Keyword Discovery Review
by Paul Flyer
Keyword Discovery is an online keyword analysis tool. From all looks and appearances it is striving to compete with the ever popular Wordtracker.
How does it compare?
For $10, a user can sign up for a one day trial. This is $2.50 more than Wordtracker.
A Keyword Discovery search covers a larger set of databases than Wordtracker.
Our database contains over 20 billion searches spanning the last 12 months. We compile this information from 37 different sources. These include major international, pay per click, meta and regional search engines.
Their data goes back for 12 months as opposed to just 2 months with Wordtracker. This is very helpful when searching for seasonal keywords like those related to holidays and such.
The search functionality is a little confusing. There is the typical text field for keyword input. Next to this field is a drop down list with various options. I chose “search”. I tried this with multiple keywords. Each and every time nothing was returned. Either their system is still in beta/something is broken/or the feature is not available to one day users. I could not find a statement anywhere on the site indicating any of the above.
The drop down list did offer numerous options for searching Overture data. I chose Overture (United States) and entered my keywords again. Finally some results. The result set initially returns possible related keywords with the number of searches. Users can click the check box next to the keywords and click the Analyze button to provide additional analysis. The analysis phase returns additional related keywords with three additional columns of information: Occurrences (number of pages on which those terms appear i.e. competition), KEI (keyword effectiveness index) this is the same formula that Wordtracker uses, and Predicted Daily (number of potential searches of that word a day).
None of this is really anything new that other tools don’t already provide. There are free tools that do similar things (see Nichebot).
One cool feature is the twelve month graph. Clicking on the number of searches produces a graph that outlines the number of searches over the last year. Users can use this to determine seasonality. I wish that the graph started with the current month on the far right. It would be easier to discern that the data was for the last twelve months. Instead, the graph shows the data starting with January on the left and December on the right. It would be easy to assume since I did this search in June that everything to the left of June was past data and everything to right is predicated data. Having read the documentation I know this isn’t the case, but the visual presentation of the graph does not make that easily discernible.
I also tried the “Related” search functionality. This is supposed to return other potential keywords that are related to one I entered. I chose “paisley socks” as my keyword. The list returned included:
paisley (of course), socks (uh huh), gymboree (huh?!?), nwt (what?), purple, pair and laugh. Not quite what I expected and not very useful. To be fair I tried other keywords and received similar irrelevant results.
Overall, the site gave me that “great idea but not ready for prime time feeling”.
UPDATE: I haven’t had time to re-review Keyword Discovery. If and when I do I will posting a separate post.
Be sure to review our list of recommended web tools.
[...] Keyword Discovery I reviewed this keyword tool before. It was less than impressive. [...]
July 19th, 2005 at 1:55 pm[...] I noticed that Nichebot has recently switched from Wordtracker to Keyword Discovery. I have heard other rumblings about others making a similar switch. I am currently re-reviewing Keyword Discovery and hope to have something posted within the next two weeks. [...]
February 9th, 2006 at 2:15 pm