CDR = Customer Data Revolution
The opportunities and pitfalls of the telcos' vast stash of CDRs (Call Detail Records) and phone bills have been a top theme here at the Telco 2.0 event. Last year, you may remember, we said on this blog that in the future, so many new applications will need contextual data to function that we'll need to think of how subscribers will take their data shadow with them when they churn. It looks like this is going to be more important than ever.
Paul Magelli, head of subscriber data management at Nokia Siemens Networks, just gave a presentation in which he argues telecoms needs to invest in understanding customers in the same way it invested in monitoring and instrumenting networks in the last 10 years; we're going from a "network driven world" to an "information driven world". Otherwise how would we know if someone, like Magelli, had two mobiles, a BlackBerry, and a laptop dongle but was still the same person?
Operators, he argued, enjoy a relationship of trust with their subscribers, as opposed to (say) Google's "relationship based on openness". The importance of contextual data is difficult to underestimate; Thomas Howe's presentation on day one was all about adding voice "as a spice" to other business processes and their hard-to-replicate data assets, and what could be achieved with an API that takes more arguments than an e164 telephone number, Martin Geddes's talk in the same session centred on integrating other kinds of context around the voice call, JP Rangaswami argued that being able to keep the context of a call -- "TiVoising voice" -- was a transforming event.
But this raises some crucial and difficult questions. Arguably, CDRs -- "the real social network" as Howe calls them -- are the creation of the subscribers, just as the content in Wikipedia or the links Google counts are. Carriers facilitate this, but only keep them for their own billing purposes (and because sinister government agencies want them to). Everyone at least mentioned the need to respect users' privacy, but there was little said about what this meant in practice. Is it really true that operators enjoy a "trusted stewardship" status in the eyes of subscribers? Just as one of the barriers to new VAS is the fear of a disastrous service launch, one of the barriers to new uses for this stuff should be the fear of a privacy or security Chernobyl that would destroy this trust once and for all.
Perhaps the guiding principle should be that operators respect subscribers' data sovereignty? That would mean subscribers would have to explicitly and effectively choose what data to release and how; it would also mean that they would have to be rewarded for uses of it that mainly benefit the operator, like ad targeting. The reward, however, doesn't have to be money. It could be quasimonetary -- lower rates -- or it could be access to new and compelling applications. Carriers would have to make it easy for churners to take their data shadow with them as they go out the door. Perhaps, as someone suggested today, this is yet another reason to deploy ENUM. That sounds grim, but the flipside is that you'd need to make it easy to import data; which is all good if you consider the CDR pile to be a strategic asset.
One business which has made its main aim to maximise its customer data pile at all costs is Amazon.com. Their CTO, Werner Vogels, spoke at Telco 2.0 yesterday. Amazon believes that its huge customer base, and its vast resource of data on their purchases, is a crucial asset. Similarly, it makes it its business to maximise its holdings of information about upstream customers -- that is to say, its catalogue. Vogels described their decision to make listings free and to open the platform to upstream customers (merchants in Amazonspeak) as being explicitly intended to increase the information pile. Eventually, he said, they aim to treat each customer as a segment to themselves. In a way, Amazon is a machine for generating and matching user profiles and SKUs and then settling the transactions that result..
In a similar way, Voice & Messaging 2.0 is all about reducing the minimum segment size it's possible to develop services for, right down to individuals. But all of this is dependent on respecting information sovereignty: if you want to create passionate users, and even more so developers, you've got to respect the work they put in creating all that data. Which means not being evil, and providing a user interface to make that data shadow manifest and controllable, and providing the APIs and terms of business that will help your upstreams to invent things with it.