Yesterday, I wrote about the trust levels of different types of searches:

• Fully Automated Search Engines, such as Google, are seeing the trust in the result slowly errode as they become victims of their own success at “indexing the worlds information.” People are beginning to question the ability for automated search results to truly understand context. This is highlighted in the study 2008 Digital Future Report by USC’s Center for the Digital Future:

“A higher percentage of Internet users reported negative views about the reliability and accuracy of information provided by search engines, such as Google. Slightly over half of Internet users — 51 percent — said that most or all of the information produced by search engines is reliable and accurate — down from the 62 percent who reported the same response in 2006.“

• Fully Human Based Search Engines, such as Mahalo, while engendering more trust that automated search engines, are light in content, and the trust only exists to the extent that the searcher trusts the editor. In this discussion, at the Digital Life Design conference, between Jason Calacanis of Mahalo and Jimmy Wales of Wikia, both attempt to apply trust to their editors differently. Jason both pays his editors and has implemented policies and procedures to improve the results. Jimmy, on the other hand, relies on the pride of ownership and participation to improve his results.

• Relationship Based Search: Lijit, the company I work for, is not a search engine. Rather, it attempts to use relationships and the inherent trusted based in those relationships to return results that are much more contextual. This was displayed in the search example from my post from yesterday.

So what does this mean? Today Marissa Meyer, of Google, was interviewed by VentureBeat. In this interview, she defines social search as:

We believe social search is any search aided by a social interaction or a social connection? Social search happens every day. When you ask a friend ?what movies are good to go see?? or ?where should we go to dinner??, you are doing a verbal social search. You?re trying to leverage that social connection to try and get a piece of information that would be better than what you?d come up with on your own.

This is an important definition. There is no way for an automated search to know that you hate horror movies when asked “what movies are good to go see?” There is no way a human powered search result will be able to answer that question, without the social context. (Which is why you see a lot of effort being put into Mahalo Social.) Lijit attempts to be able to answer that question, and Todd Vernon, Lijit’s CEO, does a great job out outlining how.

But, in truth, none of these solutions can do the perfect job. Google doesnt have the social context. Mahalo and Lijit dont have the wealth of content. So, what is the answer?

Simple, combine some element of all three components.

• Google: the sheer ability to collect the worlds information is staggering. As it begins to realize that GMail and other social indicators improve search results, it will attempt to answer the question programmatically. Unless HAL 9000 starts to run the algorithm. Google will never trump the human. Just like the fishing rod.
• Mahalo: Increase the trust in the editors themselves. Continue to do things like live track sporting events. Be the human selected bait.
• Lijit: Continue to expand distribution and improve content discovery. Be the expert who knows where to fish.

Imagine a site that had all three components. The automated collection of data to cover search queries that a human powered search engine cant. The application of trust, through the discovery of the content produced by editors, and the exploration of intrinsic and explicit relationships.

So the destination site would look something like this: it would be a mashup of Mahalo and Google, with the ability on Editor Profile and Mahalo Social pages to search the social content and social network of the humans, provided by Lijit, that are providing results. Human manage the top and topical search queries and Google covers the holes.

Now, I am no longer given the fish, and told to eat it. Rather, I am given a rod, human picked bait and the knowledge of where to fish. And, when I catch the big one, I am going to never trust another way.

Popularity: 9% [?]

Subscribe to this blog's RSS feed

Todd beat me to the punch a bit on this topic on his post Publishers! Don’t Give Up The Second Click!

I have spent the past couple of days thinking about search from a publisher’s perspective. I have had some interesting discussions with folks like Ryan Sholin, who spends a lot of time thinking about the journalist community within the publishing space, and Chris Brogan, who thinks about bloggers and journalists. I had lunch with Chris Sherman of Search Engine Land where among many things, we discussed the state of the search engine marketing world.

I came across this post, “is search getting better or worse?” where Jason Calacanis, CEO of Mahalo, a human power social search engine, discusses the reduction in trust of search results.

And, finally, there was this post, “The Fight for the Second Click” which referenced John Battelle’s post “Google Takes Aim at Wikipedia, Is Now Officially a Media Company” about Google’s KNOL project.

So, if this is true:

“A higher percentage of Internet users reported negative views about the reliability and accuracy of information provided by search engines, such as Google. Slightly over half of Internet users — 51 percent — said that most or all of the information produced by search engines is reliable and accurate — down from the 62 percent who reported the same response in 2006.”

and

“When asked to name their #1 complaint about the process (of search), 25 percent cited a deluge of results, 24 percent cited a predominance of commercial (paid) listings, 18.8 percent blamed the search engine’s inability to understand their keywords (forcing them to try again), and 18.6 percent were most frustrated by disorganized/random results.”

Then, in some ways, the search engines have become their own worst nightmare. In their attempt to index “all the world’s information,” Google has collect too much information. Perhaps by becoming so good at collecting information, Google specifically (and search engines generally) have surpassed the technical ability of their ranking algorithms to return highly relevant results.

Take this search as an example:

(click to enlarge)

The search for “nda” returns a varied amount of results. Yes, the first result is a wikipedia article on “non disclosure agreements,” but the rest are pretty varied. What does this lead to? A lack of trust, since the results are not exactly what I want (which makes me think that it behooves a search engine to lower the accuracy of the organic results to drive more clicks on the paid search). To the search engines’ credit, there is no way they can contextually determine what I am looking for when I type in “nda.”

Back, to that first result, Wikipedia. That is the exact reason that Google is working on KNOL. Because if their products are on the top of the search results, they own the first AND second click.

Todd’s post, Publishers! Don’t Give Up The Second Click!, certainly covers Lijit’s place in the battle for the second click, so I wont expound on it. After all, the re-search functionality of Lijit does absolutely capture a user, help them discover more content within your site or network, and drives pages views that a publisher can monetize, rather than a search engine.

But, there is an added benefit, that even the social search / human powered engines dont necessarily enjoy: Increased context, and therefore increased trust. Because I trust the publisher/journalist/blogger prior to searching, I am already in a mindset that I want to receive content created or linked to, by the trusted writer. Context and trust are inherent.

Mahalo search for “nda”:

(click to enlarge)

Notice that Mahalo has not written a page on “nda” yet. Thats fine, I imagine that they will write one soon, and not having a page isnt really an indication of a lack of trust, etc. I actually think for many things, Mahalo is great. Their added social components are great. But, I also know that if I am going to do certain searches, Mahalo is not the place for me.

In addition, what really concerns me, is that the line about no page on “nda” being written was hard to see. I would have love to see it bold or red. Also, look at the two first results being returned: the first was a link to a Mahalo page on LibiGel (”LibiGel is a testosterone-laden ointment that is meant to increase a female’s libido”) and the second was a link to a Mahalo page on Birth Control Pills. I suppose the first page will lead to the need for the second?

But, now I am not sure if Mahalo, like Google, understand what I am asking for. But, they have done a great job of owning the Second Click.

The same search on nda with Lijit. Now, Lijit is not a destination site, nor really a search engine in the traditional sense. But, I know Brad Feld, and I know he has written a good amount about nda’s. So, doing a search on his blog returns:

(click to enlarge)

Now, not only do I trust the publisher, but I will stay on his site for a bit longer to read through his content, and potentially, the content of his network. Is Lijit search perfect? Of course not. But, it is a more trusted solution than either automated or human powered search engines by definition. (Of course, imagine if human powered engines, like Mahalo, were to integrate the inherent trust that Lijit embodes. That might be a really unique experience).

So, who owns trust? At the end of the day, its the First and Second Click. I need to know the place I am searching on is trusted, and the results returned will continue that expectation.

At, the end of the day, that is how publishers will win. They will be bring back that level of trust newspapers and journalists used to own offline, and people will begin to read your publication more, and, more importantly, interact with the content and the publisher.

Welcome to Web 3.0. An online world that is the intersection of trust relationships, content discovery, social interaction and real people.

Popularity: 18% [?]

Working for Lijit, I spend a lot of time thinking about the collection and aggregation of data. The presentation of that data, the ability to sift through all that data, and the utility of that data.

Funny thing is that I am not a programmer or developer. Mostly, I spend my day thinking about how Lijit the product, can interact with other products to mutually improve both products.

Here is an example:

Tumblr. Awesome product. I have a tumblog at micahbaldwin.com, and a follow 30 or so other tumblogs. The minimalist design, short form post style and ease of use makes its a viable communication platform and will open up “blogging” to folks that dont want to build and maintain a full blog.

Because of the user experience, the content on people’s tumblogs grows quickly, sometimes I may post 4-5 times a day, I see others post at a much more rapid pace. Unless I stay on top of the data, there are often many good posts that I miss.

What’s missing? Some sort of search functionality. Not just the ability to search through posts, but to help me discover content from other tumblogs that would interest me, and I could grow my follow list. Reblogging (the ability to re-post someone else’s post) certainly allows the best of the best to rise to the top, but its very superficial and manual.

Enter Lijit. We offer much of that functionality, and the ability to better integrate in the Tumblr platform would certainly improve the results we provide.

Beyond interconnected search, the added benefit is that Lijit can use following and reblogging to help “rank” posts and people based on trust, which could aid in content discovery and relationship recommendations.

It is pretty clear that a relationship between Lijit and Tumblr would be mutual beneficial and provide additional functionality that would not take away from the core value of either product, but would be embraced and useful to users of either product.

Which takes us to Twitter.

After all, if Tumblr and Lijit would make a good match, wouldnt Twitter as well? It shares much of the simple user interface and ease of use of Tumblr. People use it…a…lot.

And while there are services that exist search through tweets and users, there is nothing like the interconnected search capability of Lijit, where I can search through my tweets and the tweets of my network simultaneously, as well as all the content in my and my networks combined social graph.

But tweets are different than Tumblr and blog posts. They are ephemeral, usually in context, and more and more part of a longer public conversation. Rarely, does a tweet stand on its own.

While there is intersections that Lijit and Twitter can have around the collecting, indexing and presentation of data, its certainly not as robust as other relationships.

Would I do a deal with Twitter? Absolutely. There is often ways to determine benefit to both parties without needing to implement a complete solution.

Which leads to social aggregation products. I have spoken to many of them about the value to Lijit to collect more of the user’s social graph, and the need for an interconnected search functionality to exist within their application. The biggest difficulty facing social aggregation products is the sheer amount of collected data and the rate at which that data changes.

At the end of the day, the best business development deals are the ones where both products are improved by the relationship, and therefore both companies are motivated to see the relationship grow and flourish.

SIDE THOUGHT:

After becoming quite the consumer of Tumblr, Twitter and several social aggregation products, I have also come to another belief: Tweets should stay in TwitterLand.

Because context is so important to understanding the follow of Twitter, it works best as SMS, IM in a Twitter client, or in a browser window. When mixed with Tumblr or a social aggregation product, Twitter becomes noise.

Popularity: 7% [?]