The role of humans in Google search

Randy Stross wrote an interesting article for the New York Times about search with a human touch, and I wanted to talk about the role of people in Google search.

On this post, you get not one but *two* disclaimers. It’s all part of my read-one-disclaimer, get-a-free-disclaimer program! 🙂 My disclaimers are:
– This particular post is entirely my own opinion.
– I’m really, really low on sleep. I’m up at Foo Camp 2007 this weekend. This is my first time at Foo Camp, so I stayed up until ~4 a.m. last night talking to people and discovering the crack-like addiction that is the Werewolf game. Okay, let’s begin with a question.

What is the future of search?
I see some obvious answers. For example, Google will continue to work very hard on international search so that we do just as well on a query in Japanese, German, Arabic, or Norwegian as we do in English. But what about longer-term? Will the future of search be

– personalization?
– a completely new user interface?
– semantic understanding of queries or documents?
– social search (which I’ll define as improving search by unlocking the power of people)?
– universal search, which brings in documents from non-html sources (images, videos, patents, etc.)?
– a combination of all of the above, or something entirely different?

Suffice it to say that we spend a lot of time thinking about the future of search at Google, and of course other people think about it too. Let’s take one area, social search, and delve deeper into the subject.

Social Search: the power of people

If you ask an average techie about Google, you’ll hear that we use lots of computers and algorithms. Indeed, the title of the New York Times article is “The Human Touch That May Loosen Google’s Grip.” But (in my opinion), it would be a mistake to think “Google is nothing but cold algorithms and computers; there’s no room for humans at all.” I’ll give you a few examples of the role of people over the years at Google:

– PageRank is fundamentally about the hyperlinks that people on the web create. All those people creating links help Google formulate an opinion of how important a page is.
– Google News looks at a wide variety of news sources; the decisions of human editors at thousands of news sites help Google estimate whether a particular story is significant.
– Google introduced voting buttons on the toolbar back in 2001. They look like happy/frowny faces and they let regular people send thumbs-up or thumbs-down votes to Google.
– Google has allowed users to remove results that they don’t like from Google.
– For more than five years, we’ve allowed users to report spam to Google. We’ve said for years that we reserve the right to take manual action on spam (e.g. if someone types in their name and gets off-topic porn as a result).

And of course, it’s not as if Google’s search engineers drive into the Googleplex in the morning and then spend the whole day sitting around doing nothing while the computers do all the work. 🙂 Instead, Google researchers and engineers spend our days looking for deeper insights that will let us create the next generation of search. I believe Google’s approach to search has always been pragmatic: if an approach will improve the quality of our search, we’re open to it.

“But Matt,” I hear you say, “aren’t you just saying this now because of the recent coverage of human-powered search companies such as Sproose, Mahalo, iRazoo, Bessed, etc.?” Actually, no. I think I’ve been saying similar things for a long time. I did an interview with John Battelle last year, for example. Read the full interview for my (very long) thoughts on the role of people in search, but here’s some of what I said:

I think that Google should be open to almost any signal that improves search quality. Let’s hop up to the 50,000 foot view. When savvy people think about Google, they think about algorithms, and algorithms are an important part of Google. But algorithms aren’t magic; they don’t leap fully-formed from computers like Athena bursting from the head of Zeus. Algorithms are written by people. People have to decide the starting points and inputs to algorithms. And quite often, those inputs are based on human contributions in some way. ….

So I think too many people get hung up on “Google having algorithms.” They miss the larger picture, which (to me) is to pursue approaches that are scalable and robust, even if that implies a human side. There’s nothing inherently wrong with using contributions from people–you just have to bear in mind the limitations of that data.

I believe that Google has thought about how to unlock the power of people in various ways since PageRank was invented. I’m allowed to make that claim, because more than five years ago I cared enough about leveraging social feedback that I helped write some of the Windows code for the voting buttons in the Google Toolbar. 🙂

Update, 6/26/2007: While this post is my personal opinion, I’ve noticed other Googlers confirming that Google is open to using human feedback to improve search quality. At the recent European Press Day, a journalist from Guardian Unlimited asked Marissa Mayer about this topic:

[Marissa] said that as the internet has grown, so has the need for search, At first, sites like Yahoo were listing the web by hand in the form of directories. Isn’t there now a place for human intervention again, I asked, now that the web is so full of information? I’m referring to Mahalo.com, the human-powered search engine we covered last week.

I was expecting her to say no, but she didn’t.

“When the web is as large and polluted as it is now, ultimately to need to have more sophisticated ways of searching it,” she said.

“Up to today we have relied on automation, but I believe the future will be a blend of both, combing the scale of automation and human intelligence.”

So that’s one datapoint. The other datapoint comes from Jason Calacanis, who wrote up a session at Foo Camp that Larry Page attended:

[ Larry Page just walked into the group of 12. ]

….

Larry says search is finding content… and that Wikipedia found a better way to organize information. he seems to like the model of using humans and process and machines.

So that’s another indication that Google is open to scalable and robust ways of utilizing the power of people.

106 Responses to The role of humans in Google search (Leave a comment)

  1. Matt

    The role of people in Google search.

    IMO, specially in the recent two years, Google has intensify its interaction with publishers/webmasters.

    That has been done through the communication of Matt Cutts, Adam Lasnik and Vanessa Fox with publishers/webmasters. We have seen Google asking publishers/webmasters for feedback in different occations regarding alog updates, sitemaps, paid links etc..

    All above reflects the role of people in GOOG search!

    Hopefully you enjoy Foo Camp 2007 with less night hours 🙂

  2. Matt,

    I read how you formulate the Page Rank of a page. But I really have some doubt about this blog. http://amitava.blogspot.pk/ (PR 3)

    I think not many people on the web has linked to this page. I asked many people about this and different opinion are there. Some says that this is ghost PR. Other has of the opinion that it is due to the domain blogspot.

    What is your thought. I would love to get the answer from horses mouth.

    One more question…[:)]

    What is this Foo Camp?

  3. Speaking of People Power….

    How much affect does the NOW POPULAR, Social Bookmarking sites play in a site’s popularity in Google’s algos?

    How much of a factor does CLICK POPULARITY play in the rankings of the organic SERPs?

    Are the SERPs EVER manually refined for competative keywords?

    When deciding to take action manually against Sites using Gray Hat or Black Hat Linking techniques how does one decide what length of time the penalties should exist? (one famous SEO site has returned to page one for SEARCH ENGINE OPTIMIZATION after a high profile, controversial banning a few years ago?)

    Are they Engineers and Developers who feel enough empathy and have enough courage to speak up for those small or new Web site owners
    who may feel that there is an imbalance in Google’s algos?

    How much impact does the replies on this blog and posts on SEO forums have on the Engineers when constructing new algos?

    How much impact has the uniquely adversarial, and sometimes controversially unpopular posts by SEARCHENGINESWEB had on recent adjustments made??????

  4. If Google values human signals, why has it not updated the Google Directory form the Open Directory Project / DMOZ in over a year.

    For all its faults, DMOZ is a huge resource of human signals about websites; and Google used to use it more prominently.

  5. In the fight against irrelevant results and search engine spam I do see a big role for the human touch. And things like Universal search also tend to make people more satisfied with their search results.
    But I also see many more ways to game the algorithm and the real need for an expert spam prevention team.
    So Matt, doesn’t this make your work more challenging? Please stay on as the main Google spamcop!!! You’re the only one that can tackle the problems of new social search gamers.

  6. Hi!

    my way to understand a little bit semantic understanding of queries or documents:

    We are swimming in water.
    The navy is shipping on water.
    The aquanaut dives under water.
    The albatross flies over the waters of large oceans.
    The house stands near the water.
    He dilutes the wine with water.

    The best vectors are Prepositions, because the other possibility, the verb is not so practicable

    (there are too much of them to make a system with).

    Both stand between noun and points out the content (wine with water – is the central information of

    the last sentence above).

    Prepositions describes a local, temporally, modally or causally propositions between two facts.

    I think a normal search interface has to be upgraded with an option. Than you can search like before,

    but if you want you can apply the new ways.

    Make out of that prepositions operands (buttons).
    That would make search working more variable
    (it shall look and akt like a calculator).

    Greetings Karl

  7. > I’ll give you a few examples of the role of
    > people over the years at Google …

    Additionally, a Google engineer once told me that Google also checks how often something is clicked on in results, to adjust the ranking. Considering there’s 1,000s of engineers at Google and not everyone works at web search, I’m not saying that because this person said it it makes it true… though I’d be interested to hear your take on this.

    > But algorithms aren’t magic; they don’t leap fully-formed from
    > computers like Athena bursting from the head of Zeus.

    You know that, we know that, but Google sometimes emphasizes what might be interpreted as the opposite stance — maybe that’s why Google in mainstream press isn’t always associated with the “human angle”:
    http://blog.outer-court.com/archive/2006-09-27-n79.html

  8. It’s good to hear that there’s a strong dose of human input in the Google process. After all the No Free Lunch Theorem would suggest that algorithms alone will always on average be somewhat similar in their effectiveness.

  9. I’m actually sort of on Google’s side here, but just to comment, part of what’s meant by “human touch” is manual override of algorithmic results. That is, if a specific term comes up with a bunch of not-quite-spam, just typing in ten “good” results. Now, I *KNOW* this is completely infeasible for the general problem. But what several start-ups are basing themselves on, is that a business can be made by doing tweaking on the Short Head of *profitable* search terms. That is, cream-skimming of the most profitable items.

  10. Hi

    THX 2 Barry 2 post that theorie here
    it sounds interesting that new NFL
    but reading and understanding are two different things 4 me there:

    [url]http://en.wikipedia.org/wiki/No-free-lunch_theorem[/url]

    but in another thread here I wrote:
    a genius finds, where a normal only will get the expected.

    and an average result (as you defined it above)

    suggest that algorithms alone will always on average be somewhat similar in their effectiveness

    seems to me, is a another (good and mathematical) example with the same conclusion as me.

    Greetings Karl

  11. One of the sysops who runs the MediaWiki:Spam Blacklist, a public resource, tells me that whenever he adds sites to the list, he notices that these sites drop out of Google pretty quickly, sometimes within hours. He also tells me that his list is very accurate. It’s generated by scripts that watch for spamming across all participating wikis. Volunteers trace interconnected spam sites, and check for false positives and Joe Jobs. Could this be another form of human input helping Google clean up search results?

  12. Ive been a huge fan of the way google does things for many years and its a great relief to know that there is still a human element to the search results. Webmasters the world over have been trying to figure out the google algorithms for years, and it is refreshing to know that is not just all mathematics. It has been very interesting to read some of the human roles at google and i would especially like to know more about how google detects and differentiates between spam and legitimate sites, Is this done by humans or a script because surely a script could make mistakes that could plunge a legitimate site into the abyss that is the google sandbox.. Also, What about duplicate content? There has been alot of speculation in the seo world about how google handles duplicate content so surely the “human element” could best determine what is duplicate content and what is not..

    Anyway.. Great Blog and keep up the excellent writing and insights into googlesway of doing things 😉

  13. Hi!

    Thereafter also operands like definition (of words or dimensions) or for instruction manuals as example for a sort of special documents are imaginable, as
    your farmteam is testing on [url]http://www.searchmash.com[/url] (with buttons for Images, Blogs Video and Wiki). There are the first steps; but why can I only find 2 german Blog-comments of mine? I have no answer, because my Links here are all in the Webmaster Console listed.

    Greetings Karl

  14. I have a hard time believing that Google will ever listen to people. For example I have reported a site that uses black SEO and Google still lists it on page one. The site buys all of its links, so using Matts post I reported it, nothing. That being said human touch is important and I think when you see a site with 3,000 inbound links and no outbound links an algo will not see that as a problem, but any human would say, hmm this looks odd, where is their hidden link page.

  15. I worked once for establishing ethical certification standards for affiliate programs vendors. At that time we suggested a “Certificated Badge”.

    Maybe its time for Google to issue “Google SEO Qualified Badge” for SEO firms and consultants who operate within Google Quality Guidelines. Also that Google propviding the ability to report SEO firms & consultants for spam.

  16. Hey Matt,
    I just thought I should let you know my Blog is for Sale! The details are on the homepage I entered.

  17. How much impact has the uniquely adversarial, and sometimes controversially unpopular posts by SEARCHENGINESWEB had on recent adjustments made??????

    Uniquely adversarial? You mean third-person egomaniacal ranting with obvious vested interest is uniquely adversarial?

    Multi-Worded Adam suggests that Google is broken, and that it needs to be fixed so that Multi-Worded Adam ranks #1 for every keyword and phrase he wants.

    Oh wait…MULTI-WORDED ADAM wants all that stuff, because MULTI-WORDED ADAM contributes so much.

    (10:1 odds says the point gets missed here.)

    Seriously, the biggest concern I can see with “the power of people” is whether or not the power is collected and used in aggregate or whether it’s individualized. If the latter is true, then that’s okay, but collecting data and using it in aggregate as a search engine factor (e.g. PageRank data) subjects that dataset to potential abuse.

    So…I guess what I’m wondering is how much information is individualized, and how much of it is aggregated?

  18. Foo Camp – did you get there via Larry P’s chopper? Seems funny to arrive like that and then sleep in a tent. But cool too.

    No need to defend Google against foolish ideas of human superiorities. Computers trump humans in many respects and *certainly* in 90%+ of internet search functions. Hopefully in 25 years or so the computers will be able to take over completely and we’ll see – for the first time in history – a world where everybody will be fed, educated, and healthy.

  19. Matt,

    Beyond the localized searches, and semantic relevancy algorithmic improvements, the human touch does nothing but corruption. When the pharoah says I am God and is actually given the power of God by Zeus giving him the presence of lightening rods to punish individuals and throw them to the Google hell hole rather than allowing the repenting ignorant sinners without preset automated penalty removals for first time offenses, while 90% of the other sinners are walking the google streets untouched

    Or allow non-Google public to write critical reviews of a competition to take the competition down in a search result,

    Human Manual intervention does nothing but destroy the entire idea of justice and freedom of internet as we are all sinners at the end of the day.

    Thanks for listening.

  20. Nice post, Matt. I sent you a happy-face vote from my Google toolbar. Did you get it? 🙂

  21. Google goes to the extreme of relying on algorithms and sad faces are not going to fix this!

    Sorry Matt.

    🙁

  22. Harith, that’s absolutely right; the human factor also applies in the communication we do online and at conferences; that feedback goes back to engineers who can act on it. Sometimes it’s small bug fixes (“Why doesn’t the root page of my site show up in the UK search?”) and sometimes it’s larger issues, but that feedback loop definitely helps Google (and thus searchers and site owners as well).

    igor, I deleted your link drop; please don’t do that. My comment policy at http://www.mattcutts.com/blog/guidelines-for-comments-march-24-2006/ gives more details (e.g. I remove signature links, or don’t approve those comments).

    Peter van der Graaf, chatting with friends about the role of humans in gaming and improving search is one of my favorite things to do. 🙂

    M.W.A., be nice. S.E.W. can be.. well, let’s face it — irritating sometimes. But good feedback can come from annoying comments too. At least he’s cut back on the bolding recently, yah? 🙂

    Philipp, whenever people ask me about a particular signal like that, my typical response is not to confirm or to deny it. I have stated in the past that data like that would be noisy and potentially spammy, so you’d have to be very careful about using that sort of signal or you might actually decrease relevance.

    Seth F., without going into that particular issue, I have a meta-question in response: if using N people in that way was successful, what advantage does a different company bring compared to Google? Google could use N+1 people, for example. The non-Google company would have to argue that they use the N people much more effectively than Google would be able to. But Google spends a lot of time thinking about how to pursue scalable solutions. So human “cream-skimming,” as you put it, could be an interesting niche but I don’t think that cream-skimming alone poses a mortal competitive threat to Google; there’s just so many searches in the long tail. Should we ever get a chance to chat in person, I’d enjoy talking about this more.

    Jonathan, the answer would be the same that I gave Philipp; I don’t think we would confirm or deny a particular quality signal. Out of curiosity, which blacklist was your friend talking about? Email? Open Proxy? Something else?

    scott, it could very well be that in the course of investigating the report, we took action on it that’s not visible or not visible yet.

    Matt Jones, I have my hands full just with running my own blog..

    Everfluxx, I did. 😉

    Joe Duck, I arrived at Foo Camp via the regular way; I drove. In fact, I didn’t head out from the South Bay until after 3 p.m., so I got to drive through an extra hour of traffic on the way up. Oh well — that’s how we mere mortals travel, eh? 🙂

  23. Philipp, I just got around to circling back to the url you mentioned at: http://blog.outer-court.com/archive/2006-09-27-n79.html
    where you discussed these issues last year. There’s a perfect example from that article. You discussed http://www.google.com/explanation.html and quoted the language that was on that page at that time:

    A site’s ranking in Google’s search results is automatically determined by computer algorithms using thousands of factors to calculate a page’s relevance to a given query.

    If you look at the wording now, it says “A site’s ranking in Google’s search results relies heavily on computer algorithms using thousands of factors to calculate a page’s relevance to a given query.” That language isn’t a change of our policy; it just clarifies our existing policies and lessens the odds of misinterpretation. The gist is still the same (we want to rely primarily on algorithms at Google), but the updated wording clarifies that we’re not precluded from utilizing human feedback to improve search quality, or e.g. taking manual action on spam, which is something that we’ve talked about for years and years.

    By the way, if anyone notices pages in our HTML documentation where we claim that something is generated “completely automatically,” please let me know so that we can clarify that language; I think we’ve updated most/all of the wording like that, but it’s always possible that we missed a page or so by accident.

  24. I always seem to post seconds before you respond Matt, I am wondering when Google is going to do something about people using multiple domains to completely rule a niche. Can you determine greed algorithmically? No? Then I am all for human intervention then but so far, I see no signs of it, why?

    The same greedy people are having their way with Google year after year and all it takes is multiple domains. Now with universal search you give those good at gaming gray areas even more opportunity. They will not just have the first 3-4 positions on organic search, they will have the entire first page!

    Are you guys aware of these things, it is very similar to how viruses mutate but never die.

    I was taking a break from blogging but this stuff is troubling…

  25. Forced myself to read that New York Times article and just wanted to say that human edited search engines like Mahalo suck. The article almost seemed anti-google as if the writer was being paid off by Mahalo. Expect many more of these types of articles in the future, everyone hates a winner.

    Google does need to “edit” until the point that their algorithms can do a better job, YES but as Matt stated, voting signals are sometimes extremely noisy. My bet and good wishes are with Google as always, I expect more from them.

    (sorry for talking too much)

    *out*

  26. Aaron Pratt, I definitely don’t think that Randy Stross (the author of the article) is anti-Google at all. He’s quite a good writer; I especially liked his book eBoys about Benchmark, a venture-capital firm. The one false note in Stross’ article for me was confounding spam with search engine optimization (SEO), since of course there are white-hat SEO techniques that aren’t spammy at all. But for all I know, that could have been an editor’s doing.

    I think it’s fair to ask if human-powered search will threaten Google, and I’m glad that Stross asked for Google’s take. The article seemed like a good opportunity for me to expound on the topic of “humans + Google” in more detail. Folks at SMX Seattle heard me say similar things, but it’s good to write about it on the blog for everyone else. 🙂

    By the way, my comments about “humans + Google” at SMX Seattle led to this article by Nancy Gohring at Infoworld:
    http://www.infoworld.com/article/07/06/04/Humans-get-bigger-role-in-search_1.html
    Ms. Gohring must have been in the audience because I don’t believe I talked to her during the conference. That story is really impressive for a journalist without a background in search engine optimization. So this blog post also serves to give more context to that Infoworld article as well.

  27. Hi Matt,

    Sometimes I feel Wiki does much better than Human touch Search Engines. What is your thought??

    Thanks,
    Deepthi

  28. To my observation, Google is doing a much better job than Baidu in China. Baidu is mercantile; for popular search terms, on the first results page, it mixes free listings and ad clients, sometimes all listings belong to ad clients. Baidu results are known as being heavily interfered by human, but not for better results. We once notarized our indexed pages after finding that most of the pages were gone; this just happened after we discontinue a series of ad services with Baidu. For the following two weeks, we were talking to Baidu, and they finally recovered the data, explaining out it was the search engine’s mistake.

    Google is an excellent company. It’s fair and listens to people. (Baidu listens to government.)

  29. Matt said:

    Jonathan, the answer would be the same that I gave Philipp; I don’t think we would confirm or deny a particular quality signal. Out of curiosity, which blacklist was your friend talking about? Email? Open Proxy? Something else

    This spam blacklist. Note the GFDL license. Your list of spam sites should be a superset of this list because these are just the ones that have link spammed Media Wiki installations.

    (Adding the link)

  30. I guess that Google’s human raters shouldn’t be left out in this ‘role of humans in search’ discussion. They also do a hell of a good job to improve the quality of the search results.

  31. Well yes sometimes wiki works quite well than search engines, but yet search engine is more important. It’s my thought let’s see what matt have to say on this.

  32. Hi Matt
    I think you are totally right, google help user too much with his (her?) many tools, I just tell about Google Analytic – it’s a wonderful experience to me to access of all traffic of my websites.

    Thanks
    Deb

  33. As for the future of search i personally think that in a long term less traffic would come from search engines, which would be replaced by intelligent agents based on web services, however, in order that to happen we would need a standardized representation for data. At the moment search engines have to deal with a number of errors in more or less broken html markup.

    In the short term though you don’t even need search engines to get traffic to your site. I’ve recendly done an experiment and managed to get 38k visitors in 7 days to a brand new blog. Here is more information about the experiment for those who are interested:
    http://www.ambergreeninternetmarketing.com/blog/case-study-boost-your-traffic-with-social-networks/2007/06/25/

    It just highlights the strength of social media and might be a wake up call for most of the SEO companies.

  34. I always enjoy reading other people comments and I am sure that you get a chuckle out of many of them as well. People speculating on this, and wondering on that, but to keep it fairly simple Google has the master plan. By tracking as much user data as they do, they have a much easier time with determining what is relevant, and what is not. Google keeps it simple, give the people what they want based on the user data they have. And since they have enough user data compiled to probably cover the entire galaxy in mountains of paper, it makes it much easier to give people what they are searching for. That’s why I both love and hate Google…it gives me what I want…but for the price of tracking my every move on the web. I sometimes wonder if any of this is real, or has Google already assimilated everyone into the matrix? Sometimes, if I focus REALLY hard, I can see the truth, and the binary code that is life becomes clear…then the Google agents come…

  35. Hello Matt Cutts-

    As a role as an average Internet surfer and webmaster, I’d like to inform you with a way webmasters are no plagerizing, well more-over paraphrasing WITHOUT citation, using one of Google’s powerful services.

    I have a script that allows users to do this, and I’d like to further explain it to you and show you the script.

    If you can, by any chance, email me, please do. I understand that you’re a very busy person but I don’t want to announce this “exploit” to the world..

    I’ll explain it though so you can believe me. One of Google’s services is allowing webmasters to take content from another website and it automatically rephrases the content, thus giving the webmaster unique paraphrased content.

    Please email me immediately. I think it’s immoral…a lot of webmasters are getting “quality” content just by stealing it from others this way!

  36. Matt,

    The problem I have with PageRank is that it is not the USERS of the website who are creating the links, it is a very small group of entirely untypical people called WEBMASTERS.

    This, in my opinion, is the fundamental problem with Google and one that is insurmountable whilst maintaining backlinks as the prime determinate of ranking.

    It is this reliance on backlinks (and Google’s dominance in search) that has created the paid link, PR network internet that we have today. A vast amount of the internet is simply there to spam Google.

    Pure USER determined search is the future, whatever form that might end up taking.

    For example, in my field the SERPS for the main keyword contains some sites that straight out rip-off customers (they send out no products) and many that sell substandard products.

    These sites have bought backlinks to get where they are. Even the recent introduction of some kind of user tracking into the SERPS does not fix this problem. It in fact makes it worse since these rip-off sites look amazing so users click away merrily, before getting ripped off.

    Sites like mine don’t rank very well because I don’t buy links and don’t place misleading junk on my site. It is simple and to the point.

    In product markets like mine the ONLY way forward is directly USER influenced search.

    I’d been in search since 1998 and would LOVE to move to Dublin. Ha!

  37. > If you look at the wording now, it says “A site’s ranking
    > in Google’s search results relies heavily on computer
    > algorithms using thousands of factors to calculate a
    > page’s relevance to a given query.” That language
    > isn’t a change of our policy; it just clarifies our existing
    > policies and lessens the odds of misinterpretation.

    Ah, interesting. Good to see this!

  38. Hey Matt,

    This is my first time here…I’ve heard much about you in the past from articles, or other IM’ers, but not until today did I finally google “Matt Cutts” and found what “everybody” was talking about. Aaaahhh, you are obviously a busy man.

    Now back on topic.

    I tend to think that Google has more human control with their algorithms than what they will ever tell us. Yes, I believe that obviously the core of Googles search results are generated by computers – but I’m sure there are a few tiny loopholes in the system that can be manipulated in one way or another by a human.

    Yes, I’ll admit – this is only my opinion. 🙂

    Cheers,
    Ken

  39. Richard Hearne

    Well at least when Igor is over here harassing Matt he’s not over on the Webmaster Group abusing the other members…

    Sorry Matt!

  40. Seems to me the roll of Humans in search is that the antispam team comes in and selectively punishes people for doing exactly what everyone else ranking for a term is doing.

    Example. Search Las Vegas Real Estate. MSLasVegas Removed due to affiliation with Real Estate Webmasters. Everyone one else left alone even though the majority of them are also doing recip linking from state pages.

    Its to bad the engineers at google aren’t smart enough to fix these loopholes with the algo and we have to wait for Greg Boso to turn people in, then have the anti spam team human edit and get really good websites penalized for a industry wide practise.

  41. Come on Richard, you have to be joking? Igor is the epitome of how to win friends and influence people.

  42. Monica Valentinelli

    I’ve noticed, though, that Google StopWords are different for every language. In some latin languages, like Italian, it reads that some numbers are included. Now I know through my foreign language experiences there are lots of semantics involved, but I wonder how much search differs for other languages–especially Spanish?

    Not to mention the headache of dialects…

  43. Hat Tip to Richard Hearne for finding the silver lining, at least Matt can prune comments here…oops wrong post for hat tips.

  44. Doesn’t seem fair does it Raspy1? Be thankful you were not penalized in Yahoo. I hear their penalties last longer. Then again, their search results suck more and more every day so it doesn’t really matter.

    As almost as perfect as Google is they seemed to have dropped the ball on this ordeal.

  45. >>What is the future of search?

    What is the Future of Commercial Search in Google?

  46. Hi Matt,

    Great article! I’m a little surprised there has been no mention yet of of Google Custom Search. At TopTenGlobal.com, we see Google Custom Search as the foundation for the ultimate hybrid solution, utilizing the best abilites of humans and machines in a robust and scalable solution.

    From my latest blog post:
    “…Now would be an appropriate time for me to mention that Top Ten Global has developed our own hybrid approach to social search that we call “community-powered vertical search.”

    Our solution is robust and scalable. Instead of quixotically trying to hand-craft hundreds of thousands of search-result pages, we leverage the collective intelligence of humans to continually optimize and fine-tune the algorithm. We remove spam from the index and teach the algorithm what humans consider to be relevant. Our approach blends the best of humans and machines. When you search at a Top Ten Global brand web magazine you do not need to guess if we already have a page created for your query. That would be silly. Instead, you immediately get the full power of Google search coupled with the cumulative efforts of the editorial community.

    We’re looking for feedback. Please visit one of our web magazines and let us know what you think about community-powered vertical search (a custom integration of Google Custom Search). Thank you! …”

    Jonathan

  47. I think from the human-powered search engines being covered by NYT, http://NosyJoe.com looks most promising to me, while from the algorithm based ones http://hakia.com is my favorite 🙂
    cheers

  48. Richard Barber

    Matt,

    You probably won’t remember I mentioned to you the idea of being able to “block” and “boost” sites appearing in my personalised results, using a link under the results saying “block” or “boost” at SES London.

    This would be useful data as if lots of people block a particular site, then that would help Google. You were concerned how this could be processed – i figured through the use of the desktop search and some background processing on my own machine.

    One of the things that you have said in the past is that there would not be crossover of information from adwords to organic and vice versa. If you know a user likes a certain result, the chances are they like it irrespective of whether it is in the organic or paid results. Therefore wouldn’t it make sense to promote results that do well in paid in the natural and vice versa?

    Thanks for all your great work

    Richard

  49. Nice quote Mert: “When the pharoah says I am God and is actually given the power of God by Zeus giving him the presence of lightening rods to punish individuals and throw them to the Google hell hole rather than allowing the repenting ignorant sinners without preset automated penalty removals for first time offenses, while 90% of the other sinners are walking the google streets untouched”

    One would think that if 90% of the other sinners are untouched than an algorithm change might be on its way.

  50. Matt,

    When I spoke to Randy Stross for the NYT article, I thanked him for his desire to acknowledge that, despite our small staff and lack of VC money, Bessed had actually originated the human-powered concept almost a year before Mahalo and months before the Jimmy Wales wiki-type search engine was announced. I then discussed with him one important difference I saw in what we are attempting to do versus these other sites, and how that jibes with Google.

    Mahalo has made it clear they are looking to cover the big money terms that are searched most often, and try to rid those search results of the spam seen on Google or other search engines. Problem is, those big search terms usually aren’t very spammy—due to the fact that Google uses the strength of humans via its linking algorithm to do a pretty good job of ranking sites for common searches.

    From our perspective, where Google and others are sometimes weak is in the “long tail” where the linking between sites is either rare or is totally doctored to get certain sites to the top of the heap. In these cases, it is easier for the spam to rise. For example, if you do searches for specific products, you’ll often get 7 different comparison engines in the top 10 results or get a spammy page that uses that keyword to lure you into their unrelated site. This is where an actual human looking at the results can make a big difference.

    Same thing with business-to-business sites where link factors may not come into play — companies don’t often link to their competitors, and business searchers are less likely to be linking to their ball bearings company on their blogs. Business-to-business is important, but it doesn’t often lend itself to enthusiastic linking, which makes it more difficult for Google’s algo to pick which site is best, also making it easier for the sites more tuned to SEO to finagle their way to the top.

    This is where we at Bessed hope to make a difference. We have a looong way to go and obviously the long tail is pretty long, but with a goal of eventually fixing what’s broken versus “it ain’t broke, let’s fix it,” I believe our goals are more in line with what searchers actually need (versus what we want or VCs want).

  51. I’m also interested in hearing how much a role social bookmarking sites play in a the role of SERP’s

  52. Hey, Wiki is known by tech people but, common people will know about search engines like google and Yahoo

    Great to hear that, google is going to utilize the power of people….

    Lucky people will soon join google 🙂

  53. Hi!

    Now you can have groups-buttons for local, temporally, modally or causally connections and in them you can fit the specials of this preposition (in,on,at) into ( like in graphic-programms; circle-oval, quadrat-oblong). Colors or pictograms may helpful
    by identifying it quicker. Also helpful Buttons with the operands which are now in use “like“ and + or –(without) and for “ would be ( better I think. And this brings me to a major point today: teach the user searching better! I see ever Serps like that:

    [url]http://www.google.de/search?q=sie+mag+musik+nur+wenn+sie+laut+ist+download+mp3+free&hl=de&client=firefox-a&rls=org.mozilla:de:official&start=10&sa=N[/url]

    many words and only + or spaces bring the user 71 K answers!

    But sometimes it is better to say: Nothing,
    if there are only results with no sense in it.
    But this would assume that Google has some sort of semantic KI (along the algorithm) to anticipate No-No’s, or Google let Users to the smarter search.

    [url]http://www.google.de/search?hl=de&ie=ISO-8859-1&client=firefox-a&rls=org.mozilla%3Ade%3Aofficial&q=%22sie+mag+musik+nur+wenn+sie+laut+ist%22+%22download+mp3+free%22&btnG=Suche&meta=[/url]

    will tell you the truth! And if user are too lazy, you better help them to the grouping of words
    in their questions easily. One button clich and out of
    sie+mag+musik+nur+wenn+sie+laut+ist
    wird
    “sie mag musik nur wenn sie laut ist”
    this would 3 fewer keys fort he same result and a big change in their search-achievement.

    Greetings Karl

  54. > If Google values human signals, why has it not updated the Google Directory form the Open Directory Project / DMOZ in over a year.

    This does not appear to be true : there is evidence that Google is still downloading data from the ODP and using it to improve search results.

    ODP data has been used by Google for several years, and was probably the first attempt to include human touch into search results.

    Unfortunately, the ODP has experienced serious technical problems in the past year, mainly due to the fact that it’s almost neglected by its owner, AOL/Time Warner.

    http://www.skrenta.com/2006/12/dmoz_had_9_lives_used_up_yet.html

  55. Hi Matt,

    I think you did not at all take humans in count in the last Google Algorithms.

    I have an authority site on African Art that exist since 1998. And I must tell that the last algorythm change you did at Google one week ago had dramatic changes for me, and that I saw my results drop with 75 % less visitors in that week. This is a pitty since my site had realy usefull content for my more than 6000 unique readers each week.

    Will definitly remove all not related pages on my site (+/- 10% ), to see if it reappear in the results. But the mixing from different sources (like video, news, etc..) doesn’t make me confident that for the future of webmasters trying to serve good content through Google.

    I have more than thousand pages, and that provided me an income of +/- 40$ a day. But at 13$ a day, and only traffic coming from the images on my site, If Google don’t put me in the first results I don’t see why I would work futher on my site, and set up a reliable source to get track of the African Antiques events

  56. Hi!

    And now you ask me why I wrote it here?

    I have no chance ta make it alone ever
    and it seems so, that no one on the world is interested to work with me as a Team on a project.

    But I think better seeking is a basic target for mankind today
    and thinking together about it, is better now without me in,
    than never or later and with me in.

    And to Google: maybe the first step (or the last to avoid the button-tide) You can go easily, is to show the user the figures of oher potential combinations of their
    (q=y+x+z+mp3+4+free without all canon-signs)-question
    as Clusty do 4 example.

    But I fear most normal Google-Users like it,
    when the engine shows up 1.46 million answers or more to a question.

    Greetings Karl

  57. I remember that Google never was rejecting the fact that they are using lots of “human power” to refine their results. For example, I remember a famous sample of letter sent to student who would like to work with Google and need the capabilities to refine sites in niches.

    That is why it looks pretty funny that someone thinks Google is stuck in search algo only.

  58. Hi Matt:

    While reading your post about how Google is open to human feedback – I decided to offer a little. I’ve noticed while working with Google Blog Search on some Squidoo’s lateley – the results don’t seem to be nearly as relevant (on topic) than they used to be. In fact, some of them are completely off topic.

  59. OK guys let’s back to the topic, we are talking about human touch and if it can provide more relevant and less spammy results than the ones the Google algorithm is providing.

    I have read the New York Times’s article and I have a question, I don’t know if anyone here can answer to me.

    What is Mahalo doing, that Yahoo! wasn’t already doing with his directory in 1998 ?

  60. This is fascinating – I’ve been wondering about this for a while, and I didn’t even see the NYT article.

  61. Dave (original)

    Personally, I doubt most searchers want total strangers deciding what is the best match on their behalf. Humans are simply too subjective, while algos are objective. Humans are bias, where algos aren’t. This is likely a contributing factor to the decline in users using directories.

    To top it off, humans cannot manually review all pages on the WWW. Not without the population of China or India agreeing to donate time at least 🙂

  62. SEOhNo! – I meant to tell you, your cocoon has a leak and some of the juice has been dripping down on my cocoon! So next time you unplug, please use the duct tape I left by your pod, and tape it up!
    – A friendly neigbor in a pod in section GS-3285

    ===============================
    Matt and Group,

    I am certain after watching some interesting things over time that Google 100% for sure has a human hand, (or at least finger tips) involved. At a minimum the removal of penalties are sometimes done by hand.

    I am a DMOZ editor, and really love the idea of humans being very involved with the ranking of sites. In fact, in my humble opinion, I would like to be able to flavor google to the TYPE of sites I like. When my wife searches she is looking for places that sell with lots of photos, while I am usually looking for sites with lots of text and data.

    I also have a little bit of a punk rock, edginess that I like to see on sites, whereas others much prefer sites with good spelling and animation. I think human touch should extend to your being able to tweak google, or google tweaks itself, to your preferences.

    Maybe teams of editors, with similar hair cuts, cross referenced with favorite radio stations, should sit in their own building, with their own copy of the algorythim, tweaking it, and hand coding their favorites to do well in the serps. Literally google could compete with itself. Google-pop Google-goth Google-corp. Google-clean etc.

    As far as my main site goes, the absolute best content on it, which is possibly the best content in my entire industry will never rank well, because it only has about 5 to 10 words per page spread out over 80 pages of a cartoon book. On the other hand, some of my other pages that are really not that interesting, rank well because they match what google is looking for.

    Either way, I love goog, and thanks for all the visitors!

    dk

  63. Hi Matt!

    An algorithm is only as cold as his master. So the most important human touch is tweaking the algorithm. And I guess in the future we will see more interaction between the algorithm and user input, experience and trends.

    But most of what I read here is about relevant results and spammy websites. And adsense has a major role to play here. Google pays more the higher the CTR of the ads. And that is the building foundation for all the sites that hide their content and publish adsense all over the visible part of the website.

    A quality website, with great content, but with a normal ad placement (no quality website will want to loose new visitors to adsense), will win less money per click on adsense than a spammy website. And for any advertiser, a click on a quality website, with a normal ad placement, is 100x more valuable, than any click on a website that has 10 or 20% CTR. I wouldn’t want to advertise a website on a publisher’s site that has more than 10% CTR. Those clicks are just junk!

    So, if google is really serious about relevant results, it has to pay better adsense money to publishers that have quality websites, with great content. And pay less money or even no money to junk websites. And high CTR in adwords should mean better placement. But high CTR in adsense does not justify more money. What advertiser wants a 20% CTR on his ads, on a network like google? Only a stupid one.

    And human touch has a role here. Review more adsense sites. Review them all.

    I love google. Most of the times… 🙂

  64. Dave (original)

    TomFrog, Google has started the ball rolling on made for AdSence sites. The deadline is getting ever closer 🙂

  65. Matt

    Hasn’t there been some human input/feedback for a while now?

    I’m referring to eval.google.com – which if I recall, a few years ago, GoogleGuy confirmed it existed and was used for quality scoring purposes.

    I suppose what I’m saying is, I have long thought of Google as having the ability to use human input to improve search results.

    JB

  66. Dave

    But has adsense changed? Sorry, but I didn’t notice that. I still believe that adsense pays more dollars if your CTR is very high. And that means more revenue to google in the short term. But, as a direct result, thousands of low quality websites.

    Any website that places adsense on their most valuable placements will have a higher bounce rate and I don’t believe most of these websites are quality websites.

    Will user experience change this? Why should it? Why should users help google improve the search results? Maybe a new search engine will appear with more relevant search results. Then users will just move to the new search engine.

    I don’t mind helping google, by voting that this website is cool and this one is not so cool. Or by blocking a few websites. But if a new player appears with better search results, I will use that search engine. If I want a human edited listing of websites, I will search dmoz or the yahoo directory. If I want a social search, I will search digg.

    Google’s main value is the algorithm. This human touch it’s just not google. It’s dmoz or digg or something else.

    Roy

  67. Well it only proves that automated systems still haven’t proven to be reliable enough, for even google to use more human work. If anyone I think google would be the first to use totally non-human systems, only it will take while before they are advanced enough to do everything that is needed. For now the search will stay the same, and I actually prefer it that way- getting more results and choosing myself rather than super advanced google choosing in my place…

  68. I’d like to chime in on the difference between what Yahoo! was doing a dozen years ago versus what a “human-powered” search engine is atempting to do today, or at least what we are attempting to do at Bessed.

    Randy Stross, who wrote the NYT article, asked me “Is this just a directory?” My answer was no, and here’s why. In general, a directory is a hierarchical thing that is best at finding you general information on a topic, such as decent sites for buying shoes online. Ask a directory to tell you who sells Skechers shoes otr Nike shoes or Clarks shoes or a specific shoe from a specific brand and it will always come up empty, because directories are not built to handle very specific queries. This is why search engines won out in the first place—you could type exactly what you were looking for and find relevant sites (hopefully), especially for those “long tail” searches. Yahoo realized this and shuttled its directory to a dimly-lit back room, never to be featured again.

    So the concept of a human-powered search engine should be different than a human-edited directory in that its human editors are seeking out sites for the nitty-gritty detailed searches that most people actually perform, not the general static lists of overview sites that directories generally contain.

  69. Dave (original)

    Adam, I agree there are differences, but, IMO, they are outweighed by the similarities. Particularly the point that both are subjective and have human bias and not objective without bias.

  70. Dave,

    If you buy that the Google algo uses links to decide which sites come up first, you’re already buying into human bias. All searches have human bias because humans created the algo — the very fact that Google and the others change the algo every so often and thus different results show up on the same search engine demonstrates there is a bias. Humans program machines, and the bias that the humans decide to make strongest helps determine search engine results. There is no escape from human bias.

    Also, deciding which Web sites are most relevant for a search term or keyword phrase is inherently unobjective. If I type in “George W. Bush” I might want a history of his life from a news organization while if you type in “George W. Bush” you may want the White House site so you can write him a letter. Whether a machine puts the results together or a human does, neither of them are “objective” because there is no right or wrong to which results show up first. Your opinion on how helpful you find the results that are spit back is subjective–i.e., how helpful they were to you personally.

    Again, there is no escape from human bias when it comes to finding things via a search engine (or anywhere else actually).

  71. Dave (Original)

    Links are only part of a very complex algo which are objective by design. While there is no escape from ALL human bias, Google’s algo takes out a LOT of it. They have not held the number 1 spot for Years for nothing.

    Fully Human edited SE’s and Directories have had their day in the Sun IMO. Even with all others things aside, the WWW is just too large and ever-growing for humans to manually review.

    In addition, I would bet human powered SEs use algo driven SEs to find what they THINK is the best set of results for users. There’s a not so hidden message in there 🙂

  72. HI!

    The continuance of the Button-Tide can be

    Who Biografic Data only
    Where Geogrific Data only
    When Historic Data only
    How Manuals ect.
    Excerpts who say it first?

    The searcher is better,
    ever flexible he is (and human is the most flexible searcher for ever on this earth).

    Give the people the options
    to build their own engine out of Buttons-Tide-Modules
    and delivere the special Data for the search with it.

    Thinkink in hypothetical plausibilities are in Computers only Statistics!
    Searching with it in trained human brain abnormal?

    If you are good in that
    you can say me the content of this message I received after the first comment here (but I fear she/he is wrong):

    212.×07.x16.2×0 – – [27/Jun/2007:09:40:16 +0200] “GET /robots.txt HTTP/1.1” 404 619 http://www.light2art.de “-” “Mozilla/5.0 (Windows; U; Windows NT 5.2; en-US; rv:1.8.1.4) Gecko/20070515 Firefox/2.0.0.4” “-”

    212.×07.x16.2×0 – – [27/Jun/2007:09:40:16 +0200] “GET /robots.txt HTTP/1.1” 404 619 http://www.light2art.de “-” “Mozilla/5.0 (Windows; U; Windows NT 5.2; en-US; rv:1.8.1.4) Gecko/20070515 Firefox/2.0.0.4” “-”

    212.×07.x16.2×0 – – [27/Jun/2007:09:40:17 +0200] “GET /robots.txt HTTP/1.1” 404 619 http://www.light2art.de “-” “Mozilla/5.0 (Windows; U; Windows NT 5.2; en-US; rv:1.8.1.4) Gecko/20070515 Firefox/2.0.0.4” “-”

    212.×07.x16.2×0 – – [27/Jun/2007:10:09:19 +0200] “GET /robots.txt HTTP/1.1” 404 619 http://www.light2art.de “-” “Mozilla/5.0 (Windows; U; Windows NT 5.2; en-US; rv:1.8.1.4) Gecko/20070515 Firefox/2.0.0.4” “-”

    a little helper added and the IP X-ed by me

    Greetings Karl

  73. “I’d like to chime in on the difference between what Yahoo! was doing a dozen years ago versus what a “human-powered” search engine is atempting to do today”

    @Adam Jusko
    Thank you Adam for answering to my question about the diference between what a “human-powered” search engine is doing now and what the Yahoo! Directory is doing since 1998. I explained your interessant point of view to the brazilian SEO comunity thru my blog -> http://www.netindexing.com/blog/?2007/06/29/32-adam-jusko-diretor-do-bessedcom-me-responde

  74. hi Matt Cutts, i find google toolbar is always very useful for everyday usage
    your article on role of human on google search is very impressive and i find the information very useful for begiinners user like us.
    thank you

  75. Thanks for the NYT article, more human touch is great for search engines!

  76. I never will understand, how it could be possible for Google, to develop in such a rasant manner…
    For every other company this would never be possible, but I know that YOU know this 🙂
    Greetings
    Blog Guy

  77. First off, I just stumbled onto this blog–great stuff! I’ve thought for several years that the time it takes me to sort through search results to get to the gems has gone downhill since Google arrived. Sometimes I long for the days of Mosaic when i got a listing of all the new websites that were published that day. There was no weeding through 100 pages of search engine optimized crap links with zero content then–if it was a WWW page, it was something that someone truly thought was valuable or entertaining enought to post for others to use or enjoy. There are only two ways out of the garbage bin that is the internet today–either more human involvement in search engine results, or far more intelligent algorithms within the search engines that can separate the garbage posts from the ones with real content that someone has actually thought about before posting. Glad to see Google is opening up to using human feedback!

  78. Hi!

    and 2 the 15 SlurpConfirm404’s

    I think about that

    72.×0.2×5.9x – – [30/Jun/2007:01:43:34 +0200] “GET /SlurpConfirm404/lofi/INQUIRY.htm HTTP/1.0” 404 619 http://www.light2art.de “-” “Mozilla/5.0

    as a compliment.

    Without your endless resources I could only play my Stradivari in my garage,
    I announced it here 18 months ago, that I have such a Searchulator concept.
    But is seems so, that no one can believe me ever.

    Greetings Karl

    PS: As IQ 140+ living with the social sediment for decades I made this experience:

    Most addle-brained people don’t like to become pieces of advice.
    Because they fear, that they may act as a guardian for me.
    They refuse good tips for years and tend to make antipodal
    while saying (before the tip ends): it runs here without it for years,
    so it can run along!

    So I had to learn to be silent on my few friends in that case
    but 4 you I make a exception every time!

  79. >

    But yet the publicly available service
    http://directory.google.com

    is stale and out of date by at least a year.

    Google results: stale and out of date. Is that what Matt really wants for _any_ Google service?

  80. “- a combination of all of the above, or something entirely different?”

    Something Entirely Different!!!!!!!

    The WEB 2.0 and after that the WEB 3.0,.. 4.0,.. 5.0,.. etc. will determine what search will look like in the future.

    The human factor will not so much be determined by involvement of humans, but more by how human the algorithms become.

    I don’t know how others see this, but search algorithms basically try to copy human behaviour. Order the results the way humans would do. The better they get at that, the less need for humans.

    There will come a time where we will be talking to our computers… hmm no,. that we already do… 🙂 I mean there will be a time we will be interacting with our computers through voice. And our computers will almost behave like the computer on board of the Enterprice in Start Trek. And if you follow(ed) that show a bit, you’ll know there are no keyboards or monitors anymore in the future. I don’t think we will have to wait until the 24th century before we manage to build computers that we can interact with as if they were a person. That´s more like 5 to 10 years away from now.

    So Google and possibly a next generation search company, will come up with a WEB 4.0 (I think it we have to get through some WEB 3.0 before we reach this point.) will have to come up with an interface that simply does nothing until somebody says: “Computer, what movies are playing this evening?” Where the computer says: “What do you feel like this evening?” and you reply: “Science fiction” and the computer replies: “In the cinema in the mall you can see: startrek the 6th generation, it starts at 8pm.” And ends with: “Thank you for using Google, This answer was provided to you by your local coffeeshop that is located right next to the cinema you´re going tonight!”

  81. Hi,

    I am really sad to hear that Google founders like Wikipedia and think they do good work. I really think Wikipedia sucks. Please read the article below for some details. I think Google should delete all Wikipedia pages from the Google index and not send any traffic to them. Google is helping to promote a defective site (Wikipedia) and a terrible concept and implentation. It is a big mistake by Google to rank all these poor and many times incorrect Wikipedia pages so highly. My trust of Wikipedia pages is very low, and I always advise people to not use or trust Wikipedia at all.

    http://homepage.univie.ac.at/horst.prillinger/blog/archives/2004/06/000623.html

  82. Matt, I’m so glad to read this post because I had questions for long time that I couldn’t find any answer. Here my million dollar questions 🙂
    -What’s importance of ‘voting buttons’ for google?
    -Does it really mean anything to google when it comes to define a pagerank of a website? In other words, assuming I’m a very popular person in webmasters world and I opened a website. I asked all of my webmaster friends to click on smiley face on google toolbar just once. 1000 clicks on smiley face, would return as pagerank?

    Some other questions on google toolbar and history;
    As far as I tested, if you don’t enable ‘pagerank’ option but if you choose to use ‘History’ service, google toolbar is reporting each and every website you visit. Those pages are not limited with the ones that you searched on google but everything just like a browser history.
    -Are you using history section to determine pagerank?
    -Are you using the time a visitor spent on a website to determine pagerank? In other words, I searched for ‘Matt Cutts’ and did not click on your website but next result. I spent long time on that website. If that’s the tendancy for people who are searching for ‘Matt Cutts’, should Google put that website in higher position.? Or are you already using this approach?

    Apart from Google Toolbar, when a person does a search on google and clicks on the result that he likes. Before he goes to that page, the link that he is clicking is reported to Google:
    -Are you keeping track of ‘number of visitors’ of websites that are searched from google?
    -What would be the purpose of keeping track of those clicks in terms of Human roles in search engines?

    Oh my Goodness, I have many more questions. I wish I could attend to one of your Foo Camps 🙂

    Looking forward to your answers.
    Cheers
    Mehmet

  83. Just curious Matt why you left out my article in your summary since it is the No #1 article on the subject at this time.

    http://www.1stsearchenginerankings.com/2007/06/08/google-bounce-factor-research-data-is-in/

    Since these are all “your views” then I thought i’d give mine.

    Happy 4th!!

  84. Randy,

    I read your article, I’d like to say ‘very good job’. I had those questions for quite some time just as curiosity. It is nice to hear you made experiment on this. It’d be greater to hear some official/inofficial answer from Matt though… (still hoping & waiting :))

    Thanks

  85. Thanks for the NYT article, more human touch is great for search engines!

  86. Great to hear that, google is going to utilize the power of people….

    Lucky people will soon join google

  87. Hm – You forgot to mention mind reading & search… Any potential?

  88. Hi Matt,

    I’ve read some of your stuff and it’s really great!

    Anyway, I have a question for you. Currently I’m doing some (prelim) research in the area of traffic generated by SEO to web sites for computer performance analysis and capacity planning purposes.

    I was wondering if there were any estimation formulas for this type of stuff.

    What I have done, with a colleague of mine is to come up with some estimation formulas for this type of traffic.

    The formulas are posted in a blog we have titled “SEO/SEM Theory”, with the following URL:
    http://sem-theory.blogspot.com/2007/06/how-much-traffic-does-seo-generate-to.html

    and also posted on my website at:

    http://www.cs.usm.maine.edu

    Any comments, suggestions, pointers would be great.

    Thanks,

    PMF

  89. Don’t forget to check my research on search, it is going great guns, i wish I had more resources, i start where Google ends :). http://www.idealwebtools.com/blog/mygoogle-project/

    It includes grouping of similar sites. I certainly don’t want to read websites, i want the answers and the services. The human knowledge has to be distributive.

  90. I worked once for establishing ethical certification standards for affiliate programs vendors. At that time we suggested a “Certificated Badge”.

  91. Thanks for the NYT article, more human touch is great for search engines!

  92. Don’t overlook Tim Berners-Lee’s first directory, the WWW-Virutal Library — human edited, non-commercial, with servers from the European University to University of Kansas.

    Still serves a larger purpose — unlike Wikipedia, our links are not devalued — thus, even if WWW-VL sites on not on google’s first page — sites we link to are often.

    Here’s my volunteer non-commercial take on search engines,

    WWW-VL: W3 Search Engines
    http://vlib.iue.it/history/search/

  93. Don’t overlook Tim Berners-Lee’s first directory, the WWW-Virtual Library — human edited, non-commercial, with servers from the European University Institute to the University of Kansas.

    Still serves a larger purpose — unlike Wikipedia, our links are not devalued — thus, even if WWW-VL sites on not on google’s first page — sites we link to are often.

    Here’s my volunteer non-commercial take on search engines,

    WWW-VL: W3 Search Engines
    http://vlib.iue.it/history/search/

  94. Hi Matt this is a nice article from you, that talks about human experience. But, with centillions of websites online, I don’t think it is ever possible to ever figure out a method that with live humans sorting the list. It becomes necessary for a search engine to figure out methods those can sort the importance of a website based on user inputs those are automatically fetched.

  95. Google’s use of a combination of human power and algorithm power seems to be working better than Mahalo. Sorry to self promote, but it’s worth visiting Mahalo: POS Search Engine! if only to see the screen shot of a search to find pages describing a knitting procedure (knitting twice into same stitch”). Yesterday morning, Mahalo returned “Gay Marriage” and “Asteroids” as relevant. They still are, though due to Mahalo’s quirkiness, you may need to refresh the search knitting twice into same stitch

    Enjoy!

  96. “- social search (which I’ll define as improving search by unlocking the power of people)?”

    It is nice to see some of Matt’s predictions starting to surface. Anyone who has checked out the new Google Sitemap Interface will notice a section that shows how many people are subscribing to their sites’ blog feeds. This tells me at the least Google knows an approximate number of people who are bookmarking a site’s feed. An interesting post from Matt might address how/when/if Google will integrate this into their search results. But from this post it seems clear that they will integrate this data, if they have not done so already.

  97. As the web has changed, search has changed, so to have I seen Google change. I for one am all for human interaction but to a certain degree. I think there is a fine line between the good and bad of too much interaction from a human presence in search results.

  98. I worked once for establishing ethical certification standards for affiliate programs vendors. At that time we suggested a “Certificated Badge”. yes thanks ;

  99. Thanks for the NYT article

  100. Hi Matt this is a nice article from you, that talks about human experience. But, with centillions of websites online, I don’t think it is ever possible to ever figure out a method that with live humans sorting the list. It becomes necessary for a search engine to figure out methods those can sort the importance of a website based on user inputs those are automatically fetched.

  101. A client called me today and asked why his site wasn’t appearing in
    Google. I had created his humble site http://www.nbtrans.ie a few months
    earlier. So I opened google.ie and searched for noel brennan transport
    with the ‘pages from Ireland’ option selected. Lo and behold, there he
    was, first up. That was news to him. So I searched with ‘the web’ as
    the option and there he was again, first on the list.

    However, turns out he was using the google search from his ISP
    (Eircom.net). So I tried there, and sure, his site wasn’t in the
    search results even though I used the same search words. You can see
    more of what I found on http://www.webwizards.ie/crippledgoogle.asp

    So, does google deliberately or otherwise give different (older maybe)
    search results when the search comes from partner website and partner
    website is allowing unrestricted search?

  102. Wow, same as last post for one of my sites. Pulls differently when searched via different ISP’s. Can pull on comcast here go next do and pull on at&t, results are different. Also see different results when pulled from different search engines that use googles results, but that is to be expected.

  103. I was curious what you thought about the combination of the following:

    1 – It looks like Mahalo is just using vc money to pay humans to scrape your and wikipedia’s database.

    So…

    2 – why is it that they seem to place Google Adwords in the middle of their content in a way that really blends in with everything else. They use the same hr, header style, and everything so the casual reader might just miss that it’s an ad.

    I would think you were less interested in giving them a pass where other webmasters have been flogged.

  104. Hi Matt,
    For one website, how many keywords (max) you will give preference?

    By
    Sreejith sasidharan

  105. I think google.com’s new introduction of SearchWiki is a good move. We will be surely watching the behavior of these new introduction. If Search Engine Results improve HATs off to google.com

  106. Would humans be more or less empowered by an open search algorithm?

    I would like to hear more discussion on this topic at InfoWorld and in Matt Cutts’ blog. Jonathan Ledergerber’s article Humans and Machines in the Search Equation is right on point when he refers to search that is powered by humans.

    Should search algorithms be open, with underlying logic exposed for public review? If Bing were to lead the way forward by opening up its search algorithm, would Yahoo! follow suit?

css.php