11th
Charging for API Access?
I’ve been considering the implications of charging for API access for the past two years. During this time I’ve looked at both sides of the argument, and, at times, I’ve been on the fence. Several months ago I began to understand the economics of this new model enough to formulate an opinion. Free is better.
At a certain level, a business needs to have revenue channels. Charging a couple cents per hit on an API can look like an attractive way to make money. The penny model turned Google into an ATM. This business model makes two independent assumptions: 1) A company building on top of an API will generate enough money to cover its cost. 2) a sophisticated user is willing to pay an incremental amount to gain additional functionality from a service. I say sophisticated because building apps is technical. The first scenario is the only plausible assumption. Please note that I am using plausible loosely. In a competitive market, condition number two fails. A user will probably be able to find a substitute that offers a free API (this does not account for network effects— one service might be inherently more valuable). Condition number 1 probably fails because condition two fails in a competitive market.
For so long I have looked at this problem from a revenue standpoint. A great API should increase the value of a service. It should foster innovation and communal values. Charging for access is equally as restrictive as an API that offers little access to data. Free access should be the only model. A service will attract entrepreneurs who see value in building businesses on top of a platform. Although commercial apps make money using a service’s data, they increase the value of the network and intensify the network effects. Commercial apps add functionality to a service. Functionality is all a user cares about. Consider Facebook’s perspective. Their core product is the initial network they designed— A social platform that connects you to your friends. It is not practical for them to shift their focus to installable apps. It would take company resources away from their core product. Facebook would not have been able to match the amount of apps that were created by users and businesses. Free looks like a very attractive option from a service standpoint.
A free API empowers sophisticated users to enhance the service they love. I think the best example I can find is Twitter. I’ve seen hundreds of applications built on top of the API. However, most of them will never be commercialized. Although I’d never use 95% of these apps, I can’t say that they aren’t practical. Each app was created because a user had a real need that wasn’t being filled. The utility of taking the time to build and use the app exceeds the utility of not building it at all. Free adds more value to these users who turn around and increase the value of the service.
Aside from the added functionality, which I cannot stress enough, I think we will see an enormous love effect. Love is an important concept in terms of a web service. I’m going to save this effect for another post because I don’t want to get too off topic.
Ultimately a free API can lead to incremental revenue through increased singups, inbound traffic from links, etc. But as I’ve said before, Twitter has already disrupted this model. Twitter apps don’t generally drive additional traffic to the service. This is because a non user gains no value for seeking out such an app. Maybe this changes when this particular service gets remixed with others. I think that the MyTrade application will drive additional traffic to the service. However, that still doesn’t answer the monetization question. Twitter is so interesting because it’s the one true service that can be accessed without ever visiting their website (excluding initial signup, of course). So how do you monetize a service when most of your users are accessing your product remotely? I’m not sure that this is the exact case— I suspect a fair amount of people still login to their accounts, post via web, etc. But, it doesn’t change the fact that the scenario I have described occurs. I’m going to save my thoughts on this subject for another day.
