Keyword Research

(Published Article - This article was the result of a series of questions from the Digital Marketing Institute and my answers,)

1) How important is keyword choice in SEO?

Choosing the best niche keywords is fundamental in SEO.

The majority of websites fail. The #1 reason for a website failing to achieve its objective is that the appropriate keywords have not been targeted.

A niche keyword is currently in range for a specific website to achieve top page positioning on Search Engine Results Pages (SERPS) in general and Google in particular.

Success depends on:

1. The keyword difficulty.
2. The Homepage PageRank of the website.
3. Whether it is the Homepage of the website that is competing.

The best niche keywords for a website are those that will attract the largest number of new targeted visitors.

2) Do you recommend choosing your keywords before writing a page and optimizing during the writing process, or writing the page totally from the user’s point of view and optimizing later?

Each page must be based on a theme that relies on a few related keywords. Once you select your theme, write a nice article about it for your webpage. Write for the user and not the search engines.

The search engines are increasingly capable of monitoring user signals. If users are not impressed by the content or the quality of the presentation, this will be picked up by the search engines as minimal time spent on your website by users.

From an SEO point of view, there are three categories of webpages:

1. Webpages themed on niche keywords.
2. Webpages themed on keywords that are currently out of range for top page positioning on SERPs, but could become niche keywords as the website gradually increases its authority (Homepage PageRank).
3. Webpages themed on keywords with very high keyword difficulty where top page positioning on the SERPs is highly unlikely.

Greatest effort should be given to the choice of keywords for category 1 webpages. The development of the website’s authority is almost entirely dependent on such pages.

For category 2 webpages, reasonable effort should be given to the choice of keywords, so that when the authority of the website increases to the point that these keywords come into range, additional website authority becomes achievable.

For category 3 webpages, do not worry about keywords or SEO, other than to ensure that they function correctly. Thus, if a category webpage has broken links, this will be picked up by the search engines, and the website will be marked down.

Example

Let me give an example of a new website for a car hire business in a small English fictional town that we will call “Hogwarts,” in thecountyofEssex.

A new, recently-indexed website has a Homepage PageRank of 0. This seriously restricts the keywords it can successfully compete for. Even the keywords the site can compete for are restricted to the Homepage, which will receive the #2 boost mentioned in question 5.

Experience has shown that for a completely new website, niche keywords with HPR-KD of no more than 4.3 should be targeted. Internal pages on a new website are unlikely to accomplish top page positioning for any meaningful keyword.

Initially, it is likely that the Homepage could achieve top page positioning for a small town or medium-sized borough (category 1 webpages) related keyword. Niche keywords might be “car hire Hogwarts” or “Hogwarts car hire.”

The keyword difficulty for “car hire Romford” would almost certainly be too high for our new website, but with effort, may become a niche keyword in time (Category 2 webpages).

“Car hire” would have a keyword difficulty that will almost certainly remain out of range for a small to medium-sized business website (Category 3 webpages).

I would emphasise that from an SEO point of view, Category 3 pages are not to be ignored. They will not, in their own right, bring in new visitors; however, they can entertain new visitors and increase the overall positioning of your webpages.

The time spent on your website is an important user signal, regardless of the Category of webpages followed. If our new car hire website’s Homepage has a link to a superb page on “places to visit inEssex,” this could potentially benefit the overall performance of the website.

3) Do you have any keyword choice success stories?

It is difficult to comment on unspecified claims that one big link has upgraded a website’s positioning. There are so many misunderstandings on such issues.

Some are under the illusion that a link from a major website will bring immediate benefits. It is not the “website” that matters. It is the linking webpage that is important.

Let me give the example of the link that I have as a result of my latest article on EzineArticles.com (see question 6). The Homepage PageRank of the directory is 6, but the new page when indexed will have a PR of 0.

The value of the back link (link juice) is the PR of the linking page divided by the number of links on the page. Initially, the link juice will be zero.

As EzineArticles is a popular directory, there is a good chance that my article will be read, acquire some links, and perhaps achieve a PR of 2. There are about 60 links on the page, so even with a PR of 2, the amount of link juice will not be great (2/60). I suspect that the idea of a magic bullet that will shoot you to the top is more science fiction than science.

So why bother with writing articles? There may be no immediate benefit, but every drop helps, and also multiplies.

Nobody learns from scratch how to swing a golf club and wins an Open the next day. Arnold Palmer once jested that the more he practiced, the luckier he became. So it is with SEO. There are no sudden successes – just concerted, steady efforts.

My success story relates to finding the top two factors in the Google algorithm!

I am a doctor who is self-taught in website design and SEO. I started with a website that was basically a set of webpages based on my own patient information leaflets. The website was set up for the benefit of my own patients, and it came as a surprise to find that others were finding it and reading it.

I monitor page positioning with RankTracker. Some of my webpages, on relatively small-print keywords, reached #1 on Google.

Three of my webpages, at various times, dropped from #1 to #2. The obvious question was “What did the page that was now in #1 position have that my now #2 page did not?”

From a medical point of view, the new #1 pages were not as accurate as mine. From an on-page SEO point of view, they did not seem to be high quality. I ran scoring checks with Website Auditor andIBP, and my pages had significantly better scores.

That left off-page optimisation – links from other websites – as the only option. The new #1 webpages had one or two incoming links with minimal link value (SEO SpyGlass) and PageRank 0. Again, my pages had more link juice and PageRank 1 or 2.

As far as I could see, there was no explanation for the new webpages achieving #1 position according to everything I had ever read. I have a research background and aUniversityofLondondoctorate. Here was a fascinating research opportunity.

There was one common element to the three webpages that had pushed three of my #1 webpages to #2: They were all articles published by journalists in different national newspapers’ websites.

It became apparent that it was not the authorities of the individual webpages that were pushing mine down, but the authority of their respective websites. Whilst the PageRanks of the new #1 webpages were all 0, the Homepage PageRanks ranged from 6 to 8.

There were many experts who had claimed that PageRank is irrelevant in SEO because they could quote webpages with PageRank 0 that were on the top page of Google for highly competitive keywords. When I checked these pages out, I confirmed that most still had PR 0 but their Homepage PageRanks were at least 5.

With an Excel sheet, I looked at the PageRanks of the URLs of the top 10 webpages for numerous keywords and their corresponding Homepage PageRanks. Homepage PageRank was clearly the #1 factor as far as Google was concerned; however, it did not explain a few exceptions to the rule, mainly for medium difficulty keywords.

These exceptions to the rule became the next focus for attention. There were some top positioned webpages with Homepage PageRank of 0. With few exceptions, these top page positioned webpages with PageRank 0 for medium keyword difficulty were themselves Homepages.

Google is all about “authority” as indicated by links from other websites. Google is clearly assuming that the Homepage of a website has that website’s greatest authority, and Google is providing a boost in effective PageRank to Homepages that are competing for a keyword (#2 factor in the Google algorithm).

It proved impossible to give a numerical value to this Homepage PageRank boost by manually inputting the data into Excel – it was too time-consuming.

We moved on to producing the V1.0 Keyword SEO Pro program, which put all the required data into a table that could be exported into Excel. It became apparent that the boost was between 4 and 5, but for Homepages with PageRanks of 5 or more, there was no discernable boost.

The V2.0 Keyword SEO Pro adds the Factor 2 boost to the effective Homepage PageRank to a maximum of 5. With experience, I would now recommend that a new website should only target keywords with average adjusted Homepage PageRank keyword difficulty (HPR-KD) of no more than 4.3. Only the Homepage of a newly indexed website can compete for any meaningful keyword.

4) Is keyword optimization dying a slow death as Google and the other search engines advance their algorithms for better, more diversified sets of ranking signals, especially as they look more at the user side of the algorithm with factors such as social signals and user behavioural factors?

As far as Google is concerned, I am convinced that keyword optimization is alive and kicking.

You will know that I delayed my replies to your extremely well-thought out questions until an article I had written was published. I believe that the data presented in that article (published 17^th August 2011) answers this currently topical question (http://ezinearticles.com/?Google-Panda-and-PageRank-Updates:-Top-Two-Ranking-Factors-in-SEO-Unchanged&id=6448046).

The Homepage PageRank keyword difficulty (HPR-KD) on a specific Google domain (Google.com., Google.co.uk etc.) is calculated by finding the Homepage PageRank of each of the top 10 webpages for the keyword including the #2 adjustment. The HPR-KD is the average value of the top 10 webpages for the keyword.

HPR-KD for a set of 1000 keywords was compared before the first Panda update (analysis – 24th January 2011) and again after the recent Panda update (5th July 2011 – analysis – 15th July 2011). The sum of the HPR-KDs for the 1000 keywords in January was 5878 (average 5.9 – range 2.8 to 8.8) and in July it was 5871 (average 5.9 – range 3.3 to 8.0). HPR-KD was unchanged for 121 keywords – 879 keywords changed their HPR-KD. The average change in HPR-KD was 0.36.

It is remarkable that the average HPR-KD for these thousand keywords has remained stable.

It is apparent that the primary determining factor for feasibility for top page positioning remains the Homepage PageRank. It is clear that those webpages that have lost position have been replaced by webpages with similar Homepage PageRanks.

In practice, the range for HPR-KD is surprisingly small. An HPR-KD of 4.0 equates to low keyword difficulty whereas 5.0 is difficult. This is due to the fact that the approximate number of average value backlinks to achieve PageRank of 4 is 2,500, whereas for a PageRank of 5, it is 70,000.

The relationship between PageRank and the number of average value backlinks is exponential: PageRank 6 requires 250,000 such backlinks. A keyword with HPR-KD of 5.5 or more would be in range for a tiny minority of websites.

Google is intent on presenting the best webpages that answer a search query based on merit rather than cheating. Black-hat techniques, such as artificially disseminating endless and valueless spun comments and articles around the web, will have even less and probably zero benefit to those who indulge in such malpractice.

The Panda updates to the Google algorithm that have been with us from March 2011 have undoubtedly changed the face of SEO. Websites that have been artificially promoted by black-hat techniques have suffered, and others that were a little behind are now being lifted into their appropriate positions.

This can be compared to an athlete achieving a silver medal with the aid of “medication.” When they are found out, the one with the bronze medal receives the silver and the one in fourth place receives the bronze. What does not happen is that the losers in the heats get the medals – they are not even in the frame.

5) I am writing a page and I know what I want to write about, but there are many keywords I could use.

How would I go about finding the right keyword from an SEO and user stand point?

The limiting factor is the page title tag. Most experts suggest that Google limits the number of characters in this tag that it will recognise to 65 characters and about 10 words. If your preferred keywords are not in the title tag, top page positioning is unlikely.

Google’s keyword suggestion tool provides a list of appropriate related keywords together with the global and local country volume of searches per month. The results can be downloaded into Excel.

We now have to determine the best niche keywords (see question 1). If you target keywords that are out of range, your webpage will not be found by new visitors, no natural links will be obtained and, at best, your website authority will grow more slowly than would otherwise be possible. You will be chasing rainbows.

Our list of keywords can be analysed by Keyword SEO Pro to determine their HPR-KDs. We must determine the Homepage PageRank of our website. Guidelines are available that indicate the minimum Homepage PageRank required for a webpage on our website to have even a chance of top SERP positioning.

The results of the Keyword SEO Pro analysis can similarly be exported into Excel, and we can combine the data from the Google Keyword Tool together with the HPR-KDS from Keyword SEO Pro. We can now look for the best niche keywords for our website, as we know the keywords that are in range and the number of searches for them.

As previously explained, a Homepage can often be successful when an internal webpage would fail.

6) Where are the most important parts of a page to use your keywords (meta description, h1s, alt tags, image file names, URL, etc)?

Everything is to be gained by placing our niche keywords in all of these places as well as in the body text. Furthermore, if the niche keywords are not in the title tag, top positioning is unlikely. The Page title tag is the top on-page factor and the limitation to 65 characters is always a challenge.

The Page title of the Homepage is the on-page gold dust of a website, and I cringe when I see it abused. ecently, I was invited to offer SEO advice for a major hospital and here is the page title for the Homepage:

<title>Home</title>

Much can be written about page titles, but the one additional point I would make is that for most businesses, there is little point in including a person’s name or the company name in the title tag. These can be placed in places such as the meta description, H1 tags and body text.

If somebody who already knows of you or your company searches for you by name, your website will come up. In SEO, we are mainly interested in being found by potential new visitors and clients.

7) Off the page, where is it important that you use keywords?

Backlinks provide benefits in two ways. They can bring in link juice, which will increase PageRank. Additionally, the anchor text on the linking page indicates to the search engines what others perceive your webpage to be about.

A great deal is written about the anchor text you should seek for your backlinks. With a few exceptions, you have little control over this.

The one place you have a degree of control is the URL. If you have your top keyword in the domain name, and the relevant keywords in the file names of internal webpages, these will show in the links that others place to your website.

I confess that I have not had time or, indeed, interest in pursuing social media from an SEO point of view, with the one exception of using YouTube for my videos.

I have listened to social media experts, and I have yet to be convinced that the time and effort required provide acceptable benefit. I have looked at some of the entries the social media experts have made on Facebook and Twitter. They may have hundreds or even thousands of followers.

From an SEO point of view I am not looking for sheep. Most of the social media links are “nofollow,” so they carry no link juice. The objective seems to be produce large numbers of tweets.

I suspect that the food that these media experts eat or the friends that they visit will not even be of ‘media-ocre’ interest to the search engines.

8) You have a theory that homepage PageRank is, the most powerful algorithmic factor.

However, it has been shown that other factors like Facebook likes/shares, domain names, etc have a higher correlation to high rankings than Homepage PageRank in correlation statistics and research from SEOMoz (the closest we currently have to scientific research on the algorithm); do you still believe Homepage PageRank to be the most influential factor?

First, I am not convinced that SEOMoz suggests that other factors have a higher correlation to high rankings than Homepage PageRank.

As I read it on http://www.seomoz.org/article/search-ranking-factors, “domain level link authority features” are at the top! Surely, as far as Google is concerned, domain level link authority is the Homepage PageRank.

To fully appreciate the importance of Homepage PageRank, you have to understand that while it is undoubtedly the #1 factor, you have to allow for the #2 factor – the boost given to the effective Homepage PageRank when a Homepage is competing for a keyword.

I would emphasise that determination of HPR-KD for any keyword does not require a program to verify my theory. It can be done with a spreadsheet on paper or with Excel. Once you have checked it, and, I trust, confirmed it, you will find that you need a program to do the laborious work for you.

9) Do you think Google will, at some point, move away from keywords and look more at question-type queries and more naturally written queries?

Searches are about keywords, no matter how they are written. As indicated in my answer to question 4, there has been no change to date in the importance of website authority, as identified by Homepage PageRank, even with the arrival of Google Panda.

The only way search engines can identify which webpages are likely to provide the best answers to a search will remain the user signals and, in particular, the authority allocated by web users to websites.

Typically, SERPs contain the page title and a snippet usually, but not always, taken from the meta description.

Although the meta description is no longer thought to have a direct effect on positioning, it will influence the click through rate, which will indirectly influence positioning. The meta keyword tag is no longer a factor for the majority of search engines, including Google.

Panda updates to the Google algorithm place greater emphasis on a variety of user approval signals and decrease the benefits of artificial backlinking.

Panda incorporates machine learning, which means that the program learns on the job. It is figuring out and fine-tuning the weight of user signals indicating approval or disapproval. This results in on-going re-arrangement of positioning of webpages on Google’s SERPs.

Panda examines a number of user signals, including time spent on the website, “bounce rate” (the number of visitors who fail to click on internal website links), and probably returnees.

Even the consequences of the information displayed on results pages are analysed: Positioning on results pages will be changed according to click through rate.

Panda means that Google continues to be influenced by user signals. Primarily, this is the Homepage PageRank. However, if users are not impressed by a particular webpage, as indicated by time spent on it and bounce rate, then that webpage will drop and others with similar Homepage PageRank will replace it. If you have a webpage in position 100 or lower, it is not really in the frame as far as Google is concerned.

10) What are your top 10 tips for choosing the right keywords?

1. Find the best niche keywords.
2. Start with a list of potential keywords and expand them with the Google Keyword Tool.
3. Only target keywords that are currently in range for your webpages. Keyword SEO Pro is the only keyword difficulty tool that measures HPR-KD. It uniquely identifies those keywords that are currently in range, those that will come into range, and those that are unlikely to ever achieve the coveted top positions.
4. Look for related keywords – latent semantic indexing. Google paid a fortune to acquire a program that looks at this. An article is more enjoyable to read if it uses synonyms in preference to endless repetitions of the same words.
5. At all times, give greatest attention to the keywords on your Homepage.
6. Monitor your incoming links for numbers, quality and the keywords in the anchor text – SEO SpyGlass is particularly helpful.
7. Keep an eye out for keywords that are proving to be successful. Google Analytics and Awstats often provide surprises that could influence your efforts.
8. For a new website, aim initially for the ‘low hanging’ fruit difficulty keywords. Matt Cutts, who heads the Google black-hat detection team, refers to a ‘Katamari’ philosophy for website authority development based on the Japanese game which starts with a small object that gradually rolls up ever larger ones.
9. SEO is not a one-off past time. The horizons for a website enlarge with time. As the website increases its authority, more appealing keywords come into range. It is usually preferable to add pages to compete for additional niche keywords than to make vast changes to an existing one.
10. Incorporate your targeted keywords in link-bait. Good link-bait is typically a short list of recommendations – e.g. “10 top tips for choosing the right keywords.” If your readers are impressed by the tips you provide, they will be encouraged to link to you. The outgoing link from them will not initially benefit the site providing the link. However, as with a search engine, if they are offering good links, their readers may in turn link to them as a website that provides quality information.

Original Publication - Digital Marketing Institute