9/27/05

More on Trust...

So there's more to the story with this whole Trusted Network thing.

While I do think it is oddly comforting to see a visual display of our RIA cronies on our profile pages (Flick's profile pic is my favorite), and I think aesthetically, the profile pages look much nicer now - having visually pleasing profiles is not the end game here.

And while I think it is helpful for folks browsing the site to get a feel for whether other people trust the author of a particular review, the expected benefits of this feature go far beyond that.

Here's the vision:

I think that our hand-picked Trusted Networks (those that you trust) can evolve into the foundation of a personalized recommendation system for every one of us.

Imagine if you could filter out the ratings and reviews of every hit and run poster with an agenda on RIA, and only see the recommendations of those who are connected to you via your personal trusted network, or via the trusted network of those you trust.

If this site had this functionality, you would not just be getting the unsolicited opinions of the entire universe of RateItAll reviewers - you would be getting the aggregated recommendations of only those folks who are connected to you by links of trust. This network could be extended 2 degrees out, three degrees, four degrees out, etc. - to whatever degree that you felt comfortable with and that provided sufficient data.

That's what we're shooting for here. By identifying posters that you trust, you are linking yourselves not just with those individuals, but also with the networks of those individuals. In technical terms, for every additional node on your personal network, the power of that network grows exponentially.

So what does that mean? I think in terms of RateItAll, it means that the more we are able to organize ourselves into trusted networks, the more value our community can provide in generating relevant, helpful, personalized recommendations of things like movies, TV shows, products, restaurants, travel destinations, music - and even more subjective things like politicians, blogs, and Web sites.

As many of the most prolific RIA users trust each other, the networked value of this system is not apparent yet. But I think you can see signs of its power. I've noticed several relatively unknown users showing up as trusting some RateItAll heavyweights. Clicking through to the profile pages of those unknown users shows that in some cases, that trust is returned. What can we deduce from this? One very plausible conclusion is that these little known RIA users have a real life / offline connection to the RIA heavyweight. In my mind, this makes the value of their opinions that much more significant, and makes me trust their reviews a bit more.

This is network theory in action. And I think maybe the coolest aspect of it, is that each of us controls our own Trusted Network. Only we can decide who we trust, and because RateItAll is a free service open to anyone, we have the power to pick and choose from everyone we know - online or offline - to build a powerful, customized network that we can learn from.

Stay tuned, as this is a work in process.... as always, comments are welcome.

2 comments:

  1. RIA is faced with serious credibility issues. While the site has only one purpose, to make money, it wants everyone to believe that users are respected as are their opinions. That is crap. Users can believe what they want on the issue.

    But now we have this idea of a "trusted network". The most obvious flaw for this is simple. Reverse logic.

    The importance of the "trusted network" (as stated by RIA) is NOT who trusts YOU but who YOU SELECT TO TRUST. So why list on each user page those who selected you? The proper (and more logical) thing would be to have each user page list which users that particular person has chosen to trust.

    So Castlebee's or irishgit's page would not list all the possible ass-kissers who chose to select them as trustworthy. Those user pages would instead list all the users they have chosen as trustworthy. This would allow for a more useful implementation of listing the "trusted".

    As I already stated on my own blog... Donovan is a complete moron. Corrupt and unethical. Some may disagree but for those who do agree and for me it is much more useful to know who that fool finds as trustworthy.

    Any moron can click that trustful box on a user page. Knowing that various users trust another user indicates very little. Especially if the user is an unknown entity. But it may be very informative to know which users have been "chosen".

    I could have advance warning of potential users to be wary of when I see the list of Donovans "trusted". The opposite is true as well. Viewing Castlebee's user page and seeing her chosen "trusted" would also indicate reviewers I may want to keep an eye on for quality posts. Castlebee is one of very few users who I respect. While we might disagree on who is trustworthy her list of "chosen" holds more weight than most. Especially over losers such as Donovan.

    Listing those "chosen" instead of those who chose you would also be a way to see if a user is just clicking "trustworthy" in an effort to kiss ass or otherwise gain favor.

    And this particularly item from the original post is striking:


    "I've noticed several relatively unknown users showing up as trusting some RateItAll heavyweights. Clicking through to the profile pages of those unknown users shows that in some cases, that trust is returned. What can we deduce from this? One very plausible conclusion is that these little known RIA users have a real life / offline connection to the RIA heavyweight. In my mind, this makes the value of their opinions that much more significant, and makes me trust their reviews a bit more."

    What can you deduce from unknown users clicking on a heavyweight? The most obvious and logical thing to deduce is that it is either an alias or a real life buddy. I have in fact already seen several aliases. The posting of "trusted" on user pages, in the short term, will expose several aliases. But those cheaters will soon figure things out and remove the checkmark and never click them again. At least when using an alias.

    As to the "real life" issue. It is laughable to say that it would have more weight in regard to trust. Use common sense here. If my wife was on the network and rated anonymously or rarely but chose me as "trsutworthy" that is not MORE significant. Any small use of common sense would tell you it holds absolutely no weight at all. And may in fact be detremental to the overall value.

    So a work supervisor like Donovan can get all his fellow employees to join the site. Most don't post much if at all but each one can click on him as "trustworthy". And that is more significant? hardly. Only a fool buys that. And a generally nice guy like Numbah... He gets a few of his real life friends to join and mark him as trustworthy. Is that suppose to be more significant? Drinking buddies rating a bud trustworthy. Get real. So don't buy that crap.

    RIA is in fact counting on people inviting friends to the site. And as the small cliques grow and join into large cliques the site hits will expand as will that BIG factor: revenue. But RIA knows damn good and well that a real life buddy rating another as trustworthy is about as useful or meaningful as a Sunday dump in the toilet. But alienating all that potential activity and revenue isn't going to happen.

    The bottom line though is simple. If the aspect of "trusted networks" are those users someone selects then each user page should list what users THAT user has selected as trustworthy and not the other way around.

    This would also be useful towards the "image" issue. Since these users trust each other it would be unlikely they would select a user with an offensive picture on their user page. And if a user they "trust" has an excessively large image currently in use they could recommend it be made smaller and load faster.

    ReplyDelete
  2. The site can also choose to list both those chosen by that user as trustworthy and those who them. But if only one is listed then each user page should list which users were chosen. A scale of trustwortiness should be used as well. Saying I trust Castlebee, Ignatius and Kamylienne is not the same thing as saying I trust Castlebee and rate her a 5 on that issue, I trust Ignatius and rate him a 4 on that issue, and I trust Kamylienne and rate her a 3 on that issue.

    ReplyDelete

Apture