Delicate Genius Blog

People are searching for “google” and “apple” in the images section? The mind boggles.

Do you have stats on text vs. image search?

Really sad at the @ the ‘Who to hate’. To me that means people probably went to http://www.bing.com and tried to find the results which closely matched that on BlindSearch.

That tells you Bing probably has a higher selection rating in reality.

thank you for creating an excellent ool. I for one welcome to competition of Bing for google.

I think the really important observation that can be made with this tool, is that most search engines, just like most tools/producuts in general (whether they be software or not) are _almost_ the same.

They were at one point different. Yes. But other companies have since caught up. They copied the “innovator” to a point where they’re almost indistinguishable. Copied their engineering concepts and copied their UX guidelines. Copied their marketing strategy. It feels like Coca-Cola and Pepsi or Windows and OSX.

I for one, will continue to use Google. Not because “technically” they’re the best search engine (however that technical merit is measured), but because I “feel” like they’re best search engine. I guess its almost like following your favorite football team. Through the good times and the bad.

P.S. I did one search on “web hosting” using this tool and picked Google.

I think this experiment is totally useless. Look at the most popular search terms. Mashabale? What is that thing? Mash potatoes? Today, most people search using long tail keywords. For example “how do i blah blah? Google is good at getting the satisfactory results most of the time. For the last month or so, I am going to Bing if I cannot find the thing I am looking for in the first couple of Google’s search result pages. But most of the time. But, un that scenario, Bing was also not very helpful. For shopping related search, Google gives lots of sponsed ads which are very helpful. Bing – scant sponsed ads.

Nice exercise!

As a blind test, it would be more useful if the three output columns were sorted randomly: currently, after two or three searches returning results in the order Yahoo – Bing – Google, the test is no longer blind.

Keep the ideas flowing!

I just blogged about your Blind Search (http://aces.arbita.net/blog/use-blind-search-to-prove-which-search-engine-is-best-for-you). I think the key is to plug in some very targeted kinds of searches and see if one engine performs better than another. In talent recruiting research, it’s important for us to know where to get the best/most resumes, person profiles, member directories, etc., within a given industry, job function, etc. Historically, we’ve already proven the overlap between search engines is low (just run the same query and compare the number of duplicates between any two engines at http://ranking.thumbshots.com — a tool which has been around for many years — to see visually how low it is). Now your tool lets us see the results side by side, not just what’s duplicated. One could argue that Zuula.com has given us this ability for a while, but the results for each search engine are on separate tabs there. I think Blind Search may change where you go first for a given type of search in the future.

Marketing will win at the end of the day.

In Australia, the old Coke Versus Pepsi taste tests were always won by Pepsi. Yet Coke dominates the market.

When the consumer perceives minimal difference in the experience then they tends to stay with the source they know and trust.

http://tinyurl.com/lx4ojp Here is my results on a few tests. Nice tools.

Michael, any chance of releasing the (query, preferred search engine) data? It would be interesting to see which engines people preferred for different queries.

Try searching for sex in text (not images) the parser for bing crashes, also displays an error containing full path disclosure.

Appart from that, great idea but it just confirmed what I already know; I’m a google man ;)

Google gave me the best results then Yahoo. Bing was crap. Google rules all.

as ben buchanan requested, can we have the detailed numbers for image search? :)

But what about cuil?? HA!

Decided to try it for a very focused text search (medicare changes under health care reform) that I had been looking for additional information on, and found minimal difference between two of the three search engines: Bing and Yahoo. I hadn’t tried Bing before, now I think I will add it to my list of engines to use.

I’d be great to have ask.com too (but won’t fit on my laptop screen)

Thank you mk.
At fireball.de we are screening all three search engines periodicaly. Our results are similar.
For me it gives me more confidence into our strategy.

Under that light, it seems the fusion between bing and yahoo are mainly for cost cutting and less for improving there already high search quality.

I am very surprised at the quality of Bing results. I only use Google because of it’s little things like currency conversions weather top Wikipedia results and because it’s sorta ritual. Bing is cool but Yahoo’s results sucks. Sorry but it’s true :)

For me, the main reason to use google is the fact that it is the most light weight of them. No unnecessary graphics etc.

The page loads fast and no clutter. The results are almost the same across the three providers anyway.

I know clutter, graphics and load times are not so much of an issue in the more developed internet nations. Here in Dubai, however, Internet is sadly underdeveloped and even a 1 to 4 Mbps connection is not as fast as you would like it.

Lightweight therefore is the way to go. And that means google. Same reason I use google mail to consolidate all my webmail addresses.

oh and add exalead too, to be complete

Some queries (like ‘hunchentoot’) show only bing, even though google and yahoo return good return good results.
This test is flawed.

Heya, nice job on the site.

For us fellow coders, perhaps you could write a quick blog post about how you implemented it on the back end?

What server side language? PHP? Asp? Java?

Are you using API’s from the search providers or just making HTTP requests?

If making HTTP requests, do you ‘pretend’ to be any particular browser or user?

For my keywords it is still a matter of which SE produces the least crabby results – Google and Yahoo wins 10 out of 10, so Bing needs some tuning for my niche.

I still prefer niche-surfing when it comes to quiz and trivia resources:-)

I also was surprised when I first choose Bing. I use many of Google’s tools and services and thought I would recognize the sites that came up when I searched for “news.” Yahoo is a give away when you search for anything they have a site or service for. For example, searching for “tech news” came back with only 2 of the search engines listing a result from yahoo. One was a simple url – “http://news.yahoo.com/technology” and the other had two – “http://news.yahoo.com/i/738″ and – “http://news.yahoo.com/news?tmpl=index2&cid=738.” The latter was Yahoo and the first was Google.

I thought that was interesting to see the self-interest influence the results. I like how google would place an advertisement for one of their services in a result rather than make it obvious a result has been altered.

I do enjoy Bing and currently use it as back up search. Usually I don’t ever have to ever leave Google to find what I’m looking for.

just wanted to let you know, google was the only one that had your site listed not once but twice with “blame the douche” XD

Depressing how narrow and unimaginative those who have played around with blind search seem to be. Sex and porn.

Anyway, I did 3 tests, “Lexicographical,” “Perseid Meteor Shower” and “Insomnia”

I picked Bing everytime.

I am a google person, for many of the reasons already stated–low overhead, familiarity. I actually find their news search to be annoying now with the links to you tube.

The blindsearch would be better if we could see the “next 10″ or however many results — as someone pointed out the top results are all pretty close (and I tried to pick the search that seemed to be giving me the most relevant for “Lexicographical”–since what I really wanted was “Lexicographical Order” but didn’t query it) – its the rest that kind of makes the difference–are there gems or junk in those next million hits?

Interesting experiment, thanks for sharing.

Hello,

To be fully honest, this test should present the result in the exact same way, whatever the engine. It is not the case know for images search:
The Bing images results are presented slightly bigger than the other two engine results, resulting in nicer pictures presented in a longer column. I’m sure it has an impact in the choice, when two engines are given quite good answers.

I am however surprised of two things:
-The quality of the Bing results: they are not as good or accurate as the Google ones, but still, it is a lot better than what the crappy live.com was.

-The Yahoo results are almost always far under the two other engines. It is specially visible on images search: It seems that the results are coming from far less indexed web sites and the pictures are quite old. Ex: I tested a bike model and yahoo gives pictures of the 2005 model when the other give mainly 2008/2009 models. I even can bet which image result comes from yahoo and nether miss it.

excellent idea! this is so interesting!

It’s sometimes easy to identefy the yahoo search results.

Example: search for ubuntu

Google and Bing has this as the first result:
Ubuntu Home Page | Ubuntu
http://www.ubuntu.com/

Yahoo only has this:
Ubuntu
http://www.ubuntu.com/

Yahoo misses the “Ubuntu Home Page |” and is easy to spot.

I seemed to get Yahoo when I entered one word queries and Google when I entered long queries. Can you you post some statistics on this?

I think blind search is an interesting research tool. Your results say Bing: 31% and Yahoo: 28%. That means when Bing is integrated completely with Yahoo, it will be on its way to surpass google? That’s a welcome news, I am really getting tired of Google now.

The headline results might be misleading, as some.

Across several tests, usually all three results were roughly equally good, i.e. they all got the right answer. But on a couple of trickier examples, only Google got the right answer, and at #1 spot. The other two completely missed the point and failed to get the page I wanted anywhere in the top ten. E.g. search for Aurora Fiesole because you want the homepage for that hotel in that town. Only Google gives the actual hotel homepage, the other two give a load of commercial spam sites about the hotel.

When Yahoo goes away it would be good to randomly select from one of the next top 3-5 engines and see how they rate.

This is a fun experiment. I hate to admit it, but I usually reach for Google first out of habit and just because I’m resistant to change. Then I try Bing, because the search bar is there. I usually find what I need and don’t ever get around to needing Yahoo. However, in my limited tests at home, I was equally divided between Bing and Google.

It would be nice if I could rank the results or if there was a tie button. Many times two have been fairly equal (usually seems to be google and bing) and one is just bad (seems to be yahoo) and it would be nice to be able to mark them as such.

It’s pity not to get a ‘Tie’ button when I’m indifferent to the search results.

Interesting results, but one REAL difference not included in the scope of this is presentation– e.g. the superior utility of image search results on AlltheWeb (I believe they use Yahoo data). Click on the thumbnail, and you get just the image at native size; click on the link below, and you get the whole page in context. So much cleaner than the cluttered presentation on Google or Yahoo itself. Shows that finding the information is one thing– presenting it in the most useful way is another.

My personal results show Google and Bing pretty much tied for first place, with Yahoo not that often- being a Mac fan I wasn’t prepared to be that impressed with the Microsoft product, but for a while Bing was ahead. But then I’m usually searching for technical information, not porn, or Michael Jackson.

Really, really cool idea. So damn simple we can only hit our heads agains some walls because we did not come up with this. =D
I experienced google being better with nouns and bing being better with names. But I only tested a few things.

This is the most self-serving Microsoft-nonsense I have ever seen – MSFT is as out of touch with the flow of technology as is possible – Ballmer (the bumbling idiot leading microsoft to eventual disaster) said about the horrid vista: “ahaha..guess we bit off more than we can chew.duh!” Every year MSFT puts out its Google killer and the next year is is dust – we put BING through a series of tests on searches we are very intimately aware of web content – it was truly a comic event – almost random results – so MSFT can shove another piece of crap into the web – and the web will spit it out – yes Yahoo is light years behind Google in search tech – but BING despite the curious traffic and despite MSFT’s arrogant fat wallet – will never touch Google search – but I do use it when friends are around – a great source of comic relief for my tech circle – we actually do a toast and round of applause when BING responds with anything relevant – BING is a joke – created by MSFT watching its two franchises – OFFICE and OP SYSTEMS going down the drain within 5 years…goodbye MSFT and good riddance – we have not spent money with MSFT for 5 years and after Vista we are heading to Apple on next resupply cycle….

The evidence of self-servingness of this site – just MSFT guy having fun (give me a break) and a quote from above:

“Surprise: Many were surprised that Google wasn’t always their choice. There were many instances of surprise that Google wasn’t as superior as they thought it was.”

Just as those guilty always confess – we just got the ENTIRE REASON of this deceptive site – to show people that Google isn’t their choice…

I have a question & statement:

(1) Why didn’t you mention the instances of Bing commens saying how poor it is?

(2) it is msft mentality like this & treating consumers like maleable chldren – that would keep me away from bing vista office (all inferior product) even if they one day issue responsible software..

the comments Who to Hate? portrays a near universal disgust of MSFT across anyone in technology we have ever dealt with…

MSFT btw spent millions on deception like this on mult-million dollar website/tv campaign to trick people into trying vista (by telling them it was new 0/S)…
and surprise – not everyone hated it…

This website is classic msft marketing manipulation and deception – if we ever were going to use BING – this website just shows us how the people behind MSFT products are both clueless and arrogant in their dealing with the consuming tech market…

If searches were all about text and blue links, this would be a great study. But it’s not about blue links and text anymore.

Go to Yahoo, Google, and Bing individually and do a search for “Manny Ramirez”, for example, and tell me which one you like better?

I prefer Yahoo’s and Bing’s results over Google’s.

This search has not been able to quite decide between google and bing….but one thing is for sure….Yahoo suxx…

Michael- to make the results meaningful, you clearly need to control for query mix…maybe use freq numbers from bing to weight results

Google is good but Bing is not bad either http://delicategeniusblog.com/?p=839

http://delicategeniusblog.com/?p=839 The Results are in! Google is better, but Bing is not that bad either.

8 weeks after BlindSearch: Google 41%, Bing 31%, Yahoo 28% http://bit.ly/Q9cJv. BlindSearch URL: http://blindsearch.fejus.com

Hi! I was surfing and found your blog post… nice! I love your blog. :) Cheers! Sandra. R.

http://bit.ly/E76ni – Results from blindsearch (http://bit.ly/6j51Q) which tries to compare results from search engines. Interesting.

Bing shockingly within spitting distance of Google quality in blind search tests! http://delicategeniusblog.com/?p=839 front page news!

RT @btabke: Bing within spitting distance of Google quality in blind search tests! http://delicategeniusblog.com/?p=839 front page news!

RT @btabke: Bing shockingly within spitting distance of Google quality in blind search tests! http://delicategeniusblog.com/?p=839

RT @btabke: Bing shockingly within spitting distance of Google quality in blind search tests! http://delicategeniusblog.com/?p=839

RT @btabke: Bing shockingly within spitting distance of Google quality in blind search tests! http://bit.ly/Q9cJv front page news!

I tried your BlindSearch and was very surprised about the outcome! Some months ago I used to tell myself that Bing sucks and Google holds the Holy Grail in its filthy hands. But Bing serves great results and they are different from those from Google.

A little competition rocks!

I will recommend your tool to my users. Thanks for the service!

Bing was better when I searched political questions. I found less FUD than google, but Bing sucked and google was the best when I searched computer questions.

If I typed in keywords like “directory server linux SSO” Bing & yahoo come up with a bunch of active directory pages. I already ruled out active directory, it’s too expensive, I’m looking for FOSS alternatives, and google is by far the best search engine for that. The top ten google results are all what I’m looking for when the other two search engines seem to be advertising something, or linking to corporate pages, which is NOT what I’m searching for. If it was, I wouldn’t have included the keyword “linux”. If I ran into Novell or Red Hat stuff with that search, fine, I didn’t type “FOSS”. Google seems to be the only search engine smart enough to ‘know’ active directory won’t run Linux, and that’s why it returned better results.

Another good test was “NNTP suck” suck is a program used with nntp, but because of the other uses of the word “suck” it trips up search engines. Google gave me exactly what I looking for again, and the other search engines did not. Yahoo returned crap, Bing returned “buy hosted newsgroups on rackspace” while Google returned the manual page for the program suck.

If anything this reaffirms my commitment to google, but I thought you deserve an explanation, and if you’d like to use my comment to improve bing, be my guest. That’s how those of us in the opensource community work (actually, we’d usually try to fix it ourselves, but…). Your problem defiantly isn’t people gaming the test; it’s bing.

Searching sex was great. Is google’s safe search on? If not google is running with a handicap (well, depending what people search). Google returned porn & information, yahoo returned information, and bing threw and exception. That was an easy one.

okay, i’m sorry to say this dude. I like what your trying to do and everything but your test of Yahoo vs. Google vs. Bing is probably not usually blind. On the contrary. It is likely, and very much possible, that your test is biased by fans of one of the search engines. How? Simple. Minimize the “blind” test window, and open up a new internet tab. Use whichever search engine you favor. Then, type what you searched for in the test engine and . . . Wallah. You can read the results of your favored site then find the matching results on the “blind” test tab.

I love blindsearch-thanks for taking the time to invent and maintain it.

bing and yahoo have too many ads up the top of the search, google’s so much neater.

Could you publish the updated scores? Thanks for this very useful service.

Warning: Invalid argument supplied for foreach() in /home/.taniika/mkordahi/blindsearch.fejus.com/search_funtions.php on line 143

I understand why people don’t choose Google for the sex results because of the safe search which is activated!!! They prefer hard results and it tends to biaise your statistics ;-)

@CurtinCIO Results: Google: 41%, Bing: 31%, Yahoo: 28% http://delicategeniusblog.com/?p=839. I used, Google was/is still better.

I typed my own name in and got the best results from Bing. But when I typed my name directly into Google.com, I got different results than those reported by blindsearch. This seems to undermine the credibility of blindsearch, in my mind.

I found that Yahoo consistently returned the best results for blatant searches (e.g. Tori Amos or Lady Gaga or Windows 7). Meanwhile, Google was consistently my choice for detailed, contextual searches (e.g. song lyrics, movie quotes, product numbers etc.)

Not once in over 30 tries did I ever select Bing. I found that Bing had more irrelevant results than Yahoo or Google and, more importantly, less relevant results were near the top.

Even when all the results were similar, Bing had a lot of junk near the top where as Yahoo and Google had relevant results at first and they got worse as you went down.

Overall I like Yahoo search a lot more than I ever did; I like Google just as much as ever and I’m thoroughly unimpressed with Bing.

Visually Google wins, Yahoo is too simple looking and Bing is just annoying.

thanx much ,got a lot of help by comparative study of the big demons please continue your fun so that we can see more like this i would love to make it as my home page

You should add duckduckgo.com to the test/results. :)

Another good test was “NNTP suck” suck is a program used with nntp, but because of the other uses of the word “suck” it trips up search engines. Google gave me exactly what I looking for again, and the other search engines did not.