Fred Wilson posts about inserting advertising into feeds via open API’s. In the comments, Gnip’s Eric Marcoullier doubles down on the idea. At BearHug Camp, Twitter tells us that Gnip is the gating factor in returning access to the full XMPP feed. Cisco buys Jabber. Let me know when you see a pattern developing.
For years (particularly since Google Reader blew out the feed aggregator market) the marketers have been looking at feeds with smacking lips. Indeed, so have publishers, who have vacillated between the market integrity of full text feeds and the page view games of partial feeds. More recently, the noise around full text feeds quieted. Why?
Because Twitter now dominates the marketplace for feed item notification. Nobody who understands the XMPP real time paradigm uses feeds anymore. Feeds are dead, and have been for a while. No simpler reason for the panic and withdrawal symptoms when track was withdrawn on May 23rd, and no more fundamental reason why we need to start paying close attention to what marketers want to do with our data and who controls that data.
Certainly not those who enter the data – that much is clear. If the high value target of such data is the real time conversation, the actionable envelope in which real business gets done, then as of now Twitter and its partners own that data lock stock and barrel. At BearHug Camp, Twitter executives made clear that the only impediments to returning track over IM and XMPP were business-releated, not technical. Of course, the company wants to bulletproof scalability issues that will arise as track usage ramps back up, but legal work and business relationships with partners and developers are the immediate hurdle.
Since BearHug Camp we’ve heard nothing further about developer attempts to regain a measure of functionality for access to real time track. The suggestions regarding white listing API calls at 5 second intervals or workarounds to a 1 second cache of the public feed have proven unworkable, which leaves getting the data from a third party partner such as Gnip when that company is ready to offer that service.
Let’s say that happens in the next weeks, not months, as Marcoullier told me a few days ago, validating Twitter executives’ comments at BearHug Camp that the holdup was with Gnip. If feeds are dead, or rolled up with XMPP and activity streams as Fred does in his post, then Gnip now is positioned as the owner of our data. How much they extract from us (or rather independent developers as our proxy) becomes an exercise in accounting. If today’s proposed bailout of the US finanicial infrastructure will cost us an estimated $3600 per citizen, then how we read Gnip’s lips will make for interesting conjecture over the next week(s).
Cisco’s purchase price of Jabber and its implications for influence if not control of XMPP remains undisclosed, as (so far) does the relationship between Gnip and Twitter. It’s tempting to view these moves as bigco jockeying for position in the battle for control of the real time network, but that’s not the whole story. Fred Wilson:
I hope (and pray) that this time around we don’t end up with one dominant provider of ad inventory (like adwords has become in keyword based cpc text ads). I hope that the services that provide the feeds to the audience will be able to work with a host of services that provide the feed targeting and execution to the marketers. In effect, an open exchange based on apis and data sharing.
I agree with Fred, but an open exchange based on apis and data sharing would logically include open access not just to the apis but also the underlying business relationships of vendors such as Gnip. Twitter has effectively handed off the burden of transparency to its partner, and with it the responsibility for expressing the rationale for the cost of access to the public data we contribute. Open source and other open coalitions of micromessaging services may find the answers to these questions will provide an economic framework in which their services can achieve some semblance of parity in the marketplace for our content.
If not, the time may come when open services will use licensing or other strategies to brand content with requirements for those data marts that restrict XMPP access to public streams. BearHug Camp 2.0 will focus on this issue as a high order of priority.