eBay founder Pierre Omidyar is the perfect host of a prodding discussion on the topic of online trust; as the only person I know who's made a dent in the problem (ok, maybe Jeff Bezos too), he's worth a ponder when he takes to task the new Global Trust Exchange with a thesis that perhaps trust is highly situational, and more exactly, that it depends only on the specific situation of the one specific transaction that is happening here and now:

if there was a technology that could somehow increase the level of trust in a society, that would truly be a wonderful thing. But the only thing that I've seen come close is a simple economic transaction, repeated in varied ways millions of times a day online, by and between millions of people. Each transaction reinforces the notion that strangers can be trusted to follow through on their commitments. The actual transaction is their reason for being.

The discussion has attracted the likes of Joi Ito and GTX founder Duncan Work, and, well, being the sort I am, that sort who's prone to injecting his uninformed self into the neighbouring café table's conversation, yes, I've chimed in with my own unqualified two bits with a reminder on Toshio Yamagishi's Lemons Market and a further foray into how perhaps it is not a metric of trust that sustains success in ecommerce so much as any (real or imagined) assurance of an avenue of revenge should things run sour.

And perhaps prophetically, I may get the chance to test that theory. Some may remember my example of spouses who seem to awake after decades to discover they've been betrayed, but this, I learn today, can also happen in business. It's not as icky a feeling as the matrimonial parallel, but it's icky nonetheless to have a third party show you how one of your most trusted suppliers has been secretly subverting your business, perhaps for years, and that's when it hits home that ... well ... there's nobody you can tell.

That's when the sense of total betrayal really sets in a knotted pit in your belly, when you realize you haven't the clout to affect their business success, either publically or legally. They've been merrily pissing away your business, and the most you can do is shop elsewhere. It's icky in spades.

And it's an illustration on the topic of trust: I have recommended this supplier gleefully for almost 7 years, sold their services to partners, friends and clients, never suspecting how wrong I could be. That's the flaw in GTX and the beauty of eBay in a nutshell: Now that I know the truth, if the world were an eBay-style public reputation system, I would get the chance to save others the betrayal, and my supplier would also get fair airtime to explain why they could care less; the eBay model, in ClueTrain terms, opens the conversation about trustworthiness (in the strict context of a particular set of transactions) and never tries to slap an indellible stamp of trusted on anyone.

It's also an example of trust in the blog ecology because this highly connected instant relatedness of the weblog does give some measure of eBay's conversations on trustworthiness, and the best example I can think of at the moment is to ask Technorati about the news item 'Coffee, Tea, or Should We Feel Your Pregnant Wife' which was blogged by over 60 sites within a short span of release, only a handful pausing to question the truth of it, but that handful were apparent because of Technorati and that gets us a lot closer to a perspective (consensus) on the truth than any of the credentials of any random one of those 60 authors.

It's not going to do much today only because this blog ecology of pings and consensus mining sites like Technorati is still (just look at it's name!) demographically limited to a narrow interest bandwidth. But what's going to happen (assuming it all scales well) when all this stuff is the mainstream, the first choice (or folded into Google) and J. Average Consumer suspects the reputation of a payment broker ...

To answer Pierre, perhaps yes, such a trust-improvement machine can exist and already does in a prototypical fashion that Ito might recognize from the Emergent Democracy, and perhaps the final word to the Duncan's Global Trust Exchange is the same as my final word to the b-blogs and intranets, which is that the concept is valid, but the population of recommenders in your bottle is just way too small.