A decade of open (API) innovation is ending

In software, open innovation means collaboration through open APIs. More formally, open innovation is the development of next-gen technology and business models via explicit and tacit collaboration among large companies, startups and combinations of the two. One of the first tech examples of open innovation was the IBM compatible PC revolution in the 80s and early 90s.

The 2003 to 2013 decade of rapid shifts in technology (eg, the emergence of broadband, social, mobile and SaaS) also brought a strong period of open innovations for two reasons. First, as new data assets and business models were created, the value and uses of them were initially unknown or at least unproven. Second, emerging businesses with these assets had a strong incentive to lure platform participants by sharing the data assets, business models and spoils of innovating with them.

Facebook is the perfect example. Through rapid growth between 2004 and 2007, Facebook had built a social graph of nearly inconceivable value. How to unlock this?… get developers to make things with it by launching an open platform/API (2007). This attracted external talent and mindshare to Facebook and revealed new business models, cementing Facebook as the leading Web 3.0 platform. Zynga is the perfect example of how this open innovation benefited the platform and the innovator. Zynga brought eyeballs to Facebook while creating an enormously valuable social game (a new business model) for Zynga’s founders and investors.

Fast forward to 2015, and the value of social network data assets – and of many other data assets being accumulated and created in B2C and B2B SaaS applications – is becoming clearer and clearer. Whether for 1:1 advertising, personalized health or customer success analytics, use cases for data are increasingly obvious and established. This leaves one primary reason left to open innovate: to attract more developers and users to create de facto standardship. Yet even this type of open innovation is becoming more limited and controlled. Both the Pinterest API launch and Uber API launch – arguably the most anticipated in the past few years – were careful, slow and focused on large brands/partners… not innovating startups. Once a platform has defacto standardship (or seems to be getting there with no problem), open innovation isn’t so needed.

There are other trends working against open innovation. Privacy and cyber security are both top of mind following big news stories in recent years. The upside of open innovation for emerging platforms is increasingly offset by the risk of running afoul of growing privacy sentiment or increasing the risk of a hack.

These trends are driving major reductions in API access to key technology platforms. Netflix shut down its API, effectively bringing all innovation in-house; Linkedin locked its API; the Facebook Friends API was shut down; Twitter has clamped down its fire hose; and AngelList API access is now hard to get (so sad for VCs). Access is still often available for large and/or monetizing partners, but it’s an increasingly tough world out there for startup developers.

Until the next major tech paradigm shift, we can expect both defacto standard (ex: Facebook) and successful emerging (ex: Uber) platforms to become increasingly insular. The gravy no longer spilleth from the ladle. What does this mean for startups?

If your business model is premised on API data access, cozy up but also diversify: Our best companies who depend on API access of existing or emerging tech platforms build strong person-to-person relationships with platform execs and provide demonstrable valuable (eg, income!) to the platform. As much as possible, these startups also rely on multiple data sources to manage supply risk.

Single platform exposure is increasingly risky and will restrict capital access: After loads of Linkedin API dependent startups were recently left in the cold, many VCs are rethinking their in-portfolio and future portfolio exposure to single platform risk. There are many terrific and performing platform dependent startups out there today, but to start from scratch now is a tough row to hoe for entrepreneurs and investors.

B2C models (apps?) are more risky because they don’t produce revenue to share: If you’re developing the next great B2C app to innovate on the Facebook social graph, your path to revenue is far longer and much less likely than that of a B2B sales enablement tool that integrates with Salesforce. If Facebook changes its API rules, a B2C app doesn’t have much leverage; on the other hand, integration and rev share are part of the Salesforce platform playbook.

If you can’t think small, think big: Innovating on someone else’s API is by nature incremental and now increasingly risky. If access to that API is questionable, strike out on your own and create your own platform, something big! Maybe you can have a highly sought after API someday too.