February 3, 2011

Journey to SEO Nirvana | Keyword Research (step 3 of 5)

Now that we got quantification and some basic Search Engine Optimization (SEO) housekeeping out of the way, step three in our journey to SEO nirvana is one of my favorites — keyword research.  Clients I work with often have strong ideas on what words they want to “own” on Google or the other leading search engines.

After years of experimenting and monitoring, I have found there are often too many words that could conceivably be SEO targets for optimizing their site.  The result of keyword research should be a very clear delineation as to which select words are targeted for optimization on a page-by-page basis.

Here’s a magic keyword selection formula . . . all thing’s being equal, you should target keywords that have:

  • High volumes of search traffic (in your geography)
  • Low levels of competition
  • High current ordinal rankings (for your website) on search engines
  • High levels of your website-specific search traffic (e.g. the terms visitors to your website are searching on to find you, and the search terms visitors use on your site itself)
  • High performance in Search Engine Marketing campaigns (e.g., high rates of clicks generated and/or conversions in any Google AdWords, or equivalent, campaigns)

Sounds simple, right?  Let’s walk through each of the above.

1) Search Traffic

I think this is intuitively the easiest factor to understand.  All things being equal, I’d rather optimize around keywords that have high traffic levels.  A few caveats however:

1) First, make sure that the keywords are closely related to your business or service.  Optimizing a wine bar site on security software terms like “antivirus software” is pointless.

2) Second, look at traffic levels for the market served.  Again, if I’m a local business such as a bakery (that doesn’t ship its goodies), then looking at global search volume for bakery-related keywords is pointless — unless substantial numbers of tourists visit my bakery.

3) Thirdly, SEO optimization is a “play-to-win” game, so if I don’t have a hope in hell of showing up high in the search engine rankings, the volume of searches generated by a particular keyword may be pointless. This is why I don’t look exclusively at search volume in targeting keywords, but use it as just one of several variables to consider.

Tools:  There are a number of ways to measure keyword search volumes. My favorite is the Google AdWords keyword tool – I’m most familiar with this one.  Wordtracker is another popular service.  Both services will actually help you identify a portfolio of keywords to consider based on the initial terms you generate and/or a scan of your website, or those of your competitors.

2) Competition

As noted above, all things being equal, I’d like to target keywords that have low levels of competition.  This serves as a balance to the high traffic factor. In general, high traffic keywords tend to have the highest levels of competition – not surprising.  However, this doesn’t always hold – the market is not perfect – and at times you can locate moderately searched keywords that have moderate or even low levels of competition.

Tools:  Again, I use the Google AdWords keyword tool.  It rates the level of competition for each keyword on a scale from 0 to 1.0, with 0 being the lowest level of competition and 1.0 representing the keywords with the highest level of competition.

3) Current Rankings

SEO tends to be a “winner-take-all” game, with the lion’s share of click-throughs for any keyword accruing to those sites in the 1, 2 or 3 ranked positions, followed by the rest of results on page one.  If your site is on page two or lower, the likelihood of click-throughs diminishes significantly.

The basis of taking current rankings into consideration of target keywords is as follows:

It is easier (albeit not “easy”) to move a site that currently shows up on page two of a search results (e.g. an ordinal rank of 11-20) to page one than it is to move up a site that is ranked 100th (e.g. shows up on page ten or eleven).  Adding this into the consideration mix prevents us from blindly focusing on keywords that have high traffic and low levels of competition without regard to how we currently rank.

Tools:  I use Rank Checker plug-in for Firefox.  It is a basic tool that allows you to input a URL and then check the current ranking of up 100 keywords (at a time) on Google and Yahoo.  (Need to confirm if the Yahoo results Rank Checker now returns are ostensibly Bing’s).

4) Site Search (optional)

If you have an existing website and have analytics installed that track the search terms visitors to your site used on search engines in order to find you; and the terms that visitors use to search on your website itself, then those keyword search volumes can be used as another factor in keyword research.  Basically, we want to give greater credence to terms that we know users are already attracted to.

Tools:  Google analytics and Omniture (now owned by Adobe) both can provide this data.  The benefit of Google Analytics is it is free and relatively easy to install.  Omniture is a paid service but has extensive functionality and reporting capabilities.

5) Search Engine Marketing (SEM)

If you are currently running SEM (e.g. Google Adwords or equivalent) campaigns, then they can provide yet another source of insight as to keywords you might wish to target.  A basic way to think about this information is that keywords that generate clicks to your site and/or conversions (e.g. an inquiry, sale, email newsletter sign-up, etc.) are more valuable than those that don’t.

Tools:  Google Adwords has extensive reporting.  If you have been running campaigns for some period of time (e.g. several years), you might want to consider filtering the reporting time period you look at.  Search terminology and associated performance can change over time and many businesses experience seasonality of terminology that affect how well certain keywords perform during certain times of the year.

Pulling It All Together

I’m sure brighter analytical minds than mine could build a finely tuned algorithm that appropriately weights each of the above variables to give you the “right” answer as to which keywords are best.  I tend to use a blunt force, but effective, instrument . . . called a spreadsheet!

Assign each of the variables to a column and then for each keyword give each variable for each keyword a 0-3 rating (or 1-4 if you like that).   Often when you look at the data you will see logical break points (i.e. 10,000 monthly searches for a keyword followed by the next highest volume keyword with 482 monthly searches) or simply break the keywords into quartiles.  If analyzing 400 keywords, give the 100 most attractive keywords in each of the variables (e.g. highest volume, lowest competition, highest current ranking, etc.) 3 points, the next 100 2 points, etc.

Once done use the handy “sum” function to add up all the columns in a total column, then sort by descending values to find your most attractive keywords.

Easy?

A Final Thought

The above approach works very well for a single product or narrowly focused company.  However, many companies have portfolios of products, services or customers/markets served.  In that instance, it may be helpful to add an additional column to your spreadsheet labeled “group” or “category”.  For a security software company, groups might include “virus,” “spam,” “phishing,” “malware,” etc.

This will help you sort the keywords into appropriate targets based on the subject matter of the pages you are attempting to optimize.

in our journey to SEO nirvana is one of my favorites — keyword research.  Clients I work with often have strong ideas on what words they want to “own” on Google or the other leading search engines.
After years of experimenting and monitoring, I have found there are often too many words that could conceivably be SEO targets for optimizing their site.  The result of keyword research should be a very clear delineation as to which select words are targeted for optimization on a page-by-page basis.
Here’s a magic keyword selection formula . . . all thing’s being equal, you should target keywords that have:
High volumes of search traffic (in your geography)
Low levels of competition
High current ordinal rankings (for your website) on search engines
High levels of your website-specific search traffic (e.g. the terms visitors to your website are searching on to find you, and the search terms visitors use on your site itself)
High performance in Search Engine Marketing campaigns (e.g., high rates of clicks generated and/or conversions in any Google AdWords, or equivalent, campaigns)
Sounds simple, right?  Let’s walk through each of the above.
1) Search Traffic
I think this is intuitively the easiest factor to understand.  All things being equal, I’d rather optimize around keywords that have high traffic levels.  A few caveats however:
First, make sure that the keywords are closely related to your business or service.  Optimizing a wine bar site on security software terms like “antivirus software” is pointless.
Second, look at traffic levels for the market served.  Again, if I’m a local business such as a bakery (that doesn’t ship its goodies), then looking at global search volume for bakery-related keywords is pointless — unless substantial numbers of tourists visit my bakery.
Thirdly, SEO optimization is a “play-to-win” game, so if I don’t have a hope in hell of showing up high in the search engine rankings, the volume of searches generated by a particular keyword may be pointless. This is why I don’t look exclusively at search volume in targeting keywords, but use it as just one of several variables to consider.
Tools:  There are a number of ways to measure keyword search volumes. My favorite is the Google AdWords keyword tool – I’m most familiar with this one.  Wordtracker is another popular service.  Both services will actually help you identify a portfolio of keywords to consider based on the initial terms you generate and/or a scan of your website, or those of your competitors.
2) Competition
As noted above, all things being equal, I’d like to target keywords that have low levels of competition.  This serves as a balance to the high traffic factor. In general, high traffic keywords tend to have the highest levels of competition – not surprising.  However, this doesn’t always hold – the market is not perfect – and at times you can locate moderately searched keywords that have moderate or even low levels of competition.
Tools:  Again, I use the Google AdWords keyword tool.  It rates the level of competition for each keyword on a scale from 0 to 1.0, with 0 being the lowest level of competition and 1.0 representing the keywords with the highest level of competition.
3) Current Rankings
SEO tends to be a “winner-take-all” game, with the lion’s share of click-throughs for any keyword accruing to those sites in the 1, 2 or 3 ranked positions, followed by the rest of results on page one.  If your site is on page two or lower, the likelihood of click-throughs diminishes significantly.
The basis of taking current rankings into consideration of target keywords is as follows:
It is easier (albeit not “easy”) to move a site that currently shows up on page two of a search results (e.g. an ordinal rank of 11-20) to page one than it is to move up a site that is ranked 100th (e.g. shows up on page ten or eleven).  Adding this into the consideration mix prevents us from blindly focusing on keywords that have high traffic and low levels of competition without regard to how we currently rank.
Tools:  I use Rank Checker plug-in for Firefox.  It is a basic tool that allows you to input a URL and then check the current ranking of up 100 keywords (at a time) on Google and Yahoo.  (Need to confirm if the Yahoo results Rank Checker now returns are ostensibly Bing’s).
4) Site Search (optional)
If you have an existing website and have analytics installed that track the search terms visitors to your site used on search engines in order to find you; and the terms that visitors use to search on your website itself, then those keyword search volumes can be used as another factor in keyword research.  Basically, we want to give greater credence to terms that we know users are already attracted to.
Tools:  Google analytics and Omniture (now owned by Adobe) both can provide this data.  The benefit of Google Analytics is it is free and relatively easy to install.  Omniture is a paid service but has extensive functionality and reporting capabilities.
5) Search Engine Marketing (SEM)
If you are currently running SEM (e.g. Google Adwords or equivalent) campaigns, then they can provide yet another source of insight as to keywords you might wish to target.  A basic way to think about this information is that keywords that generate clicks to your site and/or conversions (e.g. an inquiry, sale, email newsletter sign-up, etc.) are more valuable than those that don’t.
Tools:  Google Adwords has extensive reporting.  If you have been running campaigns for some period of time (e.g. several years), you might want to consider filtering the reporting time period you look at.  Search terminology and associated performance can change over time and many businesses experience seasonality of terminology that affect how well certain keywords perform during certain times of the year.
Pulling It All Together
I’m sure brighter analytical minds than mine could build a finely tuned algorithm that appropriately weights each of the above variables to give you the “right” answer as to which keywords are best.  I tend to use a blunt force, but effective, instrument . . . called a spreadsheet!
Assign each of the variables to a column and then for each keyword give each variable for each keyword a 0-3 rating (or 1-4 if you like that).   Often when you look at the data you will see logical break points (i.e. 10,000 monthly searches for a keyword followed by the next highest volume keyword with 482 monthly searches) or simply break the keywords into quartiles.  If analyzing 400 keywords, give the 100 most attractive keywords in each of the variables (e.g. highest volume, lowest competition, highest current ranking, etc.) 3 points, the next 100 2 points, etc.
Once done use the handy “sum” function to add up all the columns in a total column, then sort by descending values to find your most attractive keywords.
Easy?
A Final Thought
The above approach works very well for a single product or narrowly focused company.  However, many companies have portfolios of products, services or customers/markets served.  In that instance, it may be helpful to add an additional column to your spreadsheet labeled “group” or “category”.  For a security software company, groups might include “virus,” “spam,” “phishing,” “malware,” etc.
This will help you sort the keywords into appropriate targets based on the subject matter of the pages you are attempting to optimize.

Comments

One Response to “Journey to SEO Nirvana | Keyword Research (step 3 of 5)”

Trackbacks

Check out what others are saying about this post...
  1. [...] We looked at a case study of a security software company I had worked with on improving its search optimization.  Despite being #1 or #2 marketing position in most categories they competing in, from an search engine optimization perspective (SEO) they earned a failing grade.  Only their website titles (which aligned closely to their product names) and inbound links (which likely occurred by happenstance) received marginally passing grades.  All other on-page optimization variables were failing:  keyword selection, keyword density, meta descriptions, H1 and H2 headers.  (For more details on keyword selection, see Journey to SEO Nirvana | Step 2 of 5). [...]



Leave a Comment