Yahoo! Talk

posted by alex

After an invitation from Chris Plasser and Caterina we headed out to Yahoo! Innovation yesterday to give a talk about trust and our ongoing research. We based the presentation and accompanying slides around four different stories that all relate to trust in various ways. The first one, about eBay ratings and reputation sheds light on the rating bias and also the re-remapping of reputation that takes place. The multi-dimensional phenomena of reputation travels via a biased one-dimensional scale to becoming a social cue rather than an objective numerical value and is thus in a sense “taken back to humanness of a new multi-dimensional phenomena.

The second story, about post-ratings within intimate google groups and the conflict of large scale systems and informal interactions brings up Tönnies concept of Gesellschaft and Gemeinschaft. These two sociological “normal types” serve as an interesting framework for looking at some of the tensions within social formations and the slides lists a few dichotomies between them. The third story concerns trust as a transitive property and uses Granovetter’s ideas about weak ties to ask if transitive trust is at all useful in the scenarios where it might exist. The fourth and final story shows an example of some fine “googlestalking” and also how Internet users find new ways of interpreting the digital body and assessing social cues in order to establish trust.Apart from all the fun we had talking to the Innovation TechDev group and the always insightful Caterina we also got the chance to meet up with Yvonne French, senior product manager for the Yahoo! reputation platform, to discuss Yahoo’s point of view on trust topics. Yvonne had a good point that is worth remembering; we know that reputation is context sensitive but that doesn’t mean that the same reputation can not be used in several contexts. The issue is, of course, about finding the contexts where reputation can cross successfully. Furthermore, we can never remind ourselves too often that a context is not defined by an URL–a website might encompass several contexts in the same way that a context might entail many websites.

To finish off the already great day we met up with Tom, Ted and Bill from Opinity for a relaxed yet insightful talk about online identity, trust, philosophy and even some sci-fi… Among other things, we talked about third-party institutions that have the possibility of acting as a trust-mediator in order to guarantee a certain fact about somebody without the necessity to disclose the fact itself to the asking party. I.e. somebody could ask a trusted third party to verify that my email is indeed connected to a real person with a valid credit card without me having to reveal my credit card number or my actual name to them. In many cases this makes sense since the asking party might not be interested in the actual information per se but rather just in the verification of its existence.

On another note, we are hoping to do a small talk about trust tomorrow at BarCampStanford so if you’re in the neighborhood please come join us!

My Friend’s Boss Has Little Relevance To Me

posted by eric

trustmail.jpgSort by trust in TrustMail.In this paper Jennifer Golbeck and James Hendler outlines a “reputation network analysis” system for email filtering. The concept is simple: rate everyone you know on a scale from one to ten, and your email application will calculate a “reputation score” for each and every incoming mail message (using a local trust metric algorithm and some FOAF-magic).

Wouldn’t it be great if you could sort your inbox by “relevance”? I thought so too, but then I took a look at my own inbox while trying to think of a rating for each mail that was there. And guess what–it turned out to be quite hard. In fact, in some cases I couldn’t come up with a rating at all, and in other cases I could think of several different ratings for the same person.

Why? Well, we know already that reputation demands context. So whenever I thought of a rating, I had to artificially place the person in a context, which felt somewhat awkward, arbitrary and at times plainly wrong. Which context to choose? What if there were multiple? What if there was none at all?

This is also the reason why the system probably wouldn’t work well anyway. Reputation scores that are inferred by the algorithm are calculated without taking context into account. So let’s say my friend’s boss sends me a mail. I trust my friend, my friend trusts his boss. And yet, the mail from my friend’s boss has little relevance to me. Both reputation scores are accurate in themselves, but when we collapse the contexts around them, they loose their meanings.

So, we need a way of saying “I trust you in this particular context”. But then we run in to issues of etiquette and fingerspitzgefühl. By saying explicitly “I trust you in the context of work” I’m tacitly saying “I don’t trust you outside of work”. Apart from the high cognitive load, I’m actually being rather tactless.

Needless to say, given that I couldn’t even trust my own ratings of my own friends, I’m a bit skeptical of the Trustmail approach. The same critique is valid for many other reputation systems. Clay Shirky puts it rather harsh but well: “Almost all the work being done on reputation systems today is either trivial or useless or both, because reputations aren’t linearizable, and they’re not portable. […] The world’s best reputation management system is right here, in the brain”.

August 21st, 2006
0 Comments    

Reputation and Fingerspitzgefühl

posted by eric

Reading texts like “Manifesto for Reputation Soceity”, “Toward a Private Digital Economy” and re-reading the whuffie parts in “Down And Out in the Magic Kingdom” has made me think again about reputation. So many questions! Where to begin? Are we moving towards a sort of hyperreal version of Tönnies Gesellschaft, or a cyberspacified Weberian Iron Cage? Can we position all people on the one-and-only Great Reputation Graph, or are there in fact several over-layering, intersecting, and incommensurable systems of reputation? Is reputation portable? How to measure Halo?

We talked with Auren Hoffman (CEO, Rapleaf) about their rating service the other day. They want to break out the buyer-seller-rating component of eBay–a bold act that is not without problems (Oh no, the silos fight back!). After all, one of the most important preconditions for building reputation is context. Like Tom Dell’Aringa puts it in a comment: “I can honestly tell you that a Rapleaf score would mean zero to me in eBay. I don’t care about what some 3rd party rep system states about a person on eBay. I want to know what their rep is inside eBay!”

I have a hard time figuring out whether RapLeaf’s very Web2.0ish service (api:s, openness, componentization, you got it) is the right way to go in this case. I believe dealing with reputation takes some fingerspitzgefühl. This is a tip from the RapLeaf blog:

“Going on a date? Worried you’re on the Dontdatehimgirl website? Offset that! Send her an email with your RapLeaf badge in the signature. She’ll see that you’re a good guy.”

Wow, isn’t that like putting your credit rating in a love letter? Or am I getting something wrong? It reminds me of the story about the woman who wants to go partying with her friends. Her husband says: “I trust that you won’t cheat on me”. This story illustrates very well what happens when talk about trust gets to explicit. In this case the woman gets seriously offended–saying something like that is a clear sign of distrust. [My supervisor told me the story. I believe he found it in Lagerspetz]

A Colombian artist told me the flipside of this story yesterday. Some time ago she walked into a store–in a small village where she’d never been before–to buy some milk. When she wanted to pay she realized she didn’t have any money on her. The owner of the store immediately offered her credit. “How can you offer me credit when you don’t know me?” was the obvious question to ask. “I’d rather lose my money than my trust” was the owner’s swift reply.

A village, a marriage, a date, an eBay auction. Context seems important. Fingerspitzgefühl too.

[UPDATE: Noted that the part on the date quoted above has been removed from the blog post over at RapLeaf. Good move.]