|For any publisher, optimisation is increasingly about what you don't allow the engines to spider, as well as what you do. |
It's also about how you package information. Facts are usually in the public domain, and they're usually available from any number of sources. If you can deliver information in a form that people want to read (and that encourages people to stick around or come back, as they do with Wikipedia or Internet Movie Database or THE NEW YORK TIMES), you'll gain a lot more from search engines than you lose.
Based on lots of experience with interview material, I feel that the broad sweeping questions with which the article is framed simply don't match up with the minute specificity of the quoted answers or with my sense of Eric Schmidt as a speaker. I've heard Schmidt speak on numerous occasions, and he's simply much more articulate about visionary-type questions than the quotes used in the article suggest.
The initial question asked was...
|"What are the hard things to be solved in search in the next ten years?" |
The quote we're all kicking around, about one answer, sounds like it's a response to a follow-up question about how a natural language "Google Answers" type feature might work. I don't see it as Google's overall vision, or as an indication that Google is trying to brainwash us.
Looking at what Google has been doing... yes, the disambiguation of some queries, which in turn may limit the range of answers we can get, has been driving me nuts. I deeply hope that they will fix this.
But I also feel that Google has in fact strained to index a very broad range of information, and to present a broad and diverse spectrum of that material in its queries. That's simply not consistent with a one-answer-fits-all approach which is being predicted here. I suppose you can try to turn this inside out and say that an all-media-fits-all is essentially a one-answer-fits-all approach, but I'd have to respond that Google's attempt is to diversify, not to get simplistic about serps it returns for a query. Also, as has been pointed out, if Google were trying to maintain ad revenue, as I assume it must, it would not be distilling down its overall serps to a single-sentence, mini-Knol response.
The suggestions about data aggregation are something I'd be much more concerned about....
|So you go to a very good definitive site. And what I’d like to do is to get to the point where we could read his site and then summarize what it says, and answer the question ... Along with the citation and so forth and so on. |
This is not the place to discuss data aggregation in general, except perhaps to say that it's been happening on the web in various ways since day one (or day two, anyway ;) )... that it has both drawbacks and benefits... that it's a societal issue, not just a web issue... and that it's increasingly prevalent. The Rich Snippets and Microformats discussion is one that I expect will continue and get more heated over the years.
Digital content can be easily copied. It's harder to copy, of course, if it can't be found.
|...a viable alternative - non profit search engine |
Be careful what you wish for. ;)
I'd love to see more competition out there, btw... and I'd love to see the Library of Congress doing some of the things that Google is now doing, but that too could be problematic.
|The quote we're all kicking around, about one answer, sounds like it's a response to a follow-up question... |
Point taken, Robert. The whole thing would make more sense that way - and if that is indeed the situation, then Mr. Arrington did a pretty poor job. In fact, it is a pretty short article given that this was a one hour interview. I basically felt cheated first, and then I got upset by the quote.
|In fact, it is a pretty short article given that this was a one hour interview.... |
It's the second part of a series that Mr Arrington is dribbling out every few days.
I should say clearly that I was put off by the articles. With regard to Google, I don't know whether Arrington has an agenda or not. I can say that journalism often thrives on controversy, and artificially created controversy has become a modern disease.
I thought the "Connect It Straight To Your Brain" part of the title, eg, obviously intended by Schmidt as a joke (a joke which many friends have made as well), might have been opportunistically hyped to suggest a darker agenda. I don't know... I wasn't there... but a close reading of the article makes me, at least, think that it was milked for all it was (and wasn't) worth.
I'm waiting for Sooty to post in this thread ..or maybe Lamb Chop ..
>The absurdity of thinking that there can even BE
> "the right answer" just jumped out at me.
Of course there can be one right answer. Some times that answer may be multiple options. In ten years, computers will all have retinal or finger print scanners on them. When you set down, you will connect to "the cloud" of computing, by being scanned.
Google will know your entire 'search' history - probably your entire email history (gmail), tracked your life via gps (android), your preferences and tastes (google news), your browsing habits (chrome/toolbar) your purchasing habits (g checkout), and a host of other things there are to know about you.
Given all that, if you ask Google a question, they should be able to give you the 'one' answer you are looking for with a very high degree of certainty. There are currently around 6 billion people on the planet, and we all do the same stupid stuff. Sorry, we are not that unique. A couple thousand variations in the algo is probably all it will take to nail most of the human population. In ten years, Skynet will look like a Commodore 64.
|Of course there can be one right answer. Some times that answer may be multiple options. |
But one right answer can't have multiple options. ;)
"But one right answer can't have multiple options."
Of course it can. If I search for the sexiest woman alive, it could feed me Jessica Biel results. If the guy across the street made th same search, the right answer for him would be Megan Fox.
There could be literally billions of "right" options to answer that query.
That's the easy part of personlization. The hard part, where the engines now fail totally, is trying to personalize searches like [barack obama] or [halley's comet].
[edited by: steveb at 8:44 am (utc) on Sep. 7, 2009]
|sexiest woman alive...There could be literally billions of "right" options to answer that query. |
You can just imagine the Divorce proceedings. "My Google search told me he wasn't my ideal husband, and his says I'm not the sexiest woman for him"
Some questions with definate answers I look forward to resolving:
Will smoking kill me?
When will I die?
What is Bill Gates' Social Security number?
What is the solution to quantum gravity?
Where I come from one right answer is one right answer.
I think BT was suggesting any one person will have one right answer. "Behind the scenes" there are multiple answers, but only one get presented as determined by the personal preferences of the asker.
FWIW, I don't buy it either
"Where I come from one right answer is one right answer."
Are we the same age? The same height? Have the exact same interests?
If John in New York and Mary in Mongolia type in [local restaurants] it's absurd to pretend that the one right answer for John must be the same as the one right answer for Mary!
|If John in New York and Mary in Mongolia type in [local restaurants] it's absurd to pretend that the one right answer for John must be the same as the one right answer for Mary! |
Agree, as far as it goes. But there is no "right" [best local restaurant] single result
These results may well be personalised to the searcher, e.g. if you search for a name of a steak restaurant of a certain quality, and then search for just "fish restaurant", you may well get back results of just fish restaurants of a similar quality to the steak restaurant.
If Google went onto buy Facebook or Twitter, then they can find out even more about you. Facebook PPC targeting is scarily good. I am a fan of a band on there and sometimes ads show for a rival group...also they show me ads for things SEO related...
the sledge hammer to crack a nut scenario. While some questions are improved when you know something about the asker most answers are reduced in scope if you always try and narrow the answer to what you know of the asker.
In the real world with real AI (which is what they are trying to emulate) you would ask the question, which is the best restaurant for me and the source would then ask the pertinent questions such as do you want a local restaurant? Do you want a particular type of cuisine, what is your budget range. Thats how AI works. They do not assume the answers to all the afor mentioned detail requests first and then give you their idea of what you are probably looking for based on what the majority would have answered to those questions.
So if you type in local restaurants dont assume local is where i am now or have been in the past, just ask me. Just add a couple of steps to truly tailor the results because otherwise im having to do that anyway by adding stuff to my search term.
|it's absurd to pretend that the one right answer for John must be the same as the one right answer for Mary! |
Personally I think it's absurd to even consider that there could ever be a single answer. Even with personalisation, let's say that Google decided (in its wisdom) that I was the type of guy who liked Indian food. If I did a search for "restaurants in townname" I may feel like eating Italian or Chinese that day. How is G going to know that?
Teach people how to search effectively and there would be no need for personalisation.
Doesn't "one right answer" based on personal search history (personalization), as it might be applied in reference to serps, preclude opening a new line of inquiry into a topic the searcher had previously researched, or at leasyt make it more difficult to do so?
What if the motivation for the search changes, how long will it take for the "one right answer" to change to take the new motivation into account?
I'm not sure that 'In the future' we will be still searching for our needs on a search engine. I know that the company I work for is investing serious money into social network integration and promotion and leaving the search engines to scrap amongst themselves.
I've always wondered what percentage of G hits are by real searchers and how many are from disgruntled webmasters. If my bounce rate is anything to go by the webmasters are keeping the search business going.
they just need to ditch the principal that you input a search term/phrase and you get a serps. Theres no logic that says it needs to be a 1,2 process. You dont need refine your search suggestions but refine questions based on the search, thats true AI at work. The clever bit is knowing what questions to ASK not what to answer one time with a whole bunch of assumptions and guesses.
Hello Badcol - welcome to the forums. Interesting observation. There certainly may be a webmaster effect (I know I help to create a bounce rates all day long). There's also a bot effect, but that can be filtered out more easily.
If the questions we see on WebmasterWorld from hobbyist and newbie webmasters are any indication, turning to search engines to answer questions, is often not the first impulse ;( Just moving the general population to use search at all is a slow go in many areas. Navigational searches are so insanely high that the Chrome browser went to using the "omnibox".
|I know that the company I work for is investing serious money into social network integration and promotion and leaving the search engines to scrap amongst themselves. |
The issue is that the same phenomenon happens there - social media is overloaded with marketers of all types who appear at face value to be potential customers because they are 'following' your business :)
"How is G going to know that?"
Read what Brett posted. They intend to try and find the right answer for *you* for that specific query by reading your gmail, geotargeting the exact coordinates of your computer, assess your shopping and surfing habits, etc etc. If they further read your credit card statements, tap your phone, and follow you around, they will do a better job.
That is what Schmidt is talking about... accumulating all sorts of Big Brother-ish information about each individual searcher.
There always is a "right" answer for a specific person, and it differs for each person, and it can change over time (like a 12-year-old searching for "how tall am I in centimeters"). The thing is now that Google is a million miles from serving up the objectively, in-the-eyes-of-god best result to each searcher now. Schmidt wants to get there in a decade. I doubt they will, and I doubt most people would want them too (given the loss of privacy).
|one answer, which is exactly the right answer |
Maybe this means eventually there will be only one Adwords spot on the front page. Whoever bids the most could not just be at the top, but be the only one.
|Maybe this means eventually there will be only one Adwords spot on the front page. |
Thanks for the welcome Tedster and CainIV,
Maybe what Google needs is a more personal search service which can then flag certain websites that fit our pre-programmed ideals. Having a Google online character which, say for arguments sake, is 42, works in IT, enjoys sport, reading and socialising. Enjoys a bargain and has very little affiliation to any brand types could certainly offer Google the opportunity to narrow down the results that they provide.
I think that Google providing just one correct answer to every question could be a possibilty if Google knows more about the searchers character. Obviously, by professing to offer this service Google will open themselves to a huge amount of critism and they don't like that ... so all things considered I doubt whether it will ever happen.
In defence of Eric Schmidt, what's wrong with aiming high ?
Brett, are you suggesting that Google's goal is to become like ChaCha? Seriously, if one receives a desired and acceptable answer to a query, won't they then feel less of a need to click on an advertisement? How will Google make a profit from that approach?
So much stuff is out there that we would never come across if we did not click onto "wrong answer".
I guess people that don't do much online, don't use Gmail, and whatever, they're out of luck. ;) Lucky them - being out of luck about right answer.
I wish they could leave it alone...
I can't recollect that I ever entered a single question into the Google search box.
I put in words the website I am looking for is likely to contain.
But maybe thats the reason Google search results seem to be a load of rubish to me more and more. If they try to intpret what I am entering as question - good night google.
|I guess people that don't do much online, don't use Gmail, and whatever, they're out of luck. |
They may still have a Yahoo, Bing or Google toolbar...visit sites using Google Analytics or AdSense...have cookies from Google on their PC...maybe use Chrome as their browser when they saw it on the Google homepage...
"There always is a "right" answer for a specific person,"
You will never find it if you think its always based on knowing about the specific person at the point of his first click. How much does Stephen Hawkins need to know about me to explain that the smallest thing in the Universe is a superstring?
Personalisation works but for the searcher its knowing about the source, (whether the source is authoritative for the search or not) not the other way round. Sure you may now give some specific examples where knowing about me helps but if you are a search engine or "Answer Engine" that approach will lead to reduced efficiency/accuracy/spectrum of information over the whole spread of searches not increased. The major breakthrough will come when they relent on this one click approach and start interacting with the searcher by asking the right questions. They need know nothing about me at the start of my search, they just need to ask the right questions in some added steps and hey presto, that's when personalisation works. No guessing and no assumptions. Its the old 2.2 children in the average family paradox. The totally average family doesn't actually exists and if you cater to them you are never getting it right at the individual level.
|Seriously, if one receives a desired and acceptable answer to a query, won't they then feel less of a need to click on an advertisement? How will Google make a profit from that approach? |
Very good point. I wonder if Schmidt has though of this? :)
| This 98 message thread spans 4 pages: < < 98 ( 1  3 4 ) > > |