Sometimes you come across a set of search results that just don't make any sense. For most ordinary users, I suspect they probably just move on to the next query, but for those of us deeply embedded in the world of search and SEO, these noggin'-scratchers just keep on itchin'.
I've collected these ten over the past couple months and figured I'd share them on the blog with the hopes of getting greater group intelligence to participate in the guessing game:
#10 - Google's Favorite University
Apparently, it's the U of...
Arizona? I've seen some strangely biased sitelinks in the past, but this one is a serious enigma. Whatever the sitelinks algo is that triggers results like this, someone needs to turn down the knob.
#9 - Wikipedia as Safe Search
Try a query for guido, and note the "related searches" returns a single result - guido wikipedia (which I have a tough time imagining is a popular secondary query, but OK). Now perform that search - guido wikipedia - and note that the Wikipedia page describing the slang term has been replaced by two others for less derogatory uses of the word (and less exact matches). Bizarre, indeed...
#8 - Jobing for Unrelated Domains
I've purposely moved results 8, 9 and 10 to the top of the screenshot to illustrate result #10 - I'm not sure how Google ascribed that title to a page/domain that has no content, save the word "hello." Looks like several years ago, it contained some related content, but it's been a long time and Google's still giving an empty page front page billing.
#7 - YouTube: The 1st and 2nd Most Important Site on the WWW?
A search for inurl:www inurl:com site:com brings up an interesting listing of important domains, nearly all of which are extremely high PageRank with lots of important links, except...a weird YouTube result, the kind that normally only shows when there's a relevant YouTube video in position #2. Looking at the content and the links, I can't figure out what it's doing there. A conspiracy theorist might suggest it's evidence of Google's favoring of YouTube content, but I'm wondering if the lack of keywords in the query just brings out some odd behavior.
#6 - Sarah Jessica Reminds Us That Google Bombing is Still Alive & Well
The next time someone tells you Google-bombing is dead and gone, show them the obligatory "click here" example, but don't forget about Sex & the City's star, who's apparently "lovely," no matter what the keywords say:
I was hoping they were reading the text in the Flash file, but as that appears to be purposefully inaccessible, I'm guessing they just paid lots of attention to all that "lovely" anchor text. :-)
#5 - Your Site Isn't Here
Search for "Your Site Here" in quotes and find another disturbingly brazen example of pages ranking for a phrase that doesn't appear anywhere in the text. The weird part is not one of the top 10 results contain that keyword phrase:
It isn't until you get to result #13 that you finally find a URL containing a text match. To be honest, I'm surprised the results don't look something more like this - it's as though they automatically append an -allintitle: parameter to your query.
#4 - Gotta Love the Domain Match
I think I must have mistyped "qualit"y when I stumbled across this query for - quali. When I performed the search last month, the #1 result was for www.quali.com, which contains no content, has fewer than 5 followed links pointing to it and last featured content in 2001 (when it was, apparently, for sale).
p.s. Sadly, it's now slipped to result #13 - poor quali.com.
#3 - Where to Get Properly Hitched
I'm a busy man, and I need to rustle up a wedding pronto, so naturally I searched for marriage license seattle, wa.
Having just stopped by the government building recommended by this map, I can say that it is the proper place to acquire a marriage license in Seattle. However, why no address is provided (even clicking the map, you can't get an address) and why the link points to the wrong page (the correct URL is on Seattle.gov or Metrokc.gov) is something I can't quite figure out.
#2 - One of these Domain Queries is Not Like the Others
I'm not sure why Google's giving the query about.com the URL information treatment.
Virtually every other domain query from seomoz.org to cnn.com to porcupineliteraryarts.com gets the standard search results, most of them with sitelink. Since domain queries are also very popular searches, I suspect there's a lot of confused individuals out there wondering why About gets such odd results.
#1 - Math & the Beatles
Now, I'm well aware that a week technically contains seven days, but sometimes, that's just not enough time, particularly when you're in love. Hence, the Beatles gave us Eight Days a Week...but Google's got something strange to say about that:
I believe that's 8 divided by 7, but why it's in the calculator I can't say. I do know that it works for other numbers of days in a week as well - ten, four, eighty-one. Maybe it's just their way of being geeky.
Feel free to share your own bizarre search results in the comments, and for our US readers, welcome back from the Labor Day holiday; I hope you all enjoyed the long weekend.
#10) This doesn't seem too unusual to me. Google has learned that u = university. The universities that have the most inbound links with "u of" in the anchor text are ranked the highest, and apparently the University of Arizona is the most popular choice, so Google gave it sitelinks. #8) It looks like someone accidentally stopped redirecting glendalearenaaz.com to jobingarena.com. If you look at the Internet Archive history for glendalearenaaz.com (used the advanced search to "flag redirects"), you will see that the redirect was in place from June 14th, 2007... up until February of this year (last date recorded). I imagine it was during that time that Google picked up the title "Jobing.com Arena," and Google is still holding that data in its index somewhere, since it hasn't been presented with anything to replace the old title. #1) Danny is right about this one--Google is interpreting "a" to mean "per" or "divided by." So it's reading that query as "eight days divided by 7 days" and returning that unit-less result (the unit "days" cancels out). @Danny... Explain this one: one a day =)
And as soon as Im ready to post this Darren beats me to the punch. hehe.
A quick little analysis between the first two results for “u of”
U of A = with the operators site:arizona.edu inanchor:u of = 103,000 results
U of M = 1,100 results
U of A = with the operators site:arizona.edu inanchor:”u of” = 14,200 results
U of M = 717 results
Knowing that Google places a bit of emphasis on anchor text, and seeing that with the advanced search operator “inanchor” that Google returns 103,000 results compared to 1,100, I think this shows a great deal of information on the importance of anchor text.
If I’m seeing this incorrectly, I would appreciate the feedback.
@RyanTodd... I win! =) Other related ideas I pondered: 1.) Minnesota only has one major university, so people can refer to it (e.g. in anchor text) by a variety of phrases, without worrying about disambiguation. In other words, imagine a college football game--if someone says your team will play against "Minnesota," it's clear that they mean "University of Minnesota." On the other hand, Arizona has 2 major universities: University of Arizona and Arizona State University. This urges people to make a distinction between the two, which usually takes the form of "ASU" or "U of A." Therefore, it seems logical that U of A would have a higher % of total inbound links that include "U of" in the anchor text. 2.) How might stopwords affect this query? In other words, does U of A benefit from the fact that "A" is a stopword (as opposed to U of M, where "M" is NOT a stopword)? Just something to think about.
Good call on distingushing between the two colleges. That would definitely cause people to choose their anchor text a little more carefully.
I've been going rounds with how I think Google handles anchor text. My site is #1 for its brand name (Department of Search) which means that it's outranking government and educational TLDs that contain "department" and "search". My only explanation is that I have a few friends and a couple of other websites that link to me by my full brand name, which includes "of".
Another case in point is the search for the . My guess is that when people link to The Onion, they always include "the" in their anchor text. Otherwise, why would someone just use "onion" as the anchor.
Any thoughts?
For [Department of Search], I would say that a domain name that exactly matches a query will show up as the #1 result, 99% of the time. For [the], I would agree that anchor text made this possible, otherwise I'd assume that Wiki would be #1. Searching for [inanchor:the] supports this as well.
I might disagree for now about the 99%. When I got this domain, I thought it would honestly take a good year to rank for my brand name. It happened in about 4 months. When it happened, I only had 3 inbound links, and approx 7 posts of content. The inbound links all had the full phrase as anchor text.
When typing the querey [department of search] I figured google was ignoring the stop word (of) and focusing on "department" and "search". I would be willing to bet that Google doesn't ignore stop words within anchor text.
On a side note... why didn't Rand search the phrase [university of] ? :-)
All this just goes to show what a difficult job Google has to do to ascertain searcher intent, and how even for a simple universities search there are many variations, and then variations to those variations for special and localised reasons.
Now blow that worldwide, and then into dozens of languages, and my head hurts already.
Interesting discussion.
Darren,
That is the strangest SERP I have ever seen. No idea unless day is a weird unit of frequency...
I'll give you a hint: 1 hz = (1/sec)
yup 1 divided by the number of seconds in a day
Very nice touch on #8 Darren!
Right! And "per" and "divided by" also work in the search box."Times" works too.
I.e. 2 times a week results in 2 times (1 week) = 14 days and 2 times a month results in 2 times (1 month) = 60.8736998 days
By the way, who knows why a month is considered as 30.4368499 days? The most close formula I came up with was (365*3 + 366)/48=30.4375
Because leap years don't always occur every 4 years. If they did, the Earth would orbit the Sun every 365.25 days, which it doesn't. It actually orbits the Sun about every 365.242199 days, which is the number Google is dividing into 12 months. Since we are slightly overestimating the length of a year (by 0.007801 days per year), we make up for it by skipping a leap year 100 years (except for years that are divisible by 400). Simple, right?
From Italy a search for "u+of" returns Alberta university first,second Toronto and third Arizona. This is my first comment here,hello everyone :)
Hello Maing!
See! Even when Toronto isn't first, Canada still owns the top spot - Canadians rule :-D
I love these kinds of things. I always forget to write down weird ones I find. I should start doing that...
I'm the same way on this.
I love watching bloopers but when this stuff happens when I'm searching I am way to driven on the results I want, to write it down.
Thanks Rand
"Lovely" is the name of SJP's perfume, so that one is a bit more logical.
Exactly what I was going to say. I know this because I got it for my gf. She seems to like it. :)
My Two Cents
#8 is clearly an example of a simple 302 page hijack exploit. (https://clsc.net/research/google-302-page-hijack.htm) Appearntly someone doesn't like Jobing.com
#1 Google is multipling 8 by the ratio of days in a week.
eight (days a week) = 8 * (1/7) = 1.1428
The same logic applies to the query "5280 feet a mile" (no quotes). Google is assuming the "a" is the same mathmatically as a backslash. This makes sense because people use "a" to indicate a ratio or fraction.
OK time to get back to work...
Yeah, the Beatle's thing is definitely a quirk of Google's calculator mode. An easier to understand version would be.
But somehow a query like:
...is translated to "8 (days per week)" which Google takes to mean the same thing. No idea why, exactly.
heh, i especially like the beatles stuff. :-)
i'm sure all douglas adams fans know that one: Answer to life, the univ erse and everything
Rand thanks for the nice light hearted post to get us back in the swing of things, after having that long weekend! I lovely the little anonamolies like this that always pop up in google. Maybe you guys should have a contest, user submitted google quirks (guirks?) and the moz community can vote on the best ones.
I've got a couple of faves (warning -- links to my silly little blog):
Google trying to set me up on a date
Google questioning whether my career is a typographical error
Very strange.
The SJP result seems at odds with this site however...
Errr, except for the #1 slot which has the phrase in its URL and as alt attributes and the text of links on that page (unless I'm missing something): the rest of them do seem strange though.
Funny enough, if you search for "about.com" as an exact phrase you get the usual main URL - sitelinks - subdomains - etc pattern. It's only when doing a broad search that you get the weird result. Very interesting!
And re Sarah Jessica Parket - the site is for her perfume "Lovely" and the site has some 879 inbound links, presumably most of them with "lovely" as anchor text. I have no idea who said Google bombs are dead, examples are everywhere.
Excellent post though!
I'm glad to see this commercial term showing product-related results, although it makes me wonder where the Google Shopping results are!?
PS I bought that for my mum's birthday last year, it's quite "lovely" as is Sarah Jessica :)
I love weird Google queries -- one this i haven't figured out yet is sometime son my tracking app (clicky) it shows me site visits derived from some of the most weird/unrelated terms -- yet when i google the search term within the specific engine that delivered the visit>> i do not see the domain anywhere?
afaik google's calculator uses the standard unix units program:
:~$ units 'eight days a week' Definition: 4.1803776e-07 s^2
I like this: https://www.google.com/search?q=number+of+horns+on+a+unicorn
Interesting, "eight days a week"..
I tried to see what it says when i search "fifty days a year", no calculator for this time.
WOW :(
It looks like #3 is pulling heavily from that user map page (https://bit.ly/1kU8N7), is getting the website of https://www.kingcounty.gov/ from the "HOME" on metrokc.gov, and then bonking out on address since the geolocator can't place the pin to street coordinates (the pin is in the middle of a building) but knows it's somewhere in Seattle, WA (with a default of 98101... even though the pin is in 98104).
From the UK a search on google.com for "u+of" brings up Toronto first and Arizona 3rd.
-------------------------------
nuoc hoa | nuoc hoa nam | nuoc hoa nu
This is the kinda of things that you discover on a quite friday afternoon at work....
A search for "1" from Google.com.au shows www.cowboybooks.com.au as the #1 result out of 22,670,000,000... Acid trip 1
A search for "0" shows a strange result for #6
Plasma vs LCD: Which is right for you? - TVs
Update: Which type of flat panel display, plasma or LCD, is right for you? And which will give you more bang for your buck
im sure that there is a lot more weird results that are going to start showing up with personalised search
yeah no strange math here either for "eight days a week".
This reminds me of my good friend who is not particular computer savvy and she was looking for a local sporting good store and typed "dicks" into Google... ROFL...she is still in shock....
Perhaps "U of" is in relation to "U of Arizona" being listed first on many alphabetically ordered lists of top colleges or state schools, and Google giving preference to the first link on a page?
And the "your site here" could possibly be anchor text on many other sites pointing to these service oriented sites, "host your site here" "add your site here", "validate your site here," etc.?
Maybe?
I just checked back with "eight days a week" and result is something different.
Great post, rand.
I'm not sure if this fits the theme of this post, but;
My site is a forum. It has 4+ million posts, crap load of content and ~44,000 registered members. When you search for it on google, using the whole domain name, google does the 'sitelinks' thing, but provides links to the strangest pages... The last one that I blocked from google web tools, was a link to a users' profile that had no posts... That user had no content, at all, on the site but they managed to snake one of my sitelinks... I have never had any of my busiest sections in the sitelinks, usually members' profiles.
Search for one minus two and you get this:
one minus two = negative one
RE: #3
I've noticed this and several other misrepresented or incomplete results when looking at G-Maps listings. There seems to be an abnormally high ratio of confused data in the G-Maps department . . . I'm guessing that integrating all the different DB's isn't the easiest task in the world, but at least they're trying.
On G-Maps results I have seen:
-Blank listings as described above
-Thumbnails of images that produce totally different, random pictures when clicked.
-Local business results showing businesses that are located somewhere else (this is probably resulting from user manipulation, but G should manage it better IMO)
-Flags for buildings/addresses that are in the middle of the water
This was a brilliant blog post - no, really - I can't imagine what's going on over there at the mozplex though that brings such things to light
:-D
Really wonder if the about.com one is for legal reasons at all. seems strange
8 days is 1.1428... weeks. Makes sense to me. That one is not too bad. As for the rest, maybe some google engineers need to take a look at this post after they're done launching Chrome.
Really interesting stuff.
Maybe the about.com example has something to do with the heavy use of subdomains on about.com. I don't know why it would get special treatment though, just throwing the suggestion out there :)
I am guessing that the "about.com" query is being wrongly parsed as if it were info:about.com
Did Google ever have an "about:.." search, or is this an alias for it that gets inadvertently triggered?
Great weird results!
I've found one too. Search for yahoo cartes de voeux (the french for "yahoo greetings") in Google France . The second result points to MSN greetings provided by dromadaire.fr. Who's to blame? The result title is a perfect match for the query but yahoo doesn't appear in the source code...
#3
<address>
From the UK a search on google.com for "u+of" brings up Toronto first and Arizona 3rd.
That's because of all us crazy Canucks over in the UK
:-)
If you just search "university" you're likely to get a better mix of public and private institutions, since (at least in the US) "University of [location]" more often than not is the name of a publicly funded school.
At least from where I'm sitting (about three blocks from the campus), university gives me Harvard -- after Wikipedia, that is.
The Google Bombing algo modifications impacted linking with negative connotations. Modifying the algo to get rid of ALL bombing would butcher all rankings and is basically impossible due to the high weight linking has on rank. There will always be a way to bomb as long as there's a way to play with rankings.