Twitter and Facebook are often lauded as shining examples of what a platform does for your business. And for good reason. They have very successful developer programs. But they are also extremely popular services on their own. A developer may not build where there is not the foundation of a loyal user base. But even that has exceptions.
"Twitter had a dead-simple app and built-in REST API right out of the gate," Evan Prodromou said. Prodromou is the founder of Identica, a distributed Twitter-like service. "It wasn't popular when it became a platform; they were intertwined from the beginning," Prodromou said. Despite the link between Twitter and its API, much of its use has been through the website. Up until recently, the website was completely separate from the API. Still, would the earliest members of Twitter, many developers themselves, have stuck around to make Twitter popular if there wasn't an API to play with?
Facebook was already the social home of most college students when its groundbreaking platform launched. However, the mass adoption certainly came afterward, as developers created additional tools and features that users wanted. The social nature of the Facebook apps encouraged invitations to those who were not yet within Facebook.
Still, Facebook had a desirable service before its platform. Its API enhanced what was already there.
Before location-sharing services Foursquare and Gowalla splashed on the scene, Shizzow's similar service had a rabid following in Portland, Oregon, but it hadn't gained wide popularity. Shizzow had an API, which it saw as the competitive way forward. Why create an iPhone app when developers can do it for you? After all, this approach appeared to be working well for Twitter.
"For Shizzow, we attempted to build the platform before the service was successful," Ryan Snyder, former Shizzow CEO, said. But, Snyder thinks success can come from a platform itself, given the right situation. "Basically, if you have a unique, rich data set that is easily consumable and is interesting enough for individuals... then you can create a successful platform," Snyder said.
I write from experience, as well, having helped launch Geomena's API. The project's aim is to be an open database of geolocated WiFi. The spirit of the project inspired developers, but it's hard to keep momentum when so much is required to collect the locations of millions of access points. Many, including big companies, are interested, but only if the data was already there.
However, there are a new crop of examples of successful platforms without a popular consumer-facing service. In these cases, the platform is the service. For example, Twilio (a ProgrammableWeb sponsor) has created a telephony-in-the-cloud service that give any developer the ability to create voice and phone applications. Twilio charges pennies per minute, which is affordable for most developers.
Similar services exist for other utility functions like sending email from apps, or even finding faces in a photo. And, of course, cloud hosting more or less fits in this new category as well. The difference between these and those mentioned earlier is that they are charging developers money. The success of these platforms, then, depends on whether developers can build a business on top of them, not whether they can cater both to developers and consumers.
Not every service fits into the category that can just be a platform. After all, eventually someone needs to display information to consumers. The service, it seems, is in some ways more important than the platform. But it takes both dirt and water to grow a tree. The current web values openness and connectedness, as we see continually with the APIs and mashups in our directory.
What do you think? Can a platform succeed without a popular service?