In part one, we saw how to create a new market in 3DMarketVision. We created the “ipod” market and ended up with a list of 435 possible keywords. Now, in part two, we shall see how to filter those initial keywords down to a smaller list of relevant keywords.
In 3DMarketVision, filtering keywords can be done using three different approaches. Most of the time, we will use all three approaches in analyzing a market. One way is to gather data about the keywords, such as the number of searches done on that keyword each month, and use that data to weed out useless keywords. This is the approach we will take in the next tutorial.
Another common approach is to look at the spelling or syntax of the keywords and only look at keywords that do or do not contain certain words. We will follow this approach later on in this tutorial.
For now, we will concentrate on a third approach that is unique to 3DMarketVision, using the results of the Latent Semantic Analysis (LSA) that is done on each market. The 3DView Area summarizes the LSA results by putting keywords with similar context together. This results in clusters and towers of keywords such as the ones we can see by looking at the 3DView Area of the “ipod” market.
These towers are groups of keywords that occur in exactly the same websites in our market. If the towers are too tall, that usually means that the websites are near duplicates of each other and that keyword quality suffers because instead of being recommended by independent websites, they are only recommended by copies.
Using this knowledge, it’s easy to look for tall towers among the keywords, and then delete the keywords that make up those towers. For example, the tallest tower in the “ipod” market is about a band playing music to fight AIDS. While commendable, this doesn’t help us analyze this market. To remove this tower of keywords:
1) Click the top keyword “wi-fi music store” so it turns yellow in the 3DView Area
2) Click the Select Tower button to select closely related keywords
3) Click the Delete/UnDelete Keywords button to remove those keywords
After doing this, the number of keywords has been reduced from 435 down to 417. We can now check other towers to see if they are relevant or if they should be deleted in a similar manner, but in the interest of space we shall only delete one tower for this tutorial.
Once we have finished looking at towers, we can further reduce the number of keywords by going the more traditional route and keep just those keywords that contain the string “ipod” somewhere in the keyword.
To do this in 3DMarketVision, click on the Filter Keywords button in the keywords tab area. In the dialog box that appears, look for the “Only show key words that contain one or more of” filter option and enter “ipod” (without quotes) into the text area. Then click OK.
The number of possible keywords will be reduced from 417 down to 137 because only keywords that contain the string “ipod” remain. This is a common and useful way of reducing the number of keywords to be considered.
In 3DMarketVision, keyword filtering options include:
1) strings that must appear in each keyword
2) strings that must NOT appear in each keyword
3) the number of words in each keyword phrase
4) the searches per month of each keyword
5) custom filters based on the user-defined Tag values
Actually, we can filter on any column in the keyword table using custom filters. This is done by selecting a set of keywords, setting their tag value using the Set Tags button, and then using the Filter Keywords button to either include or exclude that tag value.
In this tutorial we have shown how to filter the initial list of keywords using two different approaches, one taking advantage of the Latent Semantic Analysis (LSA) that is done on each market, and one traditional technique based on root words. The LSA technique is unique to 3DMarketVision and opens up a new powerful filtering technique.
