People Power Vs Google

106 Comments

Will a people’s collective be able to beat Google at the search game?

That’s the proverbial $64,000 question, and venture investors are trying to answer that by funding start-ups such as Wink, that plans to go live perhaps as soon as Thursday, according to Silicon Valley sources. The company has been in a limited beta since October, and today conducted a major update to its infrastructure and interface, according to their blog.

Wink is a search engine that integrates tag results from multiple sources such as del.icio.us, Digg, Yahoo MyWeb (and we’re adding more). As more services incorporate tags and user input, new pages are added to the Wink index and ranked using our TagRank technology to deliver the best results.

Wink is backed by some serious heavy weights: Scott Kurnit of About.com; Ron Conway; Reid Hoffman, Marc Andressen along with Venky Harinarayan (of Cambrian ventures) and David Sze (of Greylock).

Wink is not alone in taking on Google. Of course, there are others like Activeweave, Jookster, Kaboodle, and Rollyo which are taking “human cycles” and trying to come-up with better results than Google.

Yahoo’s MyWeb effort, so eloquently detailed by Erick Schonfeld in Business 2.0, and its recent acquisition of Del.icio.us are part of this “people versus Google” movement.

I am not so optimistic about this trend, because I have not seen the mainstreams get interested in the bookmarking and tagging as yet. Believers can take comfort in the fact, that it is early days, but I am not so sure. I think it needs a behavioral change on part of “searchers” who have become accustomed to simply typing their queries and then trying to figure out where to go. Sort of like bad cell phone coverage : we go used to it.

Of course, conspiracy theorist like me often wonders, what if search not only became good, but great and accurate. A lot of “spare page views” generated by hit-and-miss model of today can drastically reduce page views, and advertising opportunities. Of course that would be horrible for Google….

Back to Wink…. I am going to test out their offering once they go live, and update the post. You can of course contribute your thoughts.

106 Comments

David

I’ve written a 4-part posting on the topic of Yahoo and MSN v. Google. You can see it at blogation.blogspot.com. Part three talks about del.icio.us and Flickr. I’d love to hear comments from folks.

Shrikant Joshi

Interesting. Haven’t tried out Wink yet. Am using StumbleUpon.

I am impressed by the Stumble feature which allows you to ‘stumble upon’ random pages (depending on your choice of interests) ‘tagged’ by others … The firefox extension makes it all the more better.

Wink seems to be a Web2.0 version of StumbleUpon. And SU beats del.icio.us with its quick Like-it-and-don’t-like-it buttons… unlike tags opening in pop-ups. I think I’ll stick to SU, though I will definitely try Wink out…

So, does people power involve startups and acquisitions by heavy-weights (Wink, del.icio.us, etc.) only or does it also include welterweights like SU?

Shrikant

NOTE: I am no way related to SU, I just like the service…

franck

I think one characteristic of Google is its ambition. They always try to tackle very difficult problems. And their approach which actually consists in automatically tagging the web is clearly the way forward.

Obviously you get satifactory results by having power users tag the web content as evidence with successes like del.icio.us. In pretty much the same way you used to have satisfactory results by having a group of experts catalog the web (Yahoo! version 1). But the holy grail is do that for all content without forcing people to tag anything.

And this what Google is trying to do. They have just started and experiments like Google Music or Google Movies are showing great promise.

At the end of the day, however, we should remember that Google has an advantage: they can index users’s tag and as they become popular and make them play a more and more important part in their semantic web ranking. They can leverage any ontology. Either way they win!

wsjoung

well, there is a rumor, some people believe that a Chinese search engine will beat google soon or later. because they can hire thousand of people to tagging and searching meaning information through the Internet instead of crawler, and that will be much better than any other algorithm so far.
the brain of thousands of people! that’s the real intelligence, LOL!

eric goldstein

Om,

You have started a great conversation regarding People Power vs Google. And the remarks from Michael Tanne are very poignant. It sounds like they have a great vision for offering better search results. I wanted to offer my perspective and discuss how Clipmarks is trying to tackle this in a different way.

Your question “Will a people’s collective be able to beat Google at the search game?” as well as a recent blog post by Patricia Seybold about the need to “make sense out of what’s going on and what’s out there to discover” speak to the heart of the collective innovation that has been termed web 2.0. They also directly address the purpose of Clipmarks. In our opinion, the key to offering meaningfully better results will necessitate offering a wholly different kind of results. Whether you are referring to bookmarks or search engines, links to web pages are not enough. There is too much guess work and too much clicking back and forth before you find what you’re looking for. I just can’t imagine how that will ever not be the case.

So the question is, what would these ‘different’ results offer that make them meaningfully better? In my opinion, instead of solely offering links to web pages, they would connect people with 1) specific information they are looking for; and 2) the online sources of that information; and 3) other people who share a common passion and interest in a topic. To create this platform, we are offering people a tool that lets them easily clip and tag specific content they find within the web pages they visit. This content plus a link to its online source can then be saved as a clipmark.

Like bookmarks, clipmarks can be created solely for private use or for social purposes. The difference, of course, is that with clipmarks, people are able to see the information that has been clipped and then decide whether they want to click the source link to find more information. As a result, people are making a more intelligent decision about which links to click on and which sources to visit. If the content that has been clipped does not interest you, then you won’t click the source. However, if the content is compelling, you will want to vist the source to learn more.

We hope that people will visit Clipmarks as both a destination to connect with content and people as well as a relaible point of departure to connect with sources (web pages) about any topic that interests them.

So far we have received some great feedback and suggestions from our members and we’re doing our best to make the Clipmarks experience as good as possible. With that in mind, we’re going to be introducing a number of new features over the next few days, weeks and months (years and decades too hopefully :) ) but the current version certainly illustrates what we’re trying to accomplish.

I hope this has helped add a fresh perspective to the discussion about how to offer people a better experience on the web! Oh, and we don’t have Aeron chairs either, but i sure wish we did :)

eric

Damian

Michael – since I was the one raining on your parade, I feel it is only fair to thank you for such a great and thoughtful response. Clearly, you’ve got your head on straight and have a humilty which I sometimes find lacking in SV companies. And I’m glad to hear you don’t have the Aeron chairs!

Here’s my key issue with Wink so far: I can’t figure out what to use it for. Meaning, I’m not sure it’s complete, and I’m not sure I trust the results. Therefore I, as an ordinary consumer of search, would go back to Google.

Michael Tanne

Michael Tanne chiming in. Thanks Om – you started an interesting discussion going and there are definitely some good questions raised.

It’s fair to point out that user behavior isn’t easily changed. A place in someone’s busy day is earned by delivering value. Years of experience have shown that users want to type a simple search and get the results they want yet many people use multiple sources for different needs. I know many people who did without Technorati for years – they didn’t think anyone was saying anything worthwhile in blogs – but now can’t live without it. If users find what they’re looking for at Wink a few times, we hope they’ll use Wink as one of their sources.

By integrating analysis of user input – and tags/bookmarks are just one method, albeit a growing phenomenon – with other relevance techniques we tease out the best results we can find. Both Xen and Michael have correctly pointed out you have to do much more than do a string compare across tags and hope that people use the same tag terms. There is also a distinction between people bookmarking sites they want to keep around and rating results as a passing observation. There are many other factors we try to consider. People have generally been posting that they like the results they are finding so far.

About people power – it has already become cliche, but that doesn’t mean it’s not real. When memes travel at the speed of blogs, an idea can become wired, then tired then expired before the engineers have even checked in their code into subversion. Dan Bricklin spoke of the Cornucopia of the Commons http://www.bricklin.com/cornucopia.htm on Aug 7, 2000. just because too many “Web 2.0 companies” have launched (It’s even hard to type that phrase, it’s so cliche already) doesn’t mean that Dan was wrong, or that user participation isn’t really happening. There are many examples of applications (cddb, eventful, wikipedia) where the enthusiastic 5% provides most of the information – and receive some benefit that satisfies them, even if it’s just the satisfaction of helping, or perhaps some recognition, and the 95% enjoy the results of their efforts.

On starting companies – that’s a whole different post. Are there too many starting now? Probably. Are there similarities to 1999? You bet, we were there – I was a founder of AdForce and the others on the team were at Inktomi and Epinions. A few big differences – there is a vibrant and I believe real underlying delivery of value on the Web that is expressed through the marketplaces of AdWords and Overture, et al; those of us who’ve been through it before are more cautious, but know that every new venture is a risk; the cost of doing business on a dollar per useful bit basis have been cut by a factor of 30 due to the cost/performance of hardware and the productivity of open source according to Joe Kraus http://bnoopy.typepad.com/bnoopy/2005/06/its_a_great_tim.html; and in today’s world a company faces better odds taking their product straight to the users than attempting to cut an OEM technology deal. Most bus dev people I know currently favor real market success over technology and a deal structure. So I’d still advise entrepreneurs with user-oriented breakthroughs to take it to the people before taking it to the bus dev folks.

Nothing wrong with starting with the geeks – they’ve been there at the beginning of almost every new thing. We may refer to the “Tagosphere” now, but that text is editable as the market evolves. (No doubt that word will be cliche by the time I hit “submit” – I will note that Google reports just 173 pages with the term at this moment) We’re comfortable with the idea that this may be more attractive to early adopters at first. Over time the subject matter that is adequately tagged will expand and we will try to be ready for those markets. If that happens it will be because we have done our job making the service useful.

The real test is if people like our results and honor us with a few moments of their attention. If they give is that, we will do everything in our power to deliver something of value.

Oh, and I believe the market has shown that delivering more value to the user should never be sacrificed – so we shouldn’t worry that people will find their results too fast so there won’t be enough page views to show ads on. Things that deliver value grow.

(and we don’t own any Aeron chairs :-)

Thanks for giving Wink a chance to do something interesting.

Ross Hill

Wouldn’t it be silly to assume Google don’t track user behaviour in some form? One one would be to see which links on a result page get the most hits for each keyword combination.

Om Malik

rick teach,…, you can be my copy editor any time. thank you, correction made. appreciate the eagle eyes. resolution for 2006… less calvados, more time rereading/rethinking. thanks again

seekXL

“Nothing can replace Altavista. It rocks!â€?
“Nothing can replace Yahoo! It rocks!â€?
“Nothing can replace Google. It rocks!â€?

this gag is cool :D

Anona

“For this to be succesful(useful), it needs to be appealing to the mainstream.”

Del.icio.us was not useful/appealing to the mainstream, yet it was successful in the sense that Yahoo paid money to acquire it. Isn’t that want Wink is after?

Suranga

We’ve looked a lot at tagging and community-driven search/ranking at blinkx and while we still might support it more explicitly in the future, the trials we have done thus far appear to suffer from a level of what I call ‘Bay Area bias’. The problem (as you identify, Om) is that only certain types of people are interested in cataloguing the web and bookmarking/ranking pages within it, so those people have a massively disproportionate effect on things … if you look at the world through spectacles similar to those worn by these trend-setters, everyone’s happy, but if you don’t, it can be difficult to find the right thing.

Having said all of that, Google’s early relevance success was essentially down to analysis of linking affected by webmasters across the internet — hardly a demographically diverse group, after all — and that worked really quite well. So-oo, look forward to seeing what what Wink can show us.

Incidently, Ramesh Jain has some interesting views on this sort of thing, by the way: http://ngs.ics.uci.edu/blog/?p=402

Richard Moriarty

I’m greatly interested in the concept, however I think we’re going to find only geek/web 2.0 use of Wink, and that means it’s only going to fill a small niche. For this to be succesful(useful), it needs to be appealing to the mainstream.

Jim Gilliam

People will tag and use the social bookmarking sites as soon as they are integrated into the browser and people don’t realize they are doing it. Flock is important.

Brady Joslin

Yeesh – any service that introduces itself with the word “tagosphere” is clearly making a product for geeks, not for mass consumption.

Jeremy Pepper

Well, we can look back at search history and remember people were saying the same things back then.

“Nothing can replace Altavista. It rocks!”
“Nothing can replace Yahoo! It rocks!”
“Nothing can replace Google. It rocks!”

It’s likely that something will replace Google, but it’s going to be stealth like Google was (and, well, Yahoo! as well). For all we know, there are two kids working in a garage right now that we just don’t know about, who are creating the Google competitor.

Anona

“INHO Wink has 3 disadvantages:”

Add a fourth: Marc Andressen, the kiss of death to any venture. :-)

khabri

In 2007 we find out that two of these engines are really gaining market share and somewhere in Redmond a bald old man is seen swearing at his employees “I am going to fucking KILL….” and tossing chairs at them.

RYK

I think we’re already seeing people power taking on Google: slowly and surely Wikipedia is rising to become the world’s no. 1 information source

kaveh

I’ve written a few lines on tagging here . in short, I think way too many products are starting to rely on ‘people power’. if we agreed, soon, we’d all be working for ‘web 2.0’, instead of ‘it’ working for us.

having said that, I do like the fact that wink (1) allows users to rank a given item in a result set and (2) makes grouping+saving+sharing of a result set. however, these are merely features that google & yahoo! can add (are adding?) … no need to start a company around it.

Michael

Thanks for bringing this subject up Om. As the founder of Seekum this subject area interests me very much. As Xen pointed out what one finds interesting others may not, but I believe that things like tag density, and user votes will overcome this issues as more people are involved. I agree that not everyone is going to be tagging pages, or especially bookmarking pages in the sense that Wink needs them to be bookmarked. Wink needs users to bookmark practically every page they go to. To me I am still stuck with a “Favorites” mentality sugesting that I only bookmark sites that I would go back to or even worse really like compared to similar pages. Thats why when building Seekum we use an up down philosophy. The page is either relevant to a search or not. I’m not sure if these other more complex algorithms will help search results in the end. I think users in mass just by voting will be able to though.

Xen Dolev

INHO Wink has 3 disadvantages:
a) Ranking pages is a very subjective matter – what one finds as relevant info, others might find irrelevant.
b) It’s almost impossible to get different people to tag the same page with the same tags, even if there are a fixed set of tagging words.
c) Most people won’t start work as librarians and index web pages…

But don’t get me wrong, I’m very much in favor of more accurate searches!

Damian

Excellent post Om as usual – a few comments:

1. The hardest thing to create in today’s world is a new brand. Google successfully created a top 10 brand. That will be hard to replicate.

2. As you point out, Google tapped into current user behavior. All of these other search engines are attempting to alter user behavior – which is much, much harder to do.

3. Finally, there has been such a flood of these companies, all fairly undifferentiated, that it will be very tough for them to garner significant public attention. This is a little bit back to the brand issue, but, more broadly, they will have a hard time driving the traffic. That means spending more on marketing (sound familiar) or becoming a technology provider. I believe most of these web 2.0 companies will give up on becoming a major consumer brand and start thinking about the technology provider route.

If we look at the path of a typical web 2.0 startup at this point, I believe they will follow the web 1.0 model:

– Attempt to become a consumer brand
– Attempt to partner with established consumer brands as a content or technology provider
– Become a pure technology service play
– Become an auction site for Aeron Chairs

These little Web 2.0 companies are also quickly moving to take in large amounts of VC money without any clear business models other than acquisition – sounds to me like 2000 all over again.

Jay

Explicit human endorsement of sites will definitely lead to better results than PageRank based on link popularity.

This has already been proven for years with subscription search services in the library world. For example, LexisNexis, Ovid, NLM PubMed all include citation content that has been “tagged” with terms by content experts. This leads to very high quality, precise search.

In the end though, Google can just add this as an additional layer to their search algorithms.

Dragos

one observation — while they may (or may not) take on Google they seem to earn some bucks out via Google’s Adsense. :)

Comments are closed.