I had a conversation with a web entrepreneur the other day and came upon the topic of web service APIs. Offhandedly, I asked him “When are you guys going to write an API?” assuming that the answer would be “Soon” or “We’re working on it” After all, it seems like every web service from Friendfeed to Facebook have one…and even much smaller services with very little traction.
His answer, which surprised me, was that they were not going to write one (even though their main competitor had already done so).
His theory was that there was pretty much no point in writing an API unless you were a service the scale of Facebook, MySpace, etc. After this discussion, I thought more about it and I’m starting to think that there might be something to his argument.
After all, why expend effort on developing a great API when you’re still trying to build a solid userbase?
Shouldn’t user experience and user acquisition be the core focus of your business at the early stage of any web company? Is an API actually contributing to that goal?
When is the userbase of a sufficient size where an API will be a big value-add for both users and the web sites that use the API?
Or is there another set of parameters that one should look at besides userbase size when making this kind of decision?
3 comments
API’s are becoming more and more popular. If for example he creates a Facebook application using his API, then he could theoretically increase his user base if there are users (on Facebook) who use that application.
@ahassan – I probably wasn’t as clear as I should have been in the original post. Clearly there is value in building on top of APIs of the big players (FB, MySpace, etc). What I was suggesting is that if I am a small (in userbase) webservice, that my time may be better spent developing end user functionality versus an API which few people are likely to use.
I think userbase is the wrong parameter to use here. I think the only parameter to use is how unique is the dataset. In the case of twitter / facebook et al their scale gives them unique datasets. However, if there were a site with every type of household plant then that would be a unique dataset and that alone would make it a good candidate for an API. The userbase would be irrelevant.
The site I run is the worlds largest dataset of wine (reviews, images, winemakers notes etc). Our API allows other sites to create mashups of our data, the number of users we have is not a factor.
Actually google maps is another good example, I’d use its API even if it just had 7 users…