Does Reading Level Matter in SEO?
Posted on Tuesday, November 17th, 2009 by David Waterman
Print This Post
Categories - SEO
Below are a statement (from Ted Ives) and a response (from David Waterman) on the concept of reading level as it relates to SERP rankings. Enjoy!
A statement from Ted (a wise man)
While it’s unclear whether Google utilizes reading level of web pages as an organic ranking factor explicitly, reading level clearly matters if you give it just a little thought. There are various methods of computing the reading level of a page, with pluses and minuses to each, but my favorite is the Flesch-Kincaid grade level, which takes into account the number of words per sentence and the number of syllables per word, and combines these two ratios to come up with an American grade level equivalent. A Flesch-Kincaid grade level of 8 simply means you need to be reading at an average 8th grade education level to understand the document.
Below is an analysis of a particular web page (that will remain nameless) versus the top 10 pages that rank highly for “debt consolidation”.
The analysis shows that for the target web page, its calculated grade level is 5; while high-ranking web pages for the desired term have a grade level of 6 (some of the other measures show a difference of one grade level as well). The obvious takeaway would be to rewrite the content to include longer sentences and longer words, but does grade level really matter?
It absolutely matters because, if the grade level is too high (i.e. the document is too hard to read), users will come to the web page and many of them will immediately leave and search for another document that is easier to read. Conversely, if the document is written at too low a grade level, it is probably “speaking down” to the end user who may see it as not providing enough value for them to accomplish their goal – and will immediately leave it as well after a quick scan. Thus the web page will have a high “bounce rate” relative to its competition – and Google very likely uses that bounce rate as an important factor in organic ranking.
So when you are analyzing your content from an organic ranking perspective, don’t just worry about keyword density and related words – measure your content’s reading level and make sure you are writing content at an appropriate level for your target audience.
Now a response from David (a Waterman)
Two people couldn’t agree more. If I had to put money on it, my bet would be on content reading level contributing to rankability. I’d also put a side bet on the thought that “ideal reading level” (at least according to the search engines) is different from topic to topic.
For example, if you’re writing content on the components of business intelligence and data warehousing, you’d have to assume that the search engines will expect this piece of content to have a high reading level.
However, if you’re writing a piece on Paris Hilton at the MTV Music Awards, you’d have to assume that the reading level would be…well…you get the idea.
In any case, just from a logical standpoint, having reading level as a ranking factor among the search engines just makes sense. It’s a good way to match the right content with the right audience. Of course you could make the argument that not all business intelligence content should have a high reading level (e.g. intro pieces/fundamentals of business intelligence/etc.), but then you’d also have to think that there are probably “trigger” words which might reduce the expected reading level of content written on certain topics…words like “Fundamentals” or “Basics” or “How to”. Of course we don’t know for sure if such triggers, or even if reading level, is one of the top factors of rankability. But it does make you wonder what “trigger” words would increase the expected reading level of an article on Paris Hilton.
Good discussion.
I definitely agree that bounce rate and time on site (page) are considered by Google as important factors.
As these most probably correlate with interest / understanding of on-page content (both of which Google can easily measure), it would be interesting to know the relative Google algorithm weighting of each of these factors against a complete content analysis of the page content (a more difficult endeavor IMHO).
Readability makes sense as a ranking factor – as Grant says above (and the article indicates), it’s going to be linked to bounce rate.
It’s also got to be a fairly good indicator of whether a site’s been keyword stuffing. If you cram a load of keywords into your text, that readability level is going to rise and rise.
I had never thought of this before. Flesch-Kincaid is an easy algorithm for search engines to work into their mega-algorithm. It doesn’t even have to be as complex as different levels 6 vs 7 vs 8 for different topics. If a certain topic requires longer words, for example, all pages competing for related searches will be using those longer words. Since rankings are relative, it would not matter if they were all equally affected by a one-size-fits-all readability algorithm.
Many web pages have very little text, and not necessarily even sentences, so there is no need for the search engines to pay attention to the low end. They could simply add a dampening factor for anything over 9, for example. So readability would be a very simple addition to the overall ranking algorithm.
Even if readability does not factor directly into ranking algorithms, I agree that it will affect bounce rates, and I am thoroughly convinced that bounce rates are part of their algorithms. I simply cannot believe that the search engines would ignore several bounce-related signals that expressly tell them whether the searcher is satisfied with the results they served him. It would be like a convenience store owner turning a blind eye to what people are buying and leaving the rat poison on the display up front and the chocolate bars somewhere on a shelf in the back.
hmmmmmmmm. Now here’s the next logical question (from any black hat SEO, current or recoverying)……CAN a high bounce rate as a result of poor content/bad reading level be overcome by enticing “non text” content. In other words, if I have content that’s obviously in place for the search engines, BUT I have a video or picture(s) that are entertaining to watch and help keep the viewer on the page, could this overcome a high bounce rate as a result of poor content? I’d have to imagine it definitely can (to a certain degree). So in that case, I’d have to think that Google takes a look at everything separately AND in combination. Meaning if you don’t have a high bounce rate BUT your content has an inadequate reading level grade (according to Google), you only got half the exam correct. hmmmmmmmmmmmmmmmmmm.
Great article. Poor content identified by readability tests could drive users away from a webpage. Google could therefore think, “this user has bounced from the page, the content must be poor lets push the page down the rankings” or the content could be measured according to readability metrics built into Google and judged prior to any user feedback. I’d imagine it’s a mix of both.
I’ve just written an article on exactly the same thing, which is a coincidence! Why don’t you measure the Flesch Reading Ease of this page and compare it to mine? Then see who outranks who?!
http://www.advancedinternetmarketing.co.uk/seo-readability
Thanks,
Paul
Oh another thing I picked up from my research, there was a stronger correlation between Flesch Reading Ease and a high ranking than with F/K grade level.
Paul
Magnificent blog post about search engine optimization. I’m honestly startled that that has not been told earlier.
Took me awhile to read all the comments, but I really love the article. It proved to be very helpful to me and I am sure to all the commenters here! It’s always nice when you can not only be informed, but also engaged! I’m sure you had fun writing this article.