The beauty of debates is that you get 2 individuals in relative fields that have conducted research & came up with different results for a collaborative answer, Thus setting what some may call a Standard. Our Keyword Vs. Density testing was initiated to try & find Google algorithms KeyWord Red Flag Vs. Optimal KeyWord Usage that may benefit an Overall SEO campaign. The Testing: 1st doorway attempt: In my personal research I submitted 3 doorway pages that were content rich with a 7%-9% specific key word density. Granted after 7-14 days these pages ended up in the top 10 (1st page of google). 7-14 days later they were bumped from the index entirely. 2nd Doorway attempt: This time I bumped up the keyword density to a 13%-15% for all 3 pages & made sure the pages keywords reflected a variation of my main site content (Trying to trick the engines this time into thinking they were simply additional pages to the main site theme, yet focusing on individual product lines). Again same result, indexed within 7-14 days & dropped after an additional 7-14 days. 3rd Doorway Attempt: at this point I have learned a little from the first 2 attempts. I focused on keeping my keyword density under 5% & over 3%, Continued using an added target phrase as in the 2nd attempt targeting a specific product. I also added alt & title html Tags to both the hyperlinks & images. These pages indexed within 7-14 days & 2 of these pages still remain on the first page of google for there search terms 2 months later. (Note: my experience with this may be unique, as we know everyones results may vary, However my personal Opinion on the matter is that somewhere between 5%-7% density a spamming red flag comes up in the mathematical algorithms.) Alt Title Tags: As far as the alt & title tags go: the alt tag simply provides an alternative text if an image does not appear, while the Title tag tells what the link/page in reference is about in a short sentence (Normally 5-20 words from what I have seen on the average). The debate on whether Image tags such as Alt/Title actually help has been going on for a long time. Myself, once again (Personal Opinion) leans towards Relevance simply based on how a mathematical algorithm works. (Note: I have read a lot of articles on the subject and found different theories which motivated our own research. I must encourage anyone reading this post to always conduct your own research "If possible". Basing research off an article you read or worse an article you read that they wrote about an article they read is all he said she said.) Ruling "OUT" inbound link support: I only have 1 link pointing to my research / test doorway pages (From my main site footer) & multiple outbound links in the form of navigation. So there is not really much inbound link support to the doorway itself. Ok, I'll post one of my product targeted test doorways as an example along with the keyword phrase: (Please note that this link was placed merely for supporting reference to this Testing debate) http://www.scorpionagency.com/html/usmc_template/usmc_website.html <~This doorway targets a USMC Military Website Template with the Key phrase being "USMC TEMPLATE". Now granted that search phrase only seems to be registering 5-15 hits per day (Its a test page) with a focus on keyword density between 3%-5% along with image & hyperlink alt/title tags to analyze the affects of overall ranking. The page is #1 in Google for the term. Unless the Authoritative presence of my root URL is directly influencing the doorway with one inbound link, I'm not sure that the inbound link theory plays a part for this particular doorway. (Of course I could be wrong). Now the fact that it is a sub-page from the root index along with supporting outbound links back to the root & various sub-index pages may play a roll in the equation (Again its speculative). In Conclusion: While going from failure to success may seem like unsupported evidence to some, the fact remains that the testing in the end produced Evident positive results (Unmeasurable or otherwise). So far the 3%-5% seems to be pretty safe, I am in the process of starting a 5.01% - 6.99% test to push the limits (Seeing that the 7%-9% was bumped from the index). I am open to any additional research anyone has done them self in this area (With the exception of he said she said scenarios).
hi, great research, i would like to point out about image alt tag, with this regards i have research so many time and i m confident enough to say that it help a lot go get ranking. i have used alt tag in image and no other text on that particular page and u cant believe i got ranking for that alt tag phrase.
Actually, I agree with you. I ran into a similar situation. Some of my sub-pages that have free avatars on them (Utilizing the Alt attributes) - The Google Algorithm seems to have changed again, thus putting some of my image saturated pages in top 1-4 positions (1st Page of Google). This may be a fluke (As we know they do happen), however so far for the last 20+ days the positions are solid. (Not much competing for the terms of course, maybe a 5-10 per day search term at best). In any case I would have to say that a little more weight has been shifted to image alt/title attributes in this situation. The pages them self may have a total of 1 paragraph of text (If that). It supports the Image Alt/Title attribute theory though...
Hey there, It was about time somebody did some testing like this, rather than just assuming google does this or that, +REP Can I ask what program you used to test keyword density? I have used several and they give me different results. Sincerely, Travis Walters
Agreed, I have used several as well & each have different results, as far as my Paid program I use keyworddiscovery.com . When I was researching free ones (Prior to purchasing one), I found http://www.seotoolset.com/cgi-bin/checktraffic.cgi to be quite useful in the area of longtail daily traffic averages. LongTails are Ideal for Doorways by the way .....
Interesting research Can I ask what website or tool you used to check your keyword density %? I found a good tool before but didn't bookmark it
Sure, actually for that one, you can click my top sig link - Scorpion Agency - then select Free SEO tools from the left navigation.. I use that one.. Hope it helps.
Thanks for that I've looked at my site, there's a bit too much duplication of some words there! From your work, with regards to titles, if you wanted to target (for example) yellow clothes, would you set up your title to say: yellow jumpers, yellow hats, yellow coats. (using the keyphrase yellow three times, 50% density), or would you go for yellow jumpers,hats,coats (one in four words 25% density). Im a bit concerned that my titles could be holding the site back a little
Stuffing a title tag could possibly be construed as Spam.. I would play it safe, you have a whole page of content & the keyword tag for your Keywords & key Phrases. If yellow is the color focus on that page then I might go with: "Your one stop shop for Yellow Jumpers, Hats, and Coats" You figure some people will be reading it in the browser head as well, might as well add some flare to it. Remember, you First optimize for Humans & then for engines
Actually, If you take this same testing scenario & add a few more basic SEO fundamentals along with some SEM campaigns, You can pretty much Optimize your whole website from the ground Up. (In a basic Fundamental way of course) I am doing a bit more research in this area Now & have released a few Short 2-3 paragraph tests I did on Image Alt/attribute Testing along with Forum Sig Vs. IBL testing... I'll keep you posted when i release any more..
Having looked at the various keyword Density, I've found my site to be this: (example single keyword could be yellow, variation = hat,jumper,shoes,t-shirt) single keyword 11.2% keyword variation 1: 4.2% keyword variation 2: 3.6% keyword variation 3: 3.3% keyword variation 4: 3.0% I'm happy with the key phrases (two word), but could the singular key phrase be seen as spam?
From what I gathered, as long a Keyword (Singular) is under 5.0% it raised NO red flags.... Thus having 1-2-3-4 keywords all at say 4.9% each & of course each one being one word in a Phrase like (1)Yellow(4.9%)-(2)Hat(4.9%)-(3)Jumper(4.9%)-(4)Shoes(4.9%) .. you can use any variation for your keywords & keyword Phrase, the key is that Each word in the phrase is as close to 5.0% as possible without of course taking the whole content out of it's professional essence as to read like Gibberish to the viewer. (My pages dropped around 7.0%, so there may be some leeway between 5.0% & 7.0% somewhere, just not sure where yet Until I do more testing to narrow it down) Using this formula seems to be working great for me - especially with long tail key phrases - Of course the more words you try to optimize for & utilize the content density formula to aid in weight relevancy, the harder it will be to keep the page professional rather than the overly stuffed gibberish...... So use your better judgment...
what if in my site i have only 50 words in each page, but i need to repeat the main key word at least 7 times to explain what i m teaching?
For all of you KW Density junkies, I would strongly advise loking into areas such as; Phrase Based Indexing and Retrieval Probabilistic Latent Semantic Indexing Latent Dirichlet Allocation Hidden Topic Markov Models..... Point being that search engineers have long moved away from simplistic models such as KW Density in establishing relevance and scoring (ranking). Any research into such simplistic models can be faulty... without considering the application of above type methods, you can be getting a false positive in your research. Of most interest is the patents Google put out at the beginning of the year on the phrase based stuff. Here are a bunch of Phrase Based Indexing and Retrieval resources.... 2c
Interesting, so we have a debater finally.. Ok, in light of my research & supporting documents to my testing & results. Can you provide your own testing to show otherwise? Or do i simply read an article with no circumstantial backing? The reason I ask, is because Googles algorithms change all the time & the ONLY people that truly know what an algorithm calculates is Google them self. As far as articles go, well, Anyone with a computer can write one, that doesn't make them 100% accurate either. (Nothing in this world is 100% accurate) I think that an article with documented testing showing the results + supporting evidence (Documents) would be more viable to me than merely an article. Do you happen to know where I may be able to find an article of such nature that will also have results in the SERP's to support it so i can compare my research further to it? It would be a big help, Thanks.... P.S. - as far as false positives - the testing site still remains in first spot with NO alterations - so it's now been 3 months since the originating test, the Only difference is that there are now 2 IBL's showing for that URL - Other than that, I think 3 months is a bit extensive for a fluke.....