| 1:57 am on Nov 20, 2009 (gmt 0)|
I asked this specific question of several Google engineers at a PubCon, and their responses were essentially that this kind of machine intelligence was far in the future. It seems clear to a human reader that some content is intensively created, edited, and re-written to generate polished content. But at the scale of the entire web, even sentiment analysis (is this article positive or negative) is not in use.
Now there are features in phrase-based indexing that will catch and penalize some types of artificial content creations. But I'm pretty sure Google will depend on inbound links and the "community of your peers" to supply the quality metric - and of course, if your site begins to rank for big volume query terms, their human evaluators will also be having a look.
| 10:14 pm on Dec 12, 2009 (gmt 0)|
This has always been an issue with webmasters. Finding good quality content at reasonable rates means that you have to sacrifice on quality. After all, it is easy to find a medical professional in India who will write for a few dollars but is not a native English speaker. This comes across in their content, irrespective of how well read they are or how meticulous they are about editing their work. But on the other hand native English speaking medical professionals in a developed country will not waste their time and energy writing for clients.
I believe the day will come where Google will be looking at grammar. Why should poorly written content rank high when it does not give readers much confidence in the quality of information? After all, no reader is going to trust content from a writer who cannot even structure a sentence properly. This ultimately reduces Google's value to the search engine user since they are ranking 'junk'. Google has enough foresight to prevent this in the long run.
Of course, reputable websites use reputable writers and have editors to ensure quality work. But I believe that even a reputable website's individual article will be penalized in the future if the English is not up to scratch. Just my opinion.
| 12:04 am on Dec 13, 2009 (gmt 0)|
It might become a little tricky when you consider the various types of grammar that is used on forums and other community style websites, webmasters don't have much control over that side of things.
| 3:45 am on Dec 13, 2009 (gmt 0)|
Even if possible, it would be too resource-intensive to warrant bothering with it. It would much more logical to factor in bounces and clickthroughs for new vs. repeat visitors (i.e. personalization) to be able to tell if spelling and grammar errors were blatant enough to give a negative user experience.
| 5:38 am on Dec 14, 2009 (gmt 0)|
|After all, it is easy to find a medical professional in India who will write for a few dollars but is not a native English speaker |
There are plenty of native English speakers in India - but they will probably not be the ones looking for low paid work.
@kidder, other things being equal I would rather a search found a forum where most users wrote well than one where they did not.
| 9:09 am on Dec 14, 2009 (gmt 0)|
|Why should poorly written content rank high when it does not give readers much confidence in the quality of information? |
... assuming that most readers can spot poor grammar and/or spelling. Increasingly these days, they can't.
| 4:59 am on Dec 15, 2009 (gmt 0)|
Do you know that there are more English speakers in India than there are anywhere in the US, UK, Australia, Canada, NZ, in fact the rest of the world. I am not from India, but I know they have their own Indian English, same as the US have their own American English. If G* can differentiate between Canadian French and Native French, I am sure it should also be possible for Indian v UK or US English. Now, if the content is in Indian English, G* will rate it as based in India, have you ever wondered why on Alexa and many other traffic analysis services, many US based sites, supposed to be serving US audiences showing the majority of their traffic as from India!
As to grammar and poor content and spelling, the majority of readers don't care as long as the content does make sense, grammar and spelling is a 20th century thing, we are in the 21st and anything goes, we have to accept it!
If the so called respected print media sets a bad example, we'll all learn that bad example. "Man chases cat down street, gets arrested by cops" is a normal headline news you may read on the NYT, Guardian, the Sun or any other newspaper with a large circulation despite the poor grammar, but it's a normal catchy headline and communicates the whole message in a shortened phrase but at the expense of poor quality English!
| 5:19 am on Dec 15, 2009 (gmt 0)|
|How Does Google Deal With Bad Grammar and Low Quality? |
I think with the new personalization it depends on how often you use twitter and who's on your facebook friends list...
| 7:16 am on Dec 15, 2009 (gmt 0)|
And consider some of the gems that will surface through twitter.
| 8:46 am on Dec 15, 2009 (gmt 0)|
if those "buried content pages" aren't garnering in any "actual" inbound links (from sources outside the local domain)...then this is certainly a metric of the value of the content on these pages...(or, for sure, lack of value)
| 3:02 pm on Dec 15, 2009 (gmt 0)|
bad articles attract no links. Manipulative schemes aside, the algorithm should sort this out without too much effort.