3 Alternative Uses for the Keyword Categoriser Tool

Firstly, checked it out, our bulk keyword categoriser.


Don’t let the keyword categorisation tool’s name fool you; it’s incredibly versatile and can help you to carry out tasks other than categorising keywords with incredible results.

This post assumes you already know how to use the categorisation tool.

I personally use the tool regularly to accomplish 3 other regular SEO tasks with greater efficiency:


1) Calculate Branded Vs Non Branded Traffic Accurately and Quickly

Difficulty: Beginner


What is branded website traffic?

Branded traffic includes anyone who visits your website by typing something unique to your company into a search engine. This may include your brand name, a unique product code or name as well as any misspellings or close variants. It demonstrates a user who is familiar with your company or products and wants to find your specific website.


Why Would you want to calculate branded Vs Non Branded Traffic?

Calculating the percentage of your traffic that is branded vs non branded can be important for a number of reasons:

  • If you are running a “brand awareness” campaign over a period of time, it helps you to work out how successful your efforts are. An increasing percentage in branded traffic shows your efforts are successful.
  • Alternatively, your brand might already be well known and you may wish to start ranking for generic terms. Consider an organisation like Dyson – the majority of their traffic appears to be branded:


semrush graph showing branded vs non branded traffic for Dyson

Graph produced from SEMRush data


They may start a campaign wishing to rank for more generic terms and wish to track their success there.

  • As shown in the image above, whilst we can use tools like SEMRush to get a rough split you may A) not want to splash out on paying for another tool and B) want a more accurate calculation based on first party data.
  • Working out your branded Vs Non branded traffic helps you to more accurately predict your click through rates for a given group of keywords which will allow you to forecast better.


How to Calculate Branded Vs Non Branded Traffic

You’ll need access to search console and have at least 3 months data there. Then, we can do the following:

  1. Head to a typo error generator to get some misspellings of your brand name. A good tool I use is this one: https://www.dcode.fr/typing-error-generator. Using the example of Dyson, it’ll give you a number of common misspellings for your brand name:


Misspellings for brand terms

If your brand is known for a specific product(s) it may also be worth adding these terms here too. If you have a few products, consider using a scraper plugin or screaming frog to scrape the names of all your products to plug.

  1. Navigate to the categorisation tool and download the blank data form to paste in the terms, naming them all “branded”:


Keyword categoriser data form


keyword categoriser column 1


  1. Login into search console and download the search queries which have led to clicks over your given time period. If you want to show improvements you may do this month by month. If you were doing this as part of a pitch you may wish do this over a longer time period.
  2. Run the search console queries against your “categorised” list using the categoriser


Branded keywords in Andy's keyword categoriser

Please note: Keywords have been fabricated for the purposes of this blog.


  1. Once you have hit “categorise” and the tool has done its thing, you’ll have a categorised group of keywords. All you need to do know is copy this data into your search console export (in a separate tab) and run a simple vlookup to create a list with branded keywords which also has clicks and impressions. Your finished result should like this:



Branded vs unbranded example


Now all we need to do is make a pivot table and from there we can make graphs or percentages or slice it up however you wish:


Categorised keywords pie chart

And there you have it – a quick and accurate way to calculate your companies branded and unbranded terms. As discussed earlier, it may also be useful to do this by keyword position and using a pivot table to work out a clients true click through rates which you might use for something like forecasting. You’ll end up with something like this:


Branded and unbranded click through rates


2) Deleting (or adding?) Location Based Keywords from your Keyword Research

Difficulty: Beginner


So you’ve created your keyword universe and have around 10,000 keywords you need to filter down. You want to work out the opportunity for a given category but your results are skewed as so many of the queries have “in Manchester” or “in New Delhi” appended to them. You need the keywords without locations on them. How do we mass remove these location based queries?

Lets use an example. You may be doing keyword research around “artificial intelligence” and keyword planner is doing what it does best; giving you a lot of good keywords mixed with lots of irrelevant keywords like “artificial intelligence jobs in India” and “artificial intelligence conference Birmingham” as well as “top fintech companies in California”. You could have added filters to your keyword planner, but there are so many you may as well just export all the keywords and easily filter out the irrelevant ones using the keyword categoriser. Firstly head over to our categorising list library to check if there is already a list of terms you can easily upload to the tool to use as a “filter”.


keyword categorising list screen shot


In this case the “Every Country In The World List” is exactly what I need. So this will be a CSV list of every country in the world that we simply copy and paste into the dataform sheet to get the formula before copying the formula into the keyword categorisation tool as can be seen in the steps below:

Step 1) – Use the country list and paste into one of the category columns on the data form tool to get the formula easily:


Screen shot of data form sheet


Step 2) – Paste the formula’s from the dataform sheet into the keyword categorisation tool:


Keyword categoriser screen shot


Step 3) – Hit “categorise” and you’ll have a sheet where you can easily apply a filter to delete all of the keywords to do with towns and cities:


Location filter


Of course you may want to use the lists to do the opposite and simply categorise against cities, towns and countries rather than using this to delete them.

3) Produce a Basic Customer Sentiment Analysis

Difficulty: Expert


There is a lot evidence to suggest customer sentiment is increasingly important for SEO. That is, what people are saying and writing about your company is not only good for business, but good for SEO; your online reputation matters. See these short snippets taken from Google Quality Raters Guidelines (QRG):


Screenshot of Googles Quality Rater Guidelines

If you are unaware of what Googles QRG is, in essence, every time Google makes changes to their algorithm, they send a document out to their “human testers” for feedback to see if the algo change worked. So these guidelines don’t implicitly tell us what Google’s ranking signals are, but reading what they are asking their users to look for give us a pretty good idea of what they are trying to tweak for.

If you really haven’t heard about Googles QRG and generally about a concept called E-A-T (Expertise, Authority and Trustworthiness) before there are some incredibly comprehensive resources out there which I highly recommend you learn as it’s incredibly important for SEO. Start with this article which you’ll notice features Maria Haynes who’s generally seen as the industry expert on this topic. In fact I’d highly recommend signing up for her newsletter: https://www.mariehaynes.com/ – I am personally a subscriber to the paid version and find the information invaluable as it keeps me up to date on an incredibly fast paced industry.

Back to customer sentiment analysis. So by now you hopefully agree what customers are saying about your business is incredibly important. There are some very expensive tools out there which will scan through the internet and come up with a general analysis for your site – whether the general sentiment is positive, negative or neutral. However, if you wanted a basic version to impress your clients with as a bit of value add or as part of a pitch you can use the categorisation tool with the help of a scraper such as Screaming Frog. 

Step 1) Use your scraping tool to extract all of the reviews on various platforms such as Trust Pilot and combine it with Tweets that mention your brand using a Twitter archiver. Another blog will be written soon on how to use Screaming Frog’s custom extraction to pull Trust Pilot reviews if you are unsure how to do this but for now we are focusing on the keyword categoriser tool.

Step 2) Create your category lists in the data form sheet. In category 1 may be your “positive keywords” including words like “excellent”, “amazing” and “brilliant”. Category 2 may be your negative keywords including “awful”, “terrible”, “slow” etc. And finally category 3 might be neutral so “ok” and “okay” and “wasn’t bad”. I will shortly be making a template data form sheet for you to use which you will find in the categorising list library.

Step 3) Paste the scrape of all the reviews and the twitter archives into the keyword categoriser and categorise them against the positive, negative and neutral keyword lists you’ve made.

Step 4) Visualise the data – you should now be able to create a pivot table of your brands sentiment and produce some sort of pie chart like this:

Pie chart showing sentiment analysis


In this example I didn’t categorise against any neutral keywords. But for bonus points you could even make a word cloud of all the “negative” issues to visually show your client/manager what they should be concentrating on as a business:


Negative sentiment analysis wordcloud


In this example we can see the client in question has slow delivery times and bad customer service. I often find this is a lot more impact than just saying “we have bad reviews”. You could also run your scrape and Twitter archives through a word frequency counter which will pick up how many times a certain phrase is picked up as the word cloud is only really good at showing how often a word rather than a phrase shows up.

And there you have it – 3 alternative uses you may have for the keyword categorisation tool beyond simply categorising keywords.

You may have some other ideas/improvements. I’d love to hear your thoughts in the comments section below.