BlindSearch: The Search Taste Test
UPDATE: I’ve published the results from the first 8 weeks of BlindSearch here.
—
I thought it’d be fun to see what happens when you do a search that doesn’t have branding. Are we swayed by the branding of our favourite search engines?
To figure it out, I thought I’d try my hands at building a “Blind Search”. Basically, you enter a query, get back three results in random order then select the column which you believe gave you the best results.
Will be interesting to see the results.
Check it out over at http://blindsearch.fejus.com/, Have fun!
Also, please leave feedback below.
-mk
——————-
update #1: The poll has been compromised. Probably equally due to my lack of leet skills and to the malicious person gaming it.
I’ve removed the ability to see the results until I sort this out. Meanwhile you can still have fun playing with blind search. Feel free to blame the douche for ruining it for everybody. Meanwhile, I wouldn’t take this a scientific whats-e-ma-jing, it’s just a bit of observational fun.
——————-
update #2: I’ve introduced a basic image search
I realy like the idea and it also works mostly great…
My only problem exists if i enter a querry like “USERENV(474.cf8)” the errors appear mostly with error masages with special chars in it.
else grate work.
P.s. would be interrestet in count of the search engines.
appears to be a problem viewing with i.e. 7 (i know i know). at first i did not have the window completely expanding and it was cutting off about 1/3 of the first, left, column. after maximizing the window the left column disappeared entirely and the remaining two columns shifted to hard left.
Fun test! Bing and Google seem to be doing quite good both in the blind test. I’ll stick to Google for most parts in my everyday life. Here’s why..
First impression of Bing
1) The overall layout and presentation of information – Keep it simple guys. Problem: The image in the upper left corner is quite detailed and requires decoding, this removes focus from the actual search results.
2) The Logo – Problem: The design is below average. Color scheme also strange. But what has this to do with search results you may ask? Well for me this is relevant since I have to look at this it from the corner of my eye every time I do a search.
The results are very similar for many searches, so much so that there isn’t anything really compelling about any particular set (oh, dear, the commodification of the search engine). Have you considered doing a diff of the search results and then analyzing preference based on high diff searches, i.e. when there really is a difference, which is prefered?
Another way to potentially differentiate these search engines is to look for proprietary information and how they rank.
Since Yahoo owns Flickr, images from Flickr are more likely to show up in Yahoo than the others.
Since Google owns YouTube and Google Groups formerly DejaNews, these are more likely to show up in Goolge’s listings or be ranked higher or have more of them.
This method doesn’t worke every time but it does increase the odds of one’s guess.
I tried searching for daschund, and noticed the 2 YouTube videos right away in Google’s column, but there was only one in the Bing column, and none in the Yahoo column.
Clicking on the images and finding the Flickr image helped me id which was Yahoo. Of course, that leaves the other two.
It’s absolutely awesome to get all three big search engine results at the same time. Blind Test is a great search engine. I give it my vote as the best.
Don
When I submitted
allegra!jpl
(my old uucp-style email address),
Bing came back with
Warning: Invalid argument supplied for foreach() in /home/.taniika/mkordahi/blindsearch.fejus.com/search_funtions.php on line 130
Hi,
A really nice initiative to compare the new search engine of bing to other search engines. But it got a bit broken when I treid to search `JS selection class`: http://blindsearch.fejus.com/?q=JS+selection+class&type=web
By the way: you made a IE8-accelerator. You can also make a kind of `search engine` for firefox:
Blindsearch
Blindsearch: Comparison of google/bing/yahoo!
UTF-8
Keep up the good job ;)
Ciao
Because of this blog removes tags from your post (in stead of replacing it by htmlentities). You can read about making your own search engine: http://www.whelp.nl/zoekmachine-plugin/ (sorry, it is a dutch page. Even as I am a dutchy :D).
@all, i know it’s very buggy at the moment. and doesn’t handle escapes of all kinds very well. I will get around to fixing it. I will try and address all of your feedback scenarios.
Love the idea of a “meh”/none vote.
Will also work to making it hackproof. I have some ideas, just need the time to work on them.
It’s also astonishing me how many people are using this as their search engine of choice. You’ve left me speechless.
@Jan, there is an OpenSearch provider there already, is that not sufficient?
I quite like the idea. Kind of fun. Heard about you on NPR.
The only thing that is really buggy for me is that the screen is shifted about 2″ off the left edge of my computer screen. No scroll bar. I have to guess at the first couple words of each line and I have a brand new 22″ monitor. I would use it all the time if not for that.
wow! NPR, which show?
What browser are you on?
I heard it on Tuesday so it should have been “Digital Cafe.” The show included an interview, I asumed you knew about it. The spot started out as an introduction for BING.
I am using explorer 6.0 and the columns are just not wide enough. The columns are 2 across and one under the others on the left of the screen. There is a big blank space to the right where the third colum should belong.
>>>>
@Jan, there is an OpenSearch provider there already, is that not sufficient?
>>>
Will also work to making it hackproof. I have some ideas, just need the time to work on them.
Sorry, for previous post(I didn’t notice I used brackets. So it was handled like HTML):
++++
@Jan, there is an OpenSearch provider there already, is that not sufficient?
++++
First: My name is nog Jan, but `Jan Willem` (double name). Would be John Bill in English.
Second: This combination of search-engines is not my first-choice of search page. I just like google because of the simplicity, Bing of it’s speed (for as long as it’s not going to be used by the whole world population). But I just like the idea of measuring the popularity/quality (I hope the latter) of search-engines.
Third: No, your `opensearch` is not enough. I suppose you don’t use firefox. But I do, and it’s not fit for this one (version 3.0.11). In each case FF is not in the mood of recognizing it as a search-engine or install it.
++++
Will also work to making it hackproof. I have some ideas, just need the time to work on them.
++++
Well, perhaps it may be an idea to move the vote-handling to serverside. Now I can pretell wich result will be wich search-engine, by using the error-console of firefox.
Another vote for ask.com, and possibly cuil. Though I’ve pretty much ruled out cuil. I’ve been using ask lately and would love to compare it to these results.
Nice, thanks!
I’d like to add to the votes for more search engines. The more search engines you have, the more interesting your new toy becomes!
There are actually a lot of search engines. You can find a list at http://en.wikipedia.org/wiki/List_of_search_engines and http://www.dmoz.org/Computers/Internet/Searching/Search_Engines/ .
There should be as many search engines as possible, and for each search 3 would be randomly selected.
Maybe a “Yes/No” voting system should be added where you could vote if you found what you wanted.
Excellent initiative. But don’t keep the resukts for yourself please! What makes you think your test has been compromised?
@Régis the few thousand votes in a row for yahoo from the same IP address gave it away – then the 5,000 more votes in a row for yahoo from different ip addresses made me concerned.
They have stopped gaming the system and i’ve taken a snapshot. Results are roughtly GOOG: 40%, BING: 31%, YHO: 29%.
@com adding more search engines is a good idea, it’s been requested many times, may add it in the future.
when are you going to publish the usage/vote results?
Hi, Michael,
Just an FYI. I did a search for Leo Masters banjos and then revised it by using quotes: “Leo Masters banjos”. The results were two porn sites (I didn’t actually open them – am just assuming by the descriptive sentences) that I can pretty much guarantee had nothing to do with banjos!
Regardless, congrats on your cool site! I find blind comparisons VERY interesting.
Fran Cincotta
College Librarian
+1 on styling the visited results… it’s a quick change and stops me realizing which one is my ‘usual’ search engine :)
a:visited { color: blue; }
Is each search engine always in the same column?
Neat service. Have written about it on http://www.mahesh.com/2009/06/29/forget-the-brandlogo-which-search-engine-works-for-you
Great game! Well done, sorry to say you I did quite a lot of query from egoquery to popular “barack obama” and picked every single time Google. No cheating.
very useful search engine. I removed google and made this as my home page.
Wat an amazing feeling. playing with the search engines is quiet cool.
though web search giv random result, my vote in image search always go for bing. bing rocks in image search.
good work michael
my personal results tell me that yahoo is the best in text search, followed by bing and google on par.
and this surprise me a lot.
Nice tool. Looking forward to see the stats again. For me Yahoo works best with short queries – max 2 words. For more words, Google wins.
Just a note–because of human behavior and the fact we read left to right, the left-hand column will be selected more than the other two, all things being equal.
Have you come up with any way to control for this? Perhaps weight each vote according to column? i.e. left column votes get valued at .9, center 1, and right 1.1?
Love it!
Not only is it fun to see which search engine works best but it’s also use it for actual searches… ditch all of them and use Blind Search! ;-)
It’d be very interesting to see the data acquired – which search engine is #1 when used through Blind Search?
next page button?
Hi,
Even though you can’t show us the global stats right now, it would be very nice to track individual stats (cookie-based), so that at the end of the day, I could know which search engine was giving the best results for me.
How doest it sound ?
Best regards!
greg
Oh, just another thing.
Some times I feel like all three search engines results are equally good (usually showing the same results in the same order), and still I would like to know which is which, out of curiosity.
So it would be nice to have a “All are good” button that wouldn’ cast any vote, but would just show the engine for each column.
Search results do not show those which appear in the actual search engines. Has anyone else noticed this?
I found the link to the site on Site Pro News – I had started seeing Bing creeping into the users finding our site recently. I tried to sign up with the microsoft adcentre but they have a current problem with their credit card authorisation (couldn’t save any of the work either!)
Excellent tool,
I have a question, just for fun I entered a new site I have just put up (www.problembaby.com) It came back on google, I checked with google directly and as I thought It is no where near first page google, how come your tool managed to bring my site up on top 10?
yeh! bing is rocking. i like every search engine.
but google is best and Old is Gold :)
HI,
When i enter a search query like “Passport+USA” one of the search engine always returns an error.
Seems that Bing doesn’t support these.
Am i right?
Tarak
If I search define:anything. The search engine with no result is google. :)
Since you are tracking which one each tester likes best, you ought to have a button for “none” since it is possible that none of them found what the searcher was after. Let’s be fair and show if there is (or is not) a need for better search.
Thank you for this tool, it is great and thought-inducing. May I suggest styling a:visited as a:link? It can really spoil the results of a test.
After three searches my vote was for Yahoo all three times. Which is interesting since I never use Yahoo.
I suppose everyone has noticed that if you search for “email” yahoo puts yahoo email at the top, google puts gmail at the top, while bing only puts hotmail in 4th place (still ahead of yahoo and gmail tho).
Interestingly Yahoo is the only one that puts yahoo at the top of search for “search engine”
Yahoo also manages to put Yahoo maps at the top of a search for “maps”…
Yahoo gives top spot for “news” to Yahoo news, I could go on….
these are supposed to unbiased searches, arent they?
Not very scientific i know, but it makes it look like yahoo is the most biased and bing the least
Wow! That was slick! Good Job! Google was my pick.
Try entering “CFS size limit” both in your front-end and directly in Google. The result sets for Google are different. In fact Bingo’s results look more like Google’s.
Interesting site, reminded me a lot of DogPile.com you did a nice job. Liked the “twit this” side project pretty funny stuff. Good luck with your project.
Love the site, but can you tell me how to add your search as the search for my google chrome browser? I don’t like IE (exploder).
Wow very cool. I voted and what I though would be google was bing. I guess bing is pretty good.
Hey cool blind test, was surprised at my choice of Bing! Hey do you keep stats of the which SE goes ahead? That’ll be interesting…
That was a fun test. I think it is biased to Yahoo.
I think we should not jump to conclusion only after several search.
When I first try some random words like “mail”, “orange”, “vertex cover”, “lifehacker”…. I voted for all three search engines. I’ll say I thought Yahoo wins a bit.
But, when I leave Blind Search as my search engine for the rest of the day, the result is different. When I want to find sth(well that is really what a search engine is for), I vote constantly to google.
For example, I once saw a beautifully designed wine bottle and want to search the picture. I type “Samurai design wine” for image and the result appear ONLY IN GOOGLE.
I recalled a new lifehacker post today about saying AEIOU to speed read. I type “lifehacker aeiou”. Google gives the correct post from lifehacker and from other sites referring to it. Bing and yahoo gives TOTALLY IRRELEVANT results.
The shocking thing is when I add “speedread” to more specify my query. Yahoo return NO RESULT and Bing IS BROKEN. Google gives perfect result as last search.
When I copy the post title “Repeat “A-E-I-O-U” to Read Faster – Reading – Lifehacker”, All three gives correct results. But would I remember exact wordings of tiles in order to find the post!?
I would not say google is a better search engine than the other two. Bing and Yahoo simply dosen’t qualify to be a search engine. I’d say Google is the only search engine.
hi, since yahoo is now powered by bing, why are the results different?
Not so blind.
I searched for “pudge” and the entry for my web site (#2 on Bing and Google) showed the last modified date as part of the entry. I see that all the time on Google, and not so much on Bing and Yahoo. So I figured it was Google in that column, and I was right.
Yeah, not that blind. If searching for music or shows, the ones with Youtube links are invariably Google links. But I guess that’s not so surprising. still a worthwhile experiment, especially for obscure search terms :)
Search term: what is cloud
Hurrah, i voted for Google.
3rd result of STUPID bing was “Cloud County Community College”
4th was “have you ever wanted to fly in the clouds”, STUPID Bing!
Well done & interesting.
Useful as well. Tnx to you I am moving from google to bing – 50:50…
regards
vlasto
ps.
I was looking for
“Levelised Energy for major new SEQ Water sources”
and there was nothing on google but yahoo/bing both came back with a hit. One only but the one that I was looking for… :)
Michael-
Thought you might want to know:
All 3 searches appeared all the way to the left of my screen (one under the other) with the the left margin cutting off some of the text (several words)….
XP Professional
IE6
I know a lot of people that’s trying Blind Search to get the real taste of search without brands. But some of them complain that the page render in IE7 is wrong.
Could you please fix it? Thanks a lot.
I searched for
“You’re not good, you’re not bad, you’re just nice”
with quotes and apostrophes as shown. One column gave me one good result. one column gave me a syntax error, and one column (google) was empty. So, I voted for the first column (yahoo). When I searched on the same string from the google toolbar in Firefox, I got many good hits.
Michael,
Your site may a “fun experiment” but it also is a useful site that enables students to learn that search engines are the not the same; they do return different results for the same query. Most importantly as the complexity of the search query increases; for example, from “climate change” to intitle:”climate change” AND “global warming” AND “polar ice caps” the difference is the results also increases. And the greatest teaching point: Google is not always the best search engine.
Cheers,
Paul Barron
Director of Library and Archives
George C. Marshall Foundation
Великолепный подбор информации, только почаще освежайте ее и уверен прибавится много постоянных посетителей.
Достаточно интересная статья. Думаю стоит добавить в избранное для дальнейшего изучения
“The system has many flaws that I know about already, the primary one of interest is the lack of localisation. So, all searches are going through the US as US searches.”
This is a great benefit to me in my work. I am based in the UK and I am trying to see how my sites perform on Google, Yahoo, and Bing in the US. Google results always need “de-personalising” and Yahoo.com keeps defaulting to Yahoo UK.
Your site seems to save me all the hassle of going through US proxy servers to see the results, and I now get them all on one page!
Thanks!
YAAAAH! Imma google person all teh way!!! I searched for
companion cube and google had teh bet pikshurs :D
Your search engine confirmed what I suspected.
Bing is garbage
People leave so many weird comments. I used each search engine separately, then compared the results with the blind tester – same thing (of course!). The point is to vote for the best user experience without bias. For the most part, my vote went to Google, until I started searching images, and Bing led. I had so much fun. Is it useful? I don’t know. We all know the winner, but for sure it’s really cool tool. Thanks much
I’m trying to be as objective as possible, and I’m finding that Google leads by a reasonable, margin, Bing is next. I don’t seem to choose Yahoo as often, but when I do, it’s really obviously the best, and the results are quite different from the other two.
I’m a fully programmed, washed-in-unicorn-tears Mac fangrrl, so I’m amazed at how often Bing, a product of the Evil Empire, wins. (Kind of a cute name too.)
Yahoo came in third, but I have found that when the Yahoo results are better, they are much better (and usually quite different) from Google and Bing.
I find this an interesting experiment- think I will have to start tracking my own results to see if there is any correlation between the type of query and which search engine tends to win. Have you thought of doing anything like that?
FWIW, Blind Search is now my default search engine.
why not google index my all links(each links)
in my site 1k+ links :( but google only index 9 links
It’s a nice demonstration of the fact that “organic search” has reached a plateau. Everybody gets roughly the same results.
It’s possible to do better, if you crank in some other data sources.
You should release the statistics of what your visitors prefer as their search engine for both images and web!
I tried something: I made a search, and opened another 3 tabs and searched the same term myself in bing, google and yahoo….
Out of the 3 searches presented, I could identify the bing candidate and the yahoo candidate (perfect match with the original source)… but the allege google result had nothing to do with what I got from google itself….
So it’s like the ‘Pepsi Challenge’, except that you are served root beer instead of Coca-cola and then marvel that coca-cola drinkers prefer Pepsi over root-beer!
Hi
What about the search results left out in the first page.Pl advise quiet good.Pl update and help beginers like me.
Bi
P..Murugan
Very cool tool. I searched for about 20 things and picked Google 14, Yahoo 4 and Bing 2. Then, I searched for a bunch of images and picked Google all but once, the other was Yahoo. Bing included totally non-relevant images – I searched for “dumptruck” and got photos of a concert from Bing…
Good blog content, thanks for dropping-by! Good luck ;)
Please make BlindSearch compatible with Firefox! I love this site – i almost always (65%+) choose Google but would love for this site to be in my browser instead of a bookmark! …oh, and publish the results as an up-to-date % if possible!
Am I the only one here who sees some inverse logic in this voting thing?
Vote for what you THINK is the best set of results? Huh?
Okay, nice game, but what’s the point?
Surely logic dictates that a vote AFTER you’re aware of results only makes sense?
Why woudln’t you leave your page open and raise the links in new tabs? Like a site survey. You do your will in a site and then respond to the survey you’ve agreed to earlier. Doesn’t that make sense?
Excellent! You’ve done an amazing work on Blindsearch!
Surprisingly the most accurate search results were provided by Bing and Yahoo on our domain name…
Hello, I’ve been using only your site to search for things ever since i’ve learned about it.
But today, I was very disappointed to find out that the site is vulnerable to security attacks like XSRF (try putting “alert(‘hi’)” (without the double quotes) in the search box) :(
site error :
Warning: Invalid argument supplied for foreach() in /home/.taniika/mkordahi/blindsearch.fejus.com/search_funtions.php on line 130
Hey, Now that both the search engines have matured, can you conduct a test now ?.
I’m curious to know the final comparison results (the ones you’ve posted are too old) over time.
It’d be wonderful to post the results of past 2-3 months given that Bing has matured a lot.
Thanks :)
found the site through a google search.
Nice idea but it simply doesn’t work – if you put your search into blind search, pick your choice and see the results. Then put your search into the original bing, google or yahoo you find that they don’t match at all and that blind search is deliberately making one or two of the 3 look worse. How or why i do not know but i will not be using it as a comparison
The simple search for “Sora Aoi” yields error for Google results.
Yes, surprisingly, Bing has gotten me the best results. That, coupled with the fact that google seems to hate me has made me switch to Bing for a while.
I just stumbled across your grand idea. I will continue to use it. It is one of the unsurpassed concepts, I have seen. Works for me personally and professionally!!!!
i tried your site, launching the same request to bong and your site … not the same results … so !?!
Why aren’t the results the same as if you just searched bing for it?!? also, shouldn’t the Yahoo and bing results be the same? There not!