Will Google Use Spelling & Grammar As Ranking Factors?

by Chris Silver Smith

Google’s Webmaster Help video by @MattCutts about spelling and grammar is quite interesting:

In it, Matt answers the question of whether spelling/grammer matter to them when they evaluate a site’s quality for ranking purposes.

I wrote about this exact thing in “Google Penalty For Low-Quality Writing?” over a year ago, and some commenters thought the concept of Google analyzing text to detect bad grammar and misspelling was too farfetched to believe possible. However, I’ve read some of the books on corpus linguistics, and it has seemed to me that it’s well within the realm of possibility.

Fast-forward to now, and Matt has essentially stated that some within Google have done some work on┬ádetecting bad writing and others have done work in rating the reading levels of text copy. He further states that while it’s not currently one of their ranking signals, the quality of written text correlates so closely to actual rankings that it could make sense to use.

He does acknowledge that there are significant situations where it is difficult to consistently assess the quality of writing. One example he gave was in how it’s challenging to detect what language a page may be written in, and another example was how some pages have portions of text in different languages.

However, it’s very clear that some Googlers have seriously considered using elements of writing quality as potential ranking factors, and if they improve some of the areas where it’s currently challenging for them to properly assess text quality, they might very well add these into the “secret sauce” that PageRank has evolved into.

Just on the language-detection issue alone, I’ve noticed while using Google Translate that it auto-detects the language of text on pages extremely well, and it is good at translating just the foreign language text on pages when I submit webpages that have English and foreign language text mixed on the same page. So, some of these challenges that Matt referred-to are probably moderately trivial where Google is concerned and may be very short-term impediments.

I feel pretty vindicated in the prediction and advice I gave over a year ago, however. The endsum is that Google may not be using spelling and grammar in quality assessments or rankings at the moment, but they very well may add it in the future. And, even if they are not direct ranking factors now, they are likely indirectly having an effect upon rankings because of how these elements affect user interactions with the page in terms of bookmarking, bounce rates, time on the page, linking, and social media endorsements.

(Miriam Ellis, I think you may have to bake me a cake fairly soon! ­čśë )


Tags: , , , , ,

Spammy usernames will be disallowed from posting comments!

2 Responses to “Will Google Use Spelling & Grammar As Ranking Factors?”

  1. Very interesting, I guess this means I should probably go through my site again and make sure everything is copacetic. Thanks for the post.

  2. Lucas says:

    It’s certainly possible to use corpus linguistics to judge the quality of writing to some extent. I recently did my undergraduate dissertation about how examining grammatical structure within a text can give us understanding about its explanatory power (this has implications concerning the quality of the writing). More specifically I looked at whether a policy document used paratactic or hypotactic sentence structures. What this means is that I looked at whether the passages in the document had co-ordinating (para-tactic) clauses such as “the sun was shining and the birds were singing” or subordinating clauses (hypotactic) such “as the birds were singing because the sun was shining”. Passages that have paratactic sentence structures usually have worse explanatory power since they list co-ordinating facts rather than explaining process through uses subordinating clauses such as because. What I’m saying here is that in regard to explanatory power in particular, it is worth being aware of whether a passage uses paratactic or hypotactic clauses.