Trust and Social Networking

Social networking on the web is hot. Many sites talk about trust. They tell us something about trust in cyberspace—and about trust in general.

Think four archetypes: Zagat’s, Amazon, eBay, and LinkedIn.

1. Zagat’s started off-line. It presents reviews by the "regular folks"—we trust it because the reviewers are presumably unbiased, in a way that commercial directories presumably aren’t. Zagat’s gains trust through clean motives of “people like me.”

2. Amazon has a recommender system—we trust it because we believe their massive databases dependably predict our likes.

3. eBay developed seller ratings. You can trust the person you buy from online because of their reputation among other previous buyers.

3. LinkedIn networks you to others by degrees of separation (a la the Kevin Bacon game) to others. You know Bill, I know you, ergo you can introduce me to Bill’s cousin’s friend. This works on the principle of personal relationships.

The social networks recombine these four types of DNA. combines movies, music and restaurant reviews with an “x-degrees of separation” taxonomy. LinkedIn meets Zagat’s. measures the “buzz” around products, get “points,” and also allows you to rate “trusted” reviewers. LinkedIn meets eBay.

Digg and review content a la eBay meets Amazon—combining massive databases and reputations.

Each type has drawbacks.

1. The “people like me” approach gets hard to sustain at scale; when you get thousands of “people like me,” they aren’t—they’re the crowd.

2. The crowd system—Amazon’s approach of masses of people dependably predicting our likes—depends on large numbers. Can the data be compromised? Sure. See the “let’s make my book #1” schemes, a la American Idol call storming.

3. The eBay-type ratings system is more obviously subject to hacking. See Wired for a good article on hacking reputation sites.

4. The Linked-In model has the great virtue of being based on super-high bandwidth connection—real people we actually know. The trouble is, personal trust has a high decay rate. It’s one thing to ask you to introduce me to your friend; it’s quite another to do so in order to meet his friend.

It’s in this arena that the internet falls down, even laughably, when it tries to promote trust. Consider Opinity and Rapleaf.

From Opinity’s site:

Create a powerful and portable Opinity profile and be trusted! Using your Opinity profile, you can bring your already established reputation to any new site you want. Build your reputation quickly and gain trust.

And you gotta love Rapleaf:

Rapleaf is a portable ratings system for commerce. You can look people up before you buy or sell, and rate them afterwards.
Rate people and they will be encouraged to rate you back. Before long, your Rapleaf profile will reveal you for the honest person that you are. After all, it is more profitable to be ethical.

Do you trust a website to tell you how much to trust someone who has had his friends email in to say how trustworthy he is so that he can make more money by appearing to be trusted?

Lessons from all this:

1. The best trust isn’t very transferable
2. Deep digital trust is a tradeoff for breadth; trusting you to sell me a book doesn’t mean I’ll introduce you to my daughter
3. Saying “I’m trustworthy” means you aren’t
4. As in the meatosphere, all is not what it seems.