As I read her email I could literally feel my blood pressure rising. She’d heard me speak at a webinar I did for Wordtracker about ecommerce copywriting where I said keyword density hadn’t been a factor in SEO copywriting success for years. The lump in my throat got bigger as Zoe (not her real name) explained why she thought the myth about keyword density simply wouldn’t die.
“Keyword density is going to remain a hot (contentious) topic. I just read an article in the “New Yorker” yesterday about the new AOL CEO: “Can Tim Armstrong save AOL?” Apparently AOL is going to put greater focus on being content providers. Here’s an excerpt from page 36:
‘The writing, too, is often designed to appeal more to search engines than to readers. In the list of “contributor resources” for Seed, the most prominent category is for “search engine optimization”–S.E.O.–the process of packing stories with words that will make them appear higher in the list of results that Google and Bing display when users search for terms related to the subject. Seed links to guidelines that instruct writers to pay attention to what is called “keyword density”: the number of times that certain phrases appear in a story as a percentage of total words in a piece. If you’re writing a story on herbal tea, you should use that phrase early and often.’
“So, while I’ve read articles by plenty of respected SEO experts who insist they’ve tested various keyword density models and it doesn’t correlate with returns, I have to say I’ve read at least as many articles like this that still bang the keyword density drum. Well you can see how the mixed messages can be frustrating.”
“Writing often designed to appeal more to search engines?” “Packing stories with words?” Arrgg! Give me a break! Talk about old school. Keyword density has not been a valid measure of SEO copywriting success in probably 8-10 years now.
- Do you need to include keyphrases in your copy? Yes.
- Do you need to “pack” your copy with keywords? No.
- Does your content need to appeal more to search engines than people. Absolutely not!
Yet, dreadfully, Zoe is right about one thing. There are still plenty of so-called experts out there that will swear to you copy must be written to a certain keyword density percentage. They’ll vow that this is the only way to write search engine optimized copy. To those who believe this, I say:
Oh Yeah? Prove It!
Have you ever tested it? Or are you just blindly following this outdated myth that refuses to die?
I can prove that keyword density is not an issue. Can you prove – quantifiably show me in a measurable form – that copy must have a certain keyword density to rank high? I’m sure you’ve written pages that have a 2%, 5% or even 10% keyword density ratio, but what happens if you remove some of those phrases from the copy? Does the ranking drop? Not in my experience.
In fact, clients have hired me to rewrite their previously awful-sounding copy to be more natural. While the former copy would not be called keyword stuffed, it did not flow very well at all. Rewriting it without so many keyphrase mentions not only improved conversions, but also *increased* rankings.
When writing SEO copy for my clients, I never calculate keyword density and the pages rank consistently well.
From as far back as 2006, Matt Cutts (Google’s Antispam Chief) and other officials have stated that keyword density is a non-issue. Here are just a few quotes from Matt and Google.
2006: “I’d recommend thinking more about words and variants (the “long-tail”) and thinking less about keyword density or repeating phrases.” — Matt Cutts
http://www.mattcutts.com/blog/seo-advice-writing-useful-articles-that-readers-will-love/
2008: “Keyword Density: Not really a factor. Yes keyword should be present but density is not important. Include the keyword but make writing sound natural.” — Matt Cutts
http://www.mattcutts.com/blog/join-the-webmaster-chat-today/
2009: “As long as I’ve been at Google, keyword density has not been a core factor … in either the main site text, title tag or any of the other associated tags…” — Adam Lasnik speaking at Search Masters ’09
http://www.youtube.com/watch?v=RhknZUEueKc (About 3:00 into video)
2010: “‘Keyword stuffing’ refers to the practice of loading a webpage with keywords in an attempt to manipulate a site’s ranking in Google’s search results. Filling pages with keywords results in a negative user experience, and can harm your site’s ranking. Focus on creating useful, information-rich content that uses keywords appropriately and in context.” — Google Webmaster Central
http://www.google.com/support/webmasters/bin/answer.py?answer=66358
Please, PLEASE don’t just follow along with what the majority of people online are repeating. If you read interviews from AOL saying they instruct writers to use keyword density and you also read blog posts from reliable sources telling you not to subscribe to keyword density ratios, do your own testing. Find out for yourself who’s telling the truth.
Remember what your mother used to ask you: “If your best friend jumped off a 100-foot cliff, would you do it, too?” Honestly, whether we’re talking about SEO copywriting or not, following the crowd is usually the kiss of death. Keyword density is no exception.
© 2011 Karon Thackston, All Rights Reserved
.
So true. The notion of keyword density will probably haunt SEO forever.
On another note, but slightly related. What is your experience when it comes to the number of times a keyword or variations of that keyword appears across a whole website?
I have seen some pretty clear evidence in relatively competitive areas that this is an important factor. Especially on more generic keywords for which the frontpage would be the natural landing page.
I have heard rumors of such things, Bernt. Never done any testing on my own, however. I do know for a fact that innerlinking the pages of your site using keyword-rich anchor text within the body text gives a big boost. Maybe that’s actually what’s happening with the site-wide issue. It may not actually be that the site itself is ranking. It’s actually not supposed to be that way. Each page should stand on its own. However, if you’re innerlinking with anchor text, that may give the same or similar result in the SERPs. Like I said, haven’t tested it myself.
“While the former copy was not keyword stuffed, it did not flow very well at all. Rewriting it without so many keyphrase mentions not only improved conversions, but also *increased* rankings.”
Did you mean, “While the former copy was keyword stuffed, it did not flow very well at all.”?
Hi FM,
No, I meant that the copy was not keyword stuffed. It did have a lot of keyphrases in it, but it wasn’t just filled to the brim with them either. Still, the copy did not flow well with the use of as many search terms. It had too many, but wasn’t crammed full.
Would a machine-learning-based study designed to replicate Google’s algorithm, that correctly predicted 8 out of 10 of Google’s top 10 SERP entries about 80% of the time, was done by a bunch of Ph.D’s at Northwestern and Purdue, and was published at a recent academic conference, constitute proof?
The study is referenced in this posting, and it found that Keyword Density was around 1/3 as important as PageRank, not as important as keyword in Title or Meta-Description, but slightly more important than having the keyword in an H2 tag. The study was pre-Caffeine but is likely still quite relevant:
http://www.coconutheadphones.com/does-domain-agematter/
Hi Ted,
Thanks for stopping by. After reading the study fairly carefully (I didn’t read every single word), I really didn’t see how this constitutes proof that keyword density is a major issue. Maybe I’m missing something If so, please feel free to point me in the right direction. One thing I did notice was statements such as “Google imposes negative bias toward blogs.” That sort of took me by surprise, because Google loves blogs. A post on a blog will rank highly within a matter of an hour while a traditional web page might take days or weeks. Anyway, open to hearing about anything I might have missed in the study.
Karon,
Your point about blogs makes a lot of sense because Caffeine had a lot of changes around real-time results etc., I think you’re right, that particular finding is probably not correct anymore.
The weightings are kind of buried away in the paper – if you look at the graph in figure 4b with the three lines on it – look at the 3rd iteration line, the one with the “x’s” on the data points…what that essentially is a graph of, is the “weighting” of each factor that they extracted from their system, after they trained the system to emulate Google.
If you read off the weightings on the Y axis, PageRank (“PR”) is weighted 1.0 (so if your PageRank is 4, then it’s 1.0 x 4).
Then whether the keyword is in the Hostname (“HOST”) is weighted .8 (so that would be either .8 times zero or .8 times 1) and so on.
The weighting for Keyword Density (“DENS”) looks like it’s around .3. The authors were a little vague as to what scale they used for keyword density, I’m not sure if it would be .3 times .05 for instance if your pages density was 5%, or if they scaled it somehow, but the weightings are interesting just from a relative standpoint.
Essentially the graph is an illustration of a 17-term regression equation (AX + BY + CZ + …..) that tells you what a pages ranking score would be. Then if you ran all the pages through that and compared their scores to each other it would tell you which order they would rank in.
– Ted
Yeah, it’s the vagueness that makes me wonder. I don’t find that what they are presenting clearly states that keyword density is a must. This could simply be coming from natural keyword use within the text. I certainly haven’t tried to build my own Google-like search engine, but I (and *many* of my colleagues) have tracked pages and sites over time and gotten them ranked repeatedly with no concern about keyword density at all. Another thing (un-keyword-density related) that I find very curious is that this college could built a search engine that supposedly emulated Google’s results so accurately, but yet Microsoft, Yahoo! and countless other professionals cannot? Things that make you go “mmmm?”
I think you’re talking about the “Words That Work” reports. You mean the ones that go into what words and/or phrases are most persuasive for different people in different industries/fields?
They are at http://www.copywritingcourse.com/wordswork.html
No, haven’t used Optimizely. (Funny name.) I’ll have to check it out.
Hey Karon,
What is that resource that breaks down the words different professionals respond to? CEOs. ARchitects, etc.?
Thanks, Suzanne
p.s. I like your test: see if rank drops without so many keywords. Have you all checked out “Optimizely?” It seems a little easier than Google Optimizer.
Suz