When you depend on a platform for distribution (acquiring and retaining users), you take on different risks. Specifically:
1) Oversaturation. The risk that supply of products on the platform significantly outpaces demand. This seems to have happened recently to the iOS App Store: there are over 500,000 apps and counting, and popularity tends to be highly concentrated, making it very difficult for new apps to get noticed. Oversaturation also happened to Google (organic) results in most query categories in the last 2000′s.
2) Barriers to discovery. The risk that the discovery methods on the platform aren’t meritocratic. iOS apps depend upon appearing in iTunes’ Top 25 lists, leading to a “rich get richer” bias, along with aggressive attempts to game the system. Apple has other app discovery mechanisms like its Featured Apps and Genius features, but those seem to drive far fewer downloads than the top lists. Google search has increasingly been favoring Google’s own products and also seems to heavily favor older, well-entrenched websites, making it very hard for new sites to gain significant SEO traction. Currently, social networks like Twitter and Facebook seem to have the most meritocratic discovery mechanisms, which is one reason so many startups target them for distribution.
3) Throttling. The risk that the platform will throttle distribution or monetization (for apps that rely on paid advertising, throttled monetization also means throttled distribution). Facebook started out letting apps send unfiltered notifications to users’ timelines but then introduced algorithms that heavily filtered them (thereby entrenching the position of leading app makers like Zynga). Facebook also started out letting apps charge users directly, but later changed that policy and imposed a rev-share.
If you are launching a new website or app, you should have a distribution strategy beyond just “people will love it and tell their friends about it”. Your strategy should probably involve at least one major platform. And you should think through the distribution characteristics of the platform and decide if they are a good fit for your product and how best to mitigate the risks.
Finally, it is worth noting that some of the most successful startups grew by making bets on emerging platforms that were not yet saturated and where barriers to discovery were low. Today, the most interesting new platforms are probably Android tablets and emerging social networks like Foursquare and Tumblr. Betting on new platforms means you’ll likely fail if the platform fails, but also dramatically lowers the distribution risks described above.