Hacker News Reader

Comment by wslh | Hacker News Reader

Make the firehose available for anyone at an affordable price and the earnings will jump.

Talk is cheap but in a way this is what Google did with AdWords. In the past you should have a budget of thousands dollars to advertise on popular sites like Yahoo. Google found a way to keep the required budget at a minimum.

I think Twitter haven't done that because it is always more comfortable to talk to a few customers who can pay a lot of money instead of having an horde using your service, but it is currently unacceptable to think this way if you are Twitter.

View on HN

Are there any examples of company selling their social graph data in this manner directly?

The other companies I can think of (FB, Google, Tumblr) utilize the social graph to provide ad-targeting services, but don't charge to access the graph itself. The live nature of Twitter's firehose is probably better suited to sale as a recurring commodity than FB or Tumblr's.

I wonder if there's a price point where they could achieve widespread API usage while still making money with standard infrastructure (no custom datacenters, &c.)

I was talking about the firehose, not the full social graph (although this is interesting to think about). They already sell this but their entry price is high. There other companies in this space using this data and selling additional services such as DataSift.

> I wonder if there's a price point where they could achieve widespread API usage while still making money with standard infrastructure (no custom datacenters, &c.)

What issue do you see here? can you expand?

> I was talking about the firehose, not the full social graph

Ah, I got a bit muddled in my diction here. I was trying to refer to the firehose. The firehose itself should inherently have some useful graph data though, as you would be able to see all the @reply relationships.

> What issue do you see here? can you expand?

I don't have a great handle on the economics here, but I can envision a scenario where:

cost to deliver firehose (engineering, infrastructure) + (marginal cost/user * users) > market value of firehose/user * users

Again, this may sound silly to someone with a better handle on the numbers behind engineering & maintaining this solution. My salient point was just that I'd be curious to find out what kind of value they could extract from the firehose by selling it directly.

I don't know the numbers but I know a lot of people who want to use the firehose for a purpose.

Reply on news.ycombinator.com