Ash Nallawalla21 February 2009, 1:48 PM
In the last tutorial we discussed the theoretical optimisation of a site. Now, we are ready to get down to details, beginning with keyphrase research.
Search engines are based on a simple principle – a user types a search query using everyday language; therefore, the search engine database and, by implication, the indexed websites should contain such words. In practice, websites are commissioned by people who are ignorant of this principle, which creates work for SEOs.
Keyphrases
It is easy to get caught up in keywords or keyphrases (the terms are synonymous in this context). Do not fall into the trap of counting them, e.g. “you need to optimise for 50 keyphrases”. Focus on solving the business problem for the customer (which for the SEO is an increase in targeted traffic, conversions, and RoI) and the keyphrases will fall into place.
Nevertheless, you can’t ignore keyphrases either. You must know the search terms actually used by searchers when they are looking for your kind of products or services. Then you need to make sure those terms appear on your pages in the right places and the right quantity.
Keyword Resources
There are several resources to make your job easier. Some may or may not suit your website traffic because you might be in an obscure niche for which there isn’t much third-party data available. They are fine for the majority of situations:
Google Keyword ToolGood way to test if a desired keyphrase is being searched and . Popular keyphrases will show lots of impressions.
Overture Tool (defunct) Beware of tools that claim to include its data. Gave a count of similar search terms made the previous month across Yahoo web properties.
ThesaurusFind alternative words for creating content.
SEO toolsNumerous tools for the SEO.
KeywordDiscoveryKeyword research tool (URL of free version)
WordTrackerOnline subscription tool with a basic, free version. Good for US research.
SEO Book Keyword ToolComprehensive keyword research tool.
Example of the Google Keywords Tool in use.Performing the Research
To achieve conversions and to increase traffic, SEO is one low-cost strategy but it can take up to three months or more to bear fruit. Therefore, careful keyword research is necessary, using any tool or method you like:
- Server Logs. The web server log contains every search term used to arrive at your site but it will not list the ones for which your site does not rank high in a search engine – the “you don’t know what you don’t know” dilemma.
- Competitor Sites. If you find that a competitor site ranks high for your desired keyphrase, view its source code to read the Keywords meta tag. This is not totally reliable, because an unoptimised site might contain irrelevant keyphrases in this tag.
- Brainstorming. If you have a small team of SEOs in your company, or marketers, you can get together and brainstorm the likely keywords.
- Customer Surveys. You may wish to construct a survey asking a question such as, “When you look for a camera on the Internet, what would you type in the search engine?” or you could ask a series of questions about features and characteristics they look for and then build a list of likely keywords.
- Using Keyword Tools. There are dozens of free and paid tools that we mentioned earlier.
Primary Keyphrases
Identify the main “money terms” – the ones that will bring in the money. Running a short Google PPC advertising campaign is a good way to see what terms are being searched for, as shown by the number of impressions, which is the number of times the ad is displayed, not necessarily clicked. You need to bid high enough for your ad to show on the first page of results and set an adequate budget, or else your ad will not get the right number of impressions.
For example, my research showed that “seo melbourne” received more search impressions than “melbourne seo”.
If the site has been operating for some time, there might be some data on keywords that convert versus those that bring traffic but do not convert. You should certainly choose the ones that convert, but think about the ones that do not. For many purchase decisions, the first visit or two might be to gather information, then the site might be bookmarked, and you might not associate a subsequent visit with a purchase.
Try to pick the most likely keyphrases, but be realistic. If you jammed a hundred good keyphrases into a 5-page site, you couldn’t possibly get the majority of them to rank. In fact, you’d be lucky to get any to rank because a given page would not have strong content for any single keyphrase.
When you find that there are too many likely keywords to choose from, you can pick those that have less competition online, provided that they are likely to give a return on investment.
Secondary Keyphrases
There are also the less common keyphrases – known as the “long tail” or rarely used keyphrases, you can look for opportunities to optimise content on minor pages for such terms.
Finding “long tail” keyphrases is not easy, because you might find a few obscure ones in your own logs, but you will not find the ones for which you do not rank. The phrases will exist on your site and they could be product names, model numbers, word combinations and unimaginable phrases.
Long Tail keyphrases are to the right.Searchers Are Using Longer Keyphrases
As people get familiar with the use of search engines, they learn to write better search queries, which are usually longer phrases. People are more likely to search for “Kodak XYZ1234 digital camera” than just “digital camera”.
The Buy Cycle/Search Cycle
Searchers go through four stages in what is called the Search Cycle (for an e-commerce site it is the Buy Cycle):
- Research
- Evaluate
- Decide
- Buy
Research Phase
At the research stage, say, for buying a digital camera, the searcher will have an open mind and will search for websites about digital cameras, articles about digital cameras, and so on. These are broad search terms, such as
- Digital camera discussions
- Digital camera forum
- Kodak camera
- Sony camera
- Canon camera, etc
Evaluation Phase
After visiting various websites, the searcher might have made up their mind to narrow the search, say, to a Pentax K200D SLR digital camera. Such a camera has detachable lenses, so there could be searches related to lenses, filters, etc. So the search terms could be:
- Pentax K200D
- Pentax K200D camera
- Pentax K200D forum
- Pentax K200D lenses
- Pentax lenses
- Pentax K200D reviews, etc
Decision Phase
Having confirmed that this is the camera to buy, the searcher will look for a place to buy it. Search terms could be:
- Discount Pentax K200D
- Free shipping Pentax K200D
- Cheap Pentax K200D
- Amazon Pentax K200D
- Duty free Pentax K200D
Using the Researched Keywords
What do you do with the researched keywords? There is no single, “best” way – you may be able to use some methods only for small sites but not for large, database-driven sites that would have thousands of keywords. The marketing manager will probably object to using a word such as “cheap” on the website and might suggest “inexpensive”. If your research shows that most people search for “cheap”, you will need your most convincing voice to get this point across.
The important thing is for you to find the best ones and make sure they are used in the project effectively:
- In the Title tag
- In the Keywords and Description Meta tags
- In headings, body text, Alt tags etc
Keyword Map
You can store the keywords in a spreadsheet and list the pages where you will use them. Example:
- index.html: compare camera prices
- Kodak XYZ1234: digital camera
- Kodak.html: Kodak digital cameras
- Kodak-xyz1234.html : Kodak XYZ1234 digital camera
When editing the pages for content and for the code, you would use the keyphrases appropriately (as mentioned earlier).
For a large, database-fed site, you would list the templates used and how you would pull the data into each template. You would mark up the templates with variables that would place the desired text from each data record in an appropriate location.
Commercial or Informational Intent?
For business sites, it is safe to assume that the owner wants commercial traffic; therefore, search terms should lean towards an eventual purchase rather than information for its own sake. Microsoft has a
tool that can analyse both a keyphrase or URL for transactional or informational intent.
AdSense Preview Tool
The
AdSense Preview Tool for Internet Explorer is available free from Google. It is intended for affiliate marketers, but as SEOs we can use it to determine “what the page looks like to Google’s algorithm”. The intended purpose is for a content publisher to see what ads will show on a given page.
Sometimes the tool cannot determine the subject matter of the page and shows unrelated ads. This tells you that the page needs to be rewritten to emphasise a particular theme. If Google shows ads from our competitors, it has picked up the correct context.
Benchmarks
Once you have determined your keywords and have identified a traffic measurement tool, you are ready to begin. You want to record all facets of the “status quo” as a benchmark so that you can make comparisons over the following weeks and months. You may also want to record the ranking for all desired keyphrases in various search engines, as in this example:
Links| Search Engine | Inbound Links | Pages Indexed | Domain Links |
| Google | 2 | 12 | 2 |
| Live | 4 | 0 | 1 |
| Yahoo! | 0 | 2 | 0 |
RankingKeyphrases
| Google
| Live
| Yahoo
|
Red Widgets
| 130 | 0 | 0 |
Blue Widgets
| 25 | 34 | 67 |
Consider whether you really want to measure rankings beyond a certain value. Few people search past the third page of results, so ranking #40 is as good as #400 – or is it? In a competitive niche, ranking #40 without any optimisation is actually a good sign, as it gives hope of ranking higher. Ranking #400 means there is a long wait and a lot of hard work ahead.
Advanced Keyword Research
Keyword Effectiveness Index
Some people believe in the value of the Keyword Effectiveness Index (KEI) for identifying the most effective words in a campaign. The WordTracker tool places a lot of importance on this value.
KEI = P^2/S = P/S*P
Where P = Popularity (Search Count); S = Search Results
Computing the KEI
First, get the count of times a search has been performed for a given keyword according to
WordTracker (most of its data comes from DogPile, a meta-search engine that searches all the major engines). That is its Popularity value.
The KEI for a word decreases if it becomes competitive. You must search using quote marks around the search term, e.g. “green widgets”.
Note: I don’t bother to look at the KEI for SEO but it is useful if you are looking for topics to write about for an affiliate marketing project.