On VRM and Standards

Phil Whitehouse recently served up some nuggets to stimulate conversations at next week’s VRM2008 in Munich.

I’ve been thinking a lot about VRM lately. Not so much about what it means, but rather the mechanism of how it can work.

If you’re new to VRM, it can be summarised like this: it’s the reciprocal of CRM. Rather than being bombarded with advertising, much of which is irrelevant, and the rest irritating, wouldn’t it be nice if you could just tell vendors what you want, on your terms? Without even going to the trouble of looking for them? If they’re willing and able to respond, they do so. Everyone else goes on their merry way. It’s all about sharing the data you want with the people you want.

Some examples from Doc Searls (Cluetrain Manifesto dude), who heads up the VRM project, include:
I want to:

– Buy a power convertor near St.Paul’s in the next three hours, at any price
– Buy a stroller for twins near Highway 70 in Kansas today for under $300
– Buy an Apple laptop with a 500gb HDD and weighs under five pounds, as soon as it comes out
– Buy a double decaf cappuccino at the next exit on this highway
(You can see more examples presented by Doc on this photo)

There are a few big problems that need solving. Filtering is one (both on the outbound request and the way back in), targeting is another (how do you choose which vendor to share your data with?), organisation is a third (by what mechanism do customers agree to share their data, and in what form, while retaining control over it?).
…

don’t know much about establishing standards. My erstwhile colleague Paul Downey, on the other hand, represents BT at the W3C and thus knows a bit about standards. He sez this will be a hard problem to crack, and he’s probably right. Big question: to what extent would we, the customers, allow brokers to help create this standard?

My view is this problem needs to be overcome before VRM can move forward, regardless of whether brokers are involved.

…

Good stuff. As the chair of the Standards Committee for Project VRM, it might be obvious that I think we need to create some standards.

At the end of the day, interoperability requires either standards or one-to-one interoperability engineering. The user-centric Identity movement has grown like crazy in the last few years largely because a hybrid of these approaches have been used, as OpenID, Higgins, CardSpace, and Liberty (among others) took their 1.0 products and figured out how to make them work together, leveraging standards like WS* and SAML as they did so. The nice thing about standards is that once they are in place, they reduce an O(n^2) problem, where every software vendor has to coordinate with every other vendor, to an O(n) problem where each software vendor coordinates to the standard.

The problem with standards is they are slow to develop. But once you have some apps and some standards at the 1.0 level, the efforts towards interoperability can get serious traction, like they did with the user-centric Identity movement.

I’m hoping we can engender a similar development cycle with VRM. We need both working applications and formal standards and specifications, especially with regard to data formats and communications protocols.

I’ll diplomatically disagree and agree with Bart (read his comment on the original post) regarding leaving standards to others. On the one hand, we should leverage existing work as much as possible. For example, I see Higgins and XRI playing a major enabling role for us. On the other hand, while the Dataportability and Higgins guys are doing great work, they are not necessarily solving the problems VRM has set out to tackle, namely reinventing the marketplace on behalf of individuals while creating more value for vendors.

As an example, the Dataportability movement has framed the problem in terms of Data and Portability. This brings to mind exporting and importing “my” data from vendor to vendor. That’s a start toward liberating users from vendor silos. However, I think the real win is in user-centric services, where the location of the “data” is essentially irrelevant–even as it is hosted under the control of the user–and all user-authorized vendors can access the data through approved services.

That’s the idea behind the Personal Address Manager, which we’ll be discussing in Munich. Your actual postal address isn’t that much of a problem from a dataportability perspective. It’s just a few lines to enter and no real need to “export” it from some vendor’s silo. However, when you change your address, it would be nice for the new address to automatically propagate to those authorized to get it. Or, for more sophisticated vendors, to have the address provided on demand, so that they never send postal mail to the wrong address. Such a service would be automatically discoverable by vendors using the Identity layer to authenticate and authorize exactly who gets it.

I see the job of VRM as working through these scenarios from the user’s perspective and ensuring the development of enough standards and technology for a complete implementation.

In any case, I’m looking forward to seeing Phil and Bart at VRM2008. There’s plenty of room to continue this conversation. Join us if you can; it should be fun. =)