wanderingstan

Tim Berners-Lee is talking about the Social Graph. This is good, but he and many others are conflating the age-old dream of the “Semantic Web” with the newfound value of the “Social Graph”.

“Social” has come to the web before “Meaning”, because people care more about “social” than “meaning”.

social > meaning

Who said something is infinitely more interesting than what has been said.

What does it benefit a user if the computer can understand the meaning of “Buy Viagra now for cheap! Enlarge your penis!“, but it doesn’t know who said it?

To put it in pictures:

Is it more important that meaning be derived so that the following RDF triplet can “understood” by a computer:

Or, is it more important that the computer realize that this message (no matter the “meaning” or how structured it is.) came from a person/company/source that you don’t know, don’t trust, and have no connection to? (e.g via your social graph)

This is why we have applications used by millions of people for throwing food and poo at friends, but no break-through semantic web applications.

(You might argue that people don’t care, that Wikipedia for example is not by any one person in particular. But as I said before, these sorts of facts are not the ones we care about most. You’d much rather get a pie in the face via a Facebook friend than a well-written and structured and meaning-laden bit of text that you have no relation to. (And to be more precise: Wikipedia is a trusted source/brand/company unto itself.))

Back to Berners-Lee, he sees the social graph as exciting because people are finally starting to map something and maybe now they’ll start to map everything else. My criticism is that he views social and trust relationships as only one example of relationships, no more or less important than other relationships. (Like the one in the Viagra-to-Cheap relationship above.) For example, he writes:

So, if only we could express these relationships, such as my social graph, in a way that is above the level of documents, then we would get re-use. That’s just what the graph does for us.

But social/trust relations trump others. This is what I said 3 years ago when beginning Outfoxed, and I’m happy that Facebook and others are proving it true.

Here at Lijit (the company which grew out of Outfoxed), we surface this idea in Search. But it will eventually permeate every aspect of our online lives: message exchange, discovery, shopping, advertising, religion, and more.

We have seen the rise of vertical search engines (e.g. WebMD). I’ve explained that Lijit turns each person into a vertical search engine based on trusted social connections.

But what about advertisements?

For most internet sites, you put AdSense on your site and hope the checks start coming. But you have basically zero control over what appears.

I had coffee on Monday with Matt Gerson of eConscious, who told me the story of this article about the dangers of teflon. When they had AdSense, it would show ads for teflon products. An article about why teflon is bad, and it shows advertisements for teflon products. The readers were understandably confused. AdSense isn’t there any more.

On the one hand, this violates the ethical standards of the site. On the other hand, this is very poor targeting! Everyone loses: the network gets no money, the publisher wasted valuable space, and the reader is distracted and confused.

Thus the rise of vertical ad networks like GreenAdWorks.

But what I really want is an Ad network that works like Lijit does for searches. I want a network that won’t show ads for products or companies that have been rated poorly by people that I trust, people in my social networks. I want an ad network that hilites products and companies that have been positively rated by people that I trust.

So if someone is reading a post on this blog about startups, I want to show an ad for a book about startup’s rated highly by Todd on Amazon.

This network doesn’t exist. I hope someday it does.

See also:
Social Media and Advertising

CNET reports that Email is, like, sooo dead.

When I was in Germany for grad school I was surprised when a classmate said he stayed in touch with faraway friends primarily via IM. “That’s odd”, I replied, “most of my friends aren’t on IM. They just email.”

His reply was short and to the point:

“That’s because you’re old, Stan.”

Ouch.

But it’s true. Young people do most communication via texting, IM, and social networks.

Eric Rice suggests that this is because of “social authentication. I have to approve you as a contact before you can communicate with me. If I approve someone who is a spammer, I can drop with a single click.”

I agree, but this also makes me sad. I don’t mind being contacted by non-friends (or long lost friends) who have a real reason for contact. And I don’t want to have to personally approve each potential contact. That sounds like a lot of work!

The solution is to delegate this “social authentication” out to people I trust. If someone is a friend of my friend, they should be able to contact me with no problem. Likewise, if a friend of mine has identified someone as a spammer, they should be automatically blocked for me. Visually it would look like this:

(That graph comes from my old Outfoxed days, in Keeping your network clean.)

The funny thing is that none of the big social networks employ this technique. LinkedIn comes the closest with their “ask a contact for an introduction to one of their contacts”.

Similar social networked trust solutions could be used for comment spam, product reviews, and even code execution.

That’s my prediction. We’ll see how this plays out.

TechCrunch reports today that Mitch Kapor’s Foxmarks will become a search engine, and that it wowed everyone at foo camp.

All search engines need algorithms for ranking pages. Google ranks pages based on links from other pages. Foxmarks will instead look at links in bookmarks.

No search engine has done this before, so it seems quite natural that their is a “wow” factor to the results. They’ve found an unpolluted spring, a fount of trustworthy metadata. That’s the way HTML links were back in 1996. That’s before there was money to be made from having the right constellation of links pointing to your site.

This is not the first time a search engine has been created using bookmarks. I wrote this in 2005 about Zniff:

A new search engine, Zniff, takes a step in the right direction by using publicly available social bookmarks as indicators of worth. Paradoxically, this approach is doomed to fail if it enjoys any success. If it becomes popular, it would be all to easy for tricksters to create false bookmarks for the sole purpose of inflating the ranks of chosen pages. It’s the same lesson that Google is learning now with googlebombing: You can never trust random pages on the internet. Not even social bookmark pages.

I sure was surly back then!

How solid is the “doomed to fail if it enjoys any success” principle? Even though links are no longer so trustworthy, Google still serves up good results. Will Foxmarks be able to continue wowing people once the SEO bad guys start making fake bookmarks?

Of the key insights of Google was to realize that links were not just pointers or bookmarks: they also implicitly convey an endorsement, a “vote” for the pointed-to entity.

With the advent of blogging and social networking sites, there seem to be 3 distinct flavors of links emerging. New generations of link analysis tools must be aware of these flavors in order to fully understand what information is being conveyed by the linkmakers. Lots of people seem to be talking around this and even building tools to manage it (Ziki, Wink, etc…) but I haven’t seen anyone spell it out, so here’s my shot at it.

1) The Pointer Link

This is the link we all think of, the link that was envisioned back when they were called “hyperlinks”. It’s basically a citation, like an footnote from an academic journal.

English translation: I examined the exact content I am now pointing to, and I found it to be worthwhile. It’s worthwhile to go there.

A normal pointer link in a blog post.

2) The Subscribe/Friend/Trust Link

These links are the hallmarks of RSS and social networks. These links are not endorsements of the exact content pointed to, but of the author or creative source represented by the link.

English translation: I often examine content that comes from this source, and I find it to be worthwhile. Further content from this source is probably also worthwhile.

Note that this type of link talks not only about past content from the author, but is also a prediction that future content will also be worthwhile. (As Seth once said about subscriptions, it’s a promissory note for attention that will be paid in the future.)

Blogroll on Guy Kawasaki’s blog

My friends on MySpace

3) Identity Link

These links are maybe the newest to grow in importance. With the proliferation of content creating platforms, it’s becoming more difficult to tie your online identity together. People must now create “identity indexes” that list who they are on different services.

English translation: “I create the content at this source, and take full responsibility for it.”

Guy Kawasaki identity-linking to his page on LinkedIn

Fred Wilson identity-linking to his del.icio.us profile

The identity-linking Wink-widget

Anti-Links

If I had to bet, I’d say the next important step in linking will be anti-links. In the real world, you also often recommend against things and clarify who you are not. For example:

• “Did you hear that Mary is now into multilevel marketing?”
• “Scrubs used to be funny, but now it sucks.”
• “My name is Michael Bolton. … No, not the singer.”

There are only a few parallels to this in the online world.

• rel=”nofollow” : This microformat in a link indicates that the author does not endorse what is pointed to. It was developed to solve the problems of links in blog comments. If a commented leaves a link, the author of the blog does not want to endorse whatever the commenter linked to. (Okay, not really an anti-link, more of an un-link.)
• rev=”vote-against” : This microformat in a link indicates that the author activly recommends against what is pointed to. It is not really supported, however.  (Except by Lijit!)

The most widely used anti-links come things like the anti-phishing features of Firefox and IE, the “dangerous” sites of SiteAdvisor , or the ad-blocking lists like Filterset G. (Why do we allow anyone with a web page to create positive links that contribute to (for example) PageRank, but we leave the job of anti-links to “professionals”?)

As more people get online and more and more web services are created, there are bound to be more username collisions. For example, this bfeld is certainly not this bfeld, and this stan is not this stan!

Conclusion

Google changed the world by realizing and exploiting the extra semantic information in those newfangled “hyper links”. I think successful companies of the future will learn to read a new generation of links that are even more meaning-laden that their forbears.

Of course I hope Lijit is one of those companies. We’re certainly on the right track.

The graphical shorthand we use expressing these relations and to work out complicated situations:

John Hagel wrote an excellent post, The Aphrodisiac of Attention. I insist that you read it immediately.

Trusted filters, trusted protectors, trusted concierge, human or technical, removing distractions and managing boundaries, filtering signal from noise, enabling meaningful connections, that make us feel secure, are the opportunity for the next generation. Opportunity will be the tools and technologies to take our power back.

Technology will certainly be important, but I remain convinced that technology alone will not maximize the return on our attention. We will need human intermediaries to harness the technology and adapt it to our changing needs. Some of you may remember the concept of the infomediary that Marc Singer and I wrote about in Net Worth. That concept is resurfacing in interesting ways these days.

Trust, filters, social networks, information overload. These are the important ideas.

SpikeTheVote is a way for people to game the Digg voting system. It’s very clever.

We collectively vote each other’s stories to the front page.

I act as the middle man, verifying votes and keeping everyone in line. If someone stops digging, they won’t earn enough points to get their own stories dugg.

Spike the Vote works on a point system. Each day I give you a mission with several stories to Digg. 20% of your mission involves digging stories submitted by users in this community, while 80% of of your mission is completely random. This is to eliminate footprints and keep things anonymous.

You earn 1 point for each story Dugg. Once you earn enough points, you can trade them in for Diggs on your own stories.

I agree with Fabian that this isn’t inherantly evil, but it’s obviously not something that Digg or it’s users want.

This is the same root problem that comes up again and again on the internet. If all it takes to get a “vote” is to fill out a registration form or create a link, then any “election” can be gamed. The same goes for product reviews and search results.

What’s the solution?

The feature is in it’s infancy, but the Lijit List aims to solve exactly this problem. It combines a Digg-like voting system with a social network. Votes from people closer to you are given more weight than votes from strangers.

This should also avoid the very Digg-ish problem of nerd-centricity. The core users of Digg are nerds (like me), and the content reflects that. But the Lijit solution means that your list is customized to you and your network. If lots of your friends are into knitting, then you can expect to see a lot of stories about knitting. (I do not expect to ever see a story about knitting on Digg!)

Like I said, this feature is just getting going. We need more volume in the system to really get it tuned, so why not join the beta test? Be sure to let me know how it works for you.

A great cartoon from Gaping Void. But the text that inspired it is even better:

Trust used to be something that bound small groups together. Over time we tried to scale trust. It didn’t scale. And what happened instead was Big Everything. In an Assembly-Line meets Broadcast world.
Big Everything broke trust. Big Media lied. Big Content Producer reduced our choices. Big Pipe and Big Device reduced it further. Big Firm wrongsized away. And Big Government did what it liked.

I urge you to run, not walk, and read this excellent post from JP Rangaswami immediately. Link

“Big Everything” is right. We’re just supposed to blindly trust Google that they’ll be neutral in search results, to blindly trust Norton Spyware checkers that they can’t be bought off, to blindly trust that book’s 3.42 stars average from all Amazon users, to blindly trust all other delicious users to tell me what is interesting. We can do better. We can scale trust. We can find middle ground between individual trust (e.g. blog subscriptions) and whole-world/Big trust (e.g. amazon ratings, digg, delicious-popular, Norton, PageRank).

Half the challenge of explaining Outfoxed to people is convincing them that its actually a more natural way of doing things than what we’ve been conditioned to accept as normal in the online world. What I want to show is that the online world has automated only our least preferred methods getting information, and it’s time someone automated the most preferred.

Imagine for a moment that your car breaks down while you’re on a business trip out of town to Springville.  It’s easy enough to find a mechanic online or in the Springville yellow pages, but how will you find a good one? Since you know nothing about Springville, you’ve got three choices:

1. Rely on an external trusted authority. Look for mechanics that are endorsed by the Springville Better Business Bureau, or by the American Association of Auto Mechanics.
2. Rely on the wisdom of the crowd. Look for things like “Voted best Springville Mechanic 2003-2006″ or “One of top ten trusted mechanics in Springville by Springville Times poll”
3. Rely on your network. Call your spouse and ask if she knows anyone in Springville; turns out her old college roommate lives there. Call the roommate and get a great mechanic recommendation.

Assuming all of these options threw out some recommendations, which would you take? If you’re like me, your network trumps options 1 and 2 every time.

But how do things look in the online world? What sources can you turn to?

1. Rely on an external trusted authority. Look for the “Trust-e” or “BBB” or “Verisign” seal on a page.
2. Rely on the wisdom of the crowd. If it’s a book you’re buying, look at how many stars it has on Amazon. If you’re on eBay, count the stars of the seller. 
3. Rely on your network. Call your friends and ask if they’ve heard of this website, or if they know this seller on eBay. 

The thing to notice is that the first 2 options have been completely automated by the web. External authorities were more a thing of the 1990’s Web 1.0 world, but currently Web 2.0 is positively drunk with wisdom-of-the-crowd techniques. It’s so easy! Just give every user a way to vote, and you’ve got instant crowd statistics. This is digg, this is delicious/popular, this amazon and IMDB ratings, this is YouTube stars.

But the network option sounds strange here, doesn’t it? It’s odd to bug a friend over a mere webpage, and the odds of them knowing one particular seller on eBay are staggeringling small. The problem is that this option is not yet automated. To use your network, you have to use the old slow methods of telephoning and emailing, and you have to bug a lot of people.

Source	Real-world	Online equivalent
External Authority	Better Business Bureau
Wisdom of Crowd	Newspaper Surveys
Social Network	Call your friends, ask for recommendations	???

Of course, social networking sites like MySpace, Facebook, or LinkedIn could step into that gap, but they seem to be content with only helping people find dates or business contacts.  In my next post, I’ll talk about how the current crop of sites are missing the boat.