[me_bx]

> simply compute all of this information directly rather than inefficiently via the API platform and firehose

I'm not sure to understand what you mean here. By directly, I guess you mean by crawling the web ? You're right about the fact that it would give access to more data, but, as Klout measure is people-centric, a social graph API is a more straightforward data source. Associating web pages to unique identity of persons is a challenge of its own.

> Klout were never going to gain the level of access the needed to the graphs on the various networks they utilized

Given their scale, access to twitter graph data should not have been a problem (they have enough users not to be bothered by the API's rate limits). Other social networks are much more trickier indeed.

Full disclosure: I'm co-founding an social media analytics tool providing more granular view of the interest graphs, using twitter as a data source.

[dotBen]

No, I don't mean crawl the web. I mean Twitter could just compute the data and calculate the graphs via their direct access to the data. Traversing the full corpus of all data, across all dimensions, including time, including data points not exposed via the platform.

> Given their scale, access to twitter graph data should not have been a problem

Yes, because Twitter could decide it doesn't want to provide the data or do so at costs that make it unfeasible. Like I said, this is the fallacy of the platform economy (which I was quite involved with in its infancy).

If you are co-founding anything that uses twitter platform you should consider what happens if Twitter's biz dev team decides they doesn't want you to exist. This already happened once to the Twitter Dashboards etc.

[me_bx]

Ok, now I get what you meant. You're totally right.

> If you are co-founding anything that uses twitter platform you should consider what happens if Twitter's biz dev team decides they doesn't want you to exist.

This is definitely a risk that must be mitigated in this type of projects, indeed.