We tested: How to get an extra 12 % out of your keyword research?
At the moment, Keyword Tool seems to be the best tool, but to be sure, we set up a test and analyzed the suggested keywords with four different tools. This blog post is exceptionally in English because our humble opinion is that also other than Finnish-speaking SEO experts might be interested in this topic (see the Finnish version of the post).
How does Google’s autocomplete work?
According to Google, the autocomplete suggestions are possible search terms. That is why they are relevant for the keyword research which tries to cover as many long tail keywords as possible. Actual searches have been conducted with the suggested keywords or there is content to be found with those keywords on the web.
An algorithm automatically generates the suggestions. It is based on many factors, for example, how popular the search phrases have been. It also reflects the content on the web. Even if Google says that the algorithm functions without any human involvement, Google has manually removed a set of disturbing search terms.
So the autocomplete terms are as weird as the content online is, or people searching for it are, which actually makes this feature more fun. Someone has even been creating poems with autocomplete:
How was the study made?
First we listed our favorite keyword tools. We didn’t try to find or test all of them; we only took in to account the ones we’ve been using. So this is by no means an objective test. For example, we left out ScrapeBox because it clearly was not in this league as it only gave approx. 20 suggestions. We were looking for a powerful and fast tool which easily gives a comprehensive list of suggestions.
The tools that were included in this study were:
We decided to use only the free versions of these tools. However, with Keyword Tool’s paid version it is possible to get more keywords. Key Collector is a stand-alone software (only for Windows though) and the cost for a 1-year subscription is 29$ but by sharing a review it is possible to get a free license.
We chose a seed keyword for the test, both in Finnish and in English. We wanted to test a Finnish word in order to see how the tools manage a language that has a lot of different grammatical forms and suffixes. For example, there are over 2000 different forms for a word “shop”, “kauppa” in Finnish. All forms are theoretically possible, but only a fraction of those are used in the spoken language daily.
We also wanted to include an English keyword in the test, so that we could test the tools’ performance on their home turf, as all the tools have been created outside Finland. We believe that with an English keyword we could get the most out of the tools.
The seed keywords we chose were “autovakuutus” in Finnish and “car insurance” in English. This blog post covers only the results for the English keyword, while the Finnish version deals with the Finnish results. We scraped as many keywords from every tool as it was possible in the simplest way – basically using the tools the way most users would typically do.
We then pulled the keyword lists from the keyword tools to an Excel file and marked the terms which were misspelled or which didn’t include the seed keyword (i.e. “auto insurance” is a related term to “car insurance”, but still doesn’t include the word “car” in it, therefore it was marked as “irrelevant”) as “irrelevant”. We also deleted special characters from the keywords in order to get the search volumes from Google Keyword Planner.
We also checked the search volumes for the keywords. That was done in July 2015 so current search volumes may be different due to the changes in search demands and seasonal factors.
How many suggestions did we get?
We wanted to have a very simple and rough way to evaluate which of the keywords would be useful for closer investigation. That is why we chose to use the search volume of the keyword as the only criteria for the usefulness of a keyword. This doesn’t mean that we think that a company trying to sell car insurances should necessarily rank on all those keywords. Picking the most important keywords to rank on is always a business and company specific decision.
In the chart below you can see the number of relevant keywords which had a search volume of less than 10 (i.e. zero) and those which had a search volume of equal to 10 or more. The length of the bar shows the total number of relevant keyword suggestions.
Keyword Tool gave the biggest number of both irrelevant and relevant keywords, which had a search volume 10 or more per month. Google Keyword Planner gave the greatest number of irrelevant keywords, i.e. keywords which didn’t include our seed keyword (see criteria above) but which had some search volume.
In the next table you can see the amount of unique keywords and the sum of search volumes for those keywords from the different tools. By unique we mean keywords that were only found in one specific tool and nowhere else. For this metric Google Keyword Planner ranks favorably in terms of search volume, because it sums up the search volumes of different forms and variation of the keyword.
On the other hand, in Google Keyword Planner we are never going to see keywords with special characters, which we can see in other tools: l&v car insurance, how to pay $9 for car insurance, one day car insurance 18+ etc. Those little characters can give more clues about users’ intentions.
In the table below on the first row you can see the number of relevant keywords, which don’t have a zero search volume as well as the percentage of the relevant keywords of the whole group of keywords. On the second row, the same information is presented in terms of search volume. For example, 469 unique keywords were found from Key Collector and those make 23 percent of all unique keywords.
In the terms of search volume, Key Collector is not as useful as the other tools, so it cannot be recommended as the only tool to use. In terms of search volume, 83 percent of keywords can be found from Keyword Tool. Keyword Tool really shines in English-language queries: it covers 83 percent of the unique keywords for car insurances, which is almost the same as what Google Keyword Planner covers. In terms of search volume, Keyword Tool is the best tool for English keywords. Übersuggest.org was really poor with both languages: it doesn’t give you enough data to use – definitely not a tool to go with while doing a precise keyword research.
How many extra keywords does each tool give in comparison to Google Keyword Planner? That can be seen in the next table. It seems that Keyword Tool is the best tool measured in number of keywords and in search volume.
So, using all the four tools gives you 442 percent more keywords in English than if using only Google Keyword Planner. But, in terms of search volume, you can increase the summarized search volume of the keywords in your research with “only” 12 percent – which would still be a tremendous improvement in the highly competitive car insurance business.
In some industries this might give you a bigger competitive advantage if you can find more long tail keywords, which might be easy to rank on. So, using all these tools does pay off. In conclusion, we couldn’t find a single key to happiness, but here we found an incentive to use all the tools and with their help become a more hardworking keyword researcher.
If you want to do a deep dive to our analysis, you’ll find the raw data in a Google Spreadsheet – happy analyzing!
Choosing the right keyword tool is not an easy task when there is a lot of variation between the tools. After studying what kind of and how many suggestions it is possible to get from four our favorite keyword tools, our recommendation is to continue using them all. Even if there is a great bunch of irrelevant keywords with no search volume, it is possible to find up to four times more keywords with these tools than sticking just to Google Keyword Planner, and get 12 percent more search volume to capture with your SEO efforts. – We’d say it’s worth it!
If you are interested in this topic, these two recent blog posts are worth reading: Russ Jones on Google Keyword Planner’s dirty secrets and Rand Fishkin on keyword research tools.