Sunday, October 04, 2015

More Thoughts On Launching API Startups

My experience at Integrate 2015 got me thinking about how APIs and other elements of the data supply chain can be stand-alone enterprises. I will throw my thoughts from our VC perspectives panel out on the Interwebs to see if any of them stick to entrepreneurs.

The successful API supports an app ecosystem. Other businesses in related sectors code their own apps to share affiliate revenue with the API owner for every transaction their apps process. Airlines and hotels build apps around Uber's API because their customers see value in having a short-haul transportation link. Uber's app ecosystem becomes a durable competitive advantage because the app owners would have to build a whole new app if another virtual taxi company tried to displace Uber. Digital infrastructure like apps impose some switching costs on partners, but those costs are probably not insurmountable for competitors.

One thing making APIs a core business model is the continuing integration of cloud, mobile, and Big Data solutions. Some parts of this convergence are coming along fine, like the analytics suites now part of most middleware. Other parts are not working because venture capitalists have wasted lots of money on startups with lame mobile concepts. Venture funds can force their portfolio companies to pivot from a mobile or cloud solution when something isn't working out. That's how they try to save their investments. Targeting APIs for venture investment means the least successful among them will be candidates for forced pivots.

I like startups that understand Customer Development, the Business Model Canvas, and Cloudonomics. Founders should download those white papers, hit the books, and work the equations. Revising the API business plan after working those three areas shows investors that they take de-risking seriously. Corporate development groups in particular carry marching orders form their enterprise mothership to fund startups that improve their internal KPIs and match their product lines. Startups can use Cloudonomics calculations to prove their APIs are a better investment. Cloudonomics is as compelling for IT as modern portfolio theory is for finance, because it offers a disciplined methodology for allocating limited capital among a potentially unlimited number of investment options.

I do not like startups that do not understand the new opportunities and risks of raising capital under the JOBS Act. The SEC continues to publish new rules defining what private companies can and cannot do to attract investment. Anything I blogged about in the past regarding crowdfunding will likely be obsolete by the end of 2015 as the SEC finishes its unfinished business. Startups will have to engage competent legal counsel earlier in their development process to understand JOBS Act compliance. Founders must meet all of the SEC's compliance requirements before they appear in front of pitch fest panels or hang pitch decks on crowdfunding portals. No one wants to be shut out of raising capital because they did something noncompliant.

Data sector startups can tell their stories more effectively by having a founding CEO who knows sales. It really is that simple. Technical founders love solving technical problems, but they may not know how to solve problems with financial or emotional components if they have never worked in sales. It goes back to the classic split in startup teams between the scientific co-founder who becomes the CTO and the MBA-type serial entrepreneur who becomes the co-founding CEO.

If I only had so much money to commit to a venture investment, I would prefer Big Data and cloud concepts over APIs or IoT. The industry standards for data and cloud are firmer than those for APIs and IoT, so the durable competitive advantages of a given business model will be clearer where standards are firmer.

No enterprise can ever indemnify itself by outsourcing risk. Trusting an untrustworthy partner brings a boatload of risk assumptions. Outsourcing an API means trusting someone else's coding with no supervisory input. If the API passes dirty or fraudulent data into analytics, the enterprise's compliance regime becomes a target for regulators. Don't ask for trouble by handing off mission critical API management to a third party.

Chew on all that stuff, API people. There's enough genius here for several weeks of boardroom discussions. Barriers to entry in API development are pretty low, just like in gaming, because the only immediate limit is a developer's imagination. The API businesses will be the next big thing in the mobile Big Data cloud.