Pipe symbol in page titles can cause odd search results

Forum Moderators: Robert Charlton & goodroi

Message Too Old, No Replies

Pipe symbol in page titles can cause odd search results

Sgt_Kickaxe

2:41 am on Feb 12, 2012 (gmt 0)

I'm not sure if this is new or old, I just noticed it this evening and cant find anything written about it, but the pipe symbol( | ) often used in page titles to separate the title from company name causes some odd search behavior.

"Big Blue Widgets | Widgetsite"

When searching for that type of exact match(with quotes) in Google it appears that only whatever is BEFORE the pipe is being used to find what you seek and that is often NOT the page you are looking for, not the page from the site you expected anyway.

"Big Blue Widgets Widgetsite"

When searching for this version (with quotes), which is essentially the same WITHOUT the pipe, the article with title "Big Blue Widgets | Widgetsite" may indeed be returned as the article you wanted when it wasn't using the version with the pipe.

Now this doesn't happen 100% of the time but the fact it happens at all suggests the pipe may in fact be a stop, something to consider if your company name is BEFORE the pipe on all of your titles.

lucy24

9:41 am on Feb 12, 2012 (gmt 0)

It's a separator, equivalent to OR. For a simple-but-fascinating illustration, compare search results for

amazon|wikipedia

with results for

amazon wikipedia

Sgt_Kickaxe

2:22 am on Feb 13, 2012 (gmt 0)

Yikes, Wikipedia gets sitelink status for "Amazon.com" without the pipe. Somehow I don't see Amazon appreciating that...

Anyway, using a pipe symbol in your page titles is akin to saying "OR", nice to know.

[edited by: Sgt_Kickaxe at 2:25 am (utc) on Feb 13, 2012]

Fabt

2:24 am on Feb 13, 2012 (gmt 0)

OR Google picks to use what it deems is more relevant from the page title

Sgt_Kickaxe

2:28 am on Feb 13, 2012 (gmt 0)

Webmasterworld Wikipedia returns a Pubcon article on the Wikipedia site. It seems Wikipedia wins the battle of the brands almost every time. You'd think it would return a webmasterworld article talking about wikipedia since that would more likely be what a visitor wanted?

edit: For fun I tried Google Bing and found a 3rd website offering to search them both at the same time, owned!

Anyway, back on topic, the pipe should not be used as a separator correct? It seems so basic that I don't know why I'd never read/heard about that before.

lucy24

3:36 am on Feb 13, 2012 (gmt 0)

Now I'm worried about what would happen if your page title included ? or !

Would the search engine know that you mean "I'm looking for a page called 'Whither?'" rather than "I'm looking for a page called 'Whither' or 'Whithe' or possibly nothing at all" ?

On the other hand: How often do real humans-- as opposed to you and me trawling for our own pages-- ask for something containing a | pipe? Do regular humans even think of the pipe as being part of the page title?

Marshall

4:12 am on Feb 13, 2012 (gmt 0)

Do regular humans even think of the pipe as being part of the page title?

We are regular humans, just further along the evolutionary scale ;)

But seriously, I doubt most "regular" humans even know what the pipe is, let alone it being in a title. The only punctuation marks I ever use in a title are the comma, colon and period.

Marshall

Fabt

5:31 am on Feb 13, 2012 (gmt 0)

The only discrepancy is in SE's placing more importance on the first words in the title compared to subsequent words. Pipe, comma, semi colon, or dashes have nothing to do with SE rankings or results. This is in regards to the <title> tag, not search queries or body content

EvilSaint

6:46 am on Feb 13, 2012 (gmt 0)

Slightly related...I work for a company in Australia and we are using a Pipe Character in the URL as a separator due to the nature of our data which injects other special characters into the dynamically generated URL except for the pipe, hence its usage.

But, we are noticing that Google doesnt appreciate it much because even though all our URLs are fully encoded in the sitemap.xml file, Google still has trouble executing crawls on our site and truncates the URLs that it visits and then is presented with code 500 errors.

I'm still trying to get a way for us to move away from this character.

Anyone else been in a similar situation?

g1smd

8:48 am on Feb 13, 2012 (gmt 0)

Are the pipes in the path part of the URL or in the parameter values?
There's a difference.

lucy24

9:08 am on Feb 13, 2012 (gmt 0)

and then is presented with code 500 errors

Whaa? Getting an URL wrong shouldn't result in 500 errors. 404* at the most ;) A 500 error means your server is seriously upset about something. You should look into it a little more closely.

* This reminds me that GWT "Crawl Errors" is showing pages as 404 "not found" when they are clearly and unambiguously 410 "gone". (Set in htaccess and confirmed by logs.) I do not care for this at all.

EvilSaint

7:18 am on Feb 14, 2012 (gmt 0)

@g1smd, the URL structure is such:
www.somedomainname.com.au/sub-directory/product-name/location/code|code|code|code

@lucy24 the 500 error is being shown on GWT for URLs which Google has decided to re-encode and try and execute and then not getting anything back from our server it is recording it as a 500 error because we dont have a 404 page to resolve to for us urls which do not actually exist.

lucy24

8:54 am on Feb 14, 2012 (gmt 0)

You don't need a 404 page. That's just a piece of paper. So to speak. The 404 error takes place regardless. Same thing for 403. Even if something goes wrong and the server can't display your 403 page, it's still a 403. It doesn't turn into a 500.

I don't believe robots even see the 403 or 404 page. They just note the error and move on.

aakk9999

12:11 pm on Feb 14, 2012 (gmt 0)

The HTTP 500 may be returned if you have unhandled exception in your script. For example, the URL in request is clear enough for the server to forward it to the correct script, but when the script executes, the rest of URL (parameters) cause unhandled exception in DB read/write or something like that and this ends up with HTTP 500.

I agree with Lucy that it should end up with 404 but for 404 the exception has to be handled (rather than left unhandled.)
I saw this happen often with careless programming.

g1smd

8:26 am on Feb 15, 2012 (gmt 0)

www.example.com.au/sub-directory/product-name/location/code|code|code|code

If the extra code pieces are for detailing alternative or upsell products, including that stuff in the URL leads to massive amounts of duplicate content and also to broken pages when related products go out of stock.