You can discover interesting search patterns for your SEO strategy by analyzing n-Grams in big keyword lists. Here is a real life example of how easy and fast it can be done.
This will be short how-to on applying n-Gram analysis on search queries extracted from Google keyword planner. The main goal of this analysis is to better understand patterns in big keyword lists. In my example I will use the monthly search volume and the indexed competition as KPIs. Use our free N-Gram Analyzer to get the exact same output I’m showing in this post.
I created a keyword list in Google keyword planner and imported the CSV data to Google Sheets. This is how my data looks like:
Instead of looking at the complete queries I’m interested in single words (1-Grams). Maybe there are some query patterns that appear over hundreds of queries – none of the single queries would have caught my attention when looking on search volume. By summing up the volume on n-Grams interesting search patterns can be discovered. For running the N-Gram Analysis, copy your table columns to your clipboard.
Then paste the data into our free N-Gram Analyzer tool:
Download the CSV file and import it to Google Sheets or Excel and your result of the n-Gram Analyzer is looking like this:
- Keyword: This is the result of the n-gram transformation: Instead of full queries single words appear.
- Avg. monthly searches / Competition: These 2 columns where part of our file input – they get aggregated by default. In the case of competition we need a small adjustment, which we’ll come to in a second.
- Count: This is the number of queries where the n-Gram appeared.
- AvgCompetition: This is the adjustment I was talking about. To get correct averages I take the aggregated competition numbers and divide it by the “Count” of the queries.
This was just one small example of how you can use n-Gram analysis in real world use cases. It is not only great for discovering interesting search patterns for your SEO strategy – it is also great to identify negative keywords for your PPC accounts. But this is another story for a standalone post…