Keywords, or key phrases, (should) form the basis of any SEO campaign. They should provide a roadmap for everything from designing websites to content development, and are the key to understanding your customers. However, starting an SEO campaign based on the wrong research is a recipe for disaster. By disaster, I mean low traffic, low rankings, and potentially a Google penalty.
So, how can you make sure your keyword research is “right”? Traditional keyword research, where important optimisation decisions were made using a vague combination of search volume and competition, is simply no longer going to cut it. Consumer behaviour and the modern purchase journey have become much more complex, so our processes for analysing keywords must evolve as well. In short, we must also be analysing user intent, and placing these queries at the right point in the conversion funnel.
There are plenty of other excellent guides out there that explain how to do keyword research properly, however it is often useful to follow a real life example from start to finish.
In this post, we’ll look at a cat product comparison site as an example, as this happens to be a project I am working on. The general framework I am going to go through for conducting keyword research (or search query analysis), can be applied to any industry.
Before you dive into the various keyword research tools out there, it’s crucial to get a better understanding of your customers, and how they are talking about your products or services online. There are plenty of ways you can start collecting data on your target user personas, whilst also identifying some potentially profitable keywords.
All seasoned SEO’s know that forums can be a gold mine when it comes to finding great keyword ideas. They are essentially live focus groups, and contain an incredibly accurate example of how people talk – and therefore, search.
The easiest way to search for forums related to your niche is to use simple search strings in Google. Here are a few examples I used:
"cat product" + "forum" "best cat products"+ "forums" "cat advice" + "forum" "cat product" + "boards"
The petforums.co.uk site was frequently mentioned in my research, so I thought I’d take a peek.
The first thing to notice is how the forum is divided up into sections or categories. These sections are often niche topics, and can be very useful for when it comes to building your own category structures.
In a number of seconds, I’ve identified 2 (or 4) possible niche topics for my cat product comparison site:
Digging into the threads a little deeper, I find an A-Z of wet food for cats thread, a thread on the best food for neutered kittens, and insight into their love of cardboard boxes. Automatically I’ve got hundreds of keyword ideas, all related to one of the major niche topics on my site, and a better understanding as to the quirks of our feline friends.
Delving into each and every thread to better understand your user personas can get a little time consuming. What if we could automatically scrape relevant forums to find questions your target audience are asking, for free?
Welcome to my first tool suggestion: FAQ Fox. I came across this gem of a tool a few months ago, and have been using it to aid my keyword research ever since. While it’s not a typical keyword research tool, it’s great for understanding real user problems or queries. Let’s see it in action:
The tool lets you search by keyword, and then auto populates a bunch of forum sites for scraping, depending on the category chosen. I have also added in several cat forums that I found during my previous research phase. At a glance we can see a few more good threads for finding potential keyword opportunities (amongst a couple of totally unrelated threads). It is evident from this small amount of research that cat health and food are popular topics amongst the audience.
Now that we have a better understanding of our audience and the various niche topics, we can start generating some initial keyword ideas.
A good place to start is by rolling up your sleeves and trying to think of keyword ideas on your own. It’s typically easier to think of broader topics than long-tail keywords, but it’s good to begin with some brainstorming before digging into all of the research data. Try to think about other niche topics, product names, brand names etc and you should be able to come up with a good list.
I have managed to create a seed list of 70 keyword ideas.
Doing some competitive analysis is the next best place to start. As well as learning about your audience, it is crucial to learn about your competitors. SEMRush is one of the most valuable search marketing tools for this, as it provides highly valuable organic and paid keyword and traffic data.
One of the most valuable tools to aid your keyword research in the SEMRush suite is the paid search keyword data. You can see exactly what keywords your competitors are bidding on, as well as the costs, giving you a good indication of the potential profitability of each keyphrase. Competitor research data doesn’t get much more sexy than that!
Here’s an example of the data I can get for a large price comparison website’s ‘cat product’ related organic ranking and search volume data.
SEMRush allows you to export these keyword lists. So, my next step is to run this tool for various competitors, export the keyword data, and add it straight into my keyword ideas spreadsheet. Due to the nature of a comparison site, researching retailers sites also generated some good keyword ideas.
This brings my seed list up to 1,372 keyword ideas. Note: there may be multiple duplicates at this point.
Both tools are free, and use Google’s Autocomplete to generate hundreds of relevant long-tail keyword ideas. Simply input your base keyword, and the tools will extract Google suggestions based on how often users have searched for a particular term, amongst other factors.
For the sake of this example, I used KeywordTool.io as it’s my personal preference out of the two options. I ran this tool several times using different head terms as my base keywords, for example:
As well as keyword suggestions, KeywordTool.io also generates frequently asked ‘questions’. The result is a very useful set of data for long-tail content development ideas.
This brings my seed list up to 2,730 keyword ideas.
If you want to generate an even larger list of keywords, Keyword Snatcher is the answer to your prayers.
Simply open the tool, enter your base keyword, and click on ‘Get Suggestions’. You might need to wait for some time, but this tool will return thousands of keywords from your single seed term. If you don’t want to wait too long, you can stop the tool at any point, at which point you’ll already have generated a useful list of terms.
For this search, I’m going to focus on the ‘comparison’ feature of the site and use ‘compare cat’ as my seed term. After 5 minutes, the tool has already generated over 1,900 keywords. I’m going to stop it there, copy the list and paste it straight into my keyword ideas spreadsheet. Other keywords I can use for my seed term are, for example, ‘best cat’, ‘cheap cat’ etc.
This brings my seed list up to 4,625 keywords.
Keyword Spy is a great tool for getting a quick lay of the land. With access to PPC ad data, related keywords, and competitor data, it’s almost a definite one for the keyword research toolbox!
I like to round up my keyword idea generation phase with this tool as it has a ‘misspell’ feature, which provides keyword suggestions based on common misspellings. Unfortunately in the case of my example, there weren’t too many useful misspelling suggestions.
This leaves my seed list at a total of 4,625 keywords.
If the site has been around for a while and you have historical data from your Analytics or Search Console accounts, then this would be one of your final stages before de-duplicating the list and running your keyword ideas through Google’s Keyword Planner to get monthly search volume and competitive data.
The more established your site is, the more data you will have to play with, and the more potential keywords you will find. In this example, the cat product comparison site is still in its development phase so we do not have access to this useful data.
Before I take my list and extract search volume data, I am going to run the final stage of my keyword idea generation phase through the Google Keyword Planner ‘search for new keywords’ option.
As you can see, Keyword Planner has suggested some very useful ad group ideas, which I can also use as ideas to build my site architecture and category structure. For the purpose of generating keyword ideas, I am simply going to click on the ‘Keyword Ideas’ tab and download the 801 results generated from my seed term ‘cat products’.
This brings my final seed list up to 5,426 keywords.
Don’t be overwhelmed. During the the next phase of our keyword research process we will remove thousands of keywords from our master list by using a few simple tactics.
The first thing we want to do with our keyword ideas list is remove any duplicates we may have generated throughout the research phase. You can do this in Excel by simply selecting the data range, clicking on the ‘Data’ tab, and selecting ‘Remove Duplicates’ from the Data Tools group.
In this example, Excel has found and removed 418 duplicates, leaving a master list of 5,009 keywords.
Before moving onto the next step, we also want to remove any special characters as Keyword Planner doesn’t process results for these.
For this step we will go back to Google’s Keyword Planner. While various tools can provide you with this kind of information, it’s important to run your whole list of keywords through the same tool to get the most accurate data. Google’s tool is understandably the most powerful.
Google’s Keyword Planner has a limit on the number of keywords it will return, so we are going to break up our list into chunks of 800 and repeat the process for each keyword chunk.
There are various possible paths within Keyword Planner, however for this purpose we are going to select ‘Get search volume data and trends’.
I downloaded the results and then consolidated them into a single Excel sheet, leaving me with 4,943 keywords to analyse.
I instantly removed any keyword on the list with an estimated monthly search volume less than 100. I then browsed through the remaining terms to make sure there weren’t any irrelevant terms that might have been picked up during the keyword generation phases. If found, these were also removed. This included anything to do with ‘dogs’ or other animals which are not going to be the focus of a cat product site (duh!).
This brings my final list to 529, with a great selection of long-tail keywords.
The next step, and arguably the most important, is to group the keywords into “buckets” according to their user intent. The 4 buckets I use are:
This process can be done manually, where you go through each keyword and tag them appropriately. Or, you can automate the whole process by using an Excel macro. If you don’t know any Visual Basics, or don’t particularly want to learn, then you can buy Nick Eubanks ready-made search intent macro-enabled Excel template. Instructions on how to use it here.
With a few tweaks to the tagging criteria, you can auto populate your keyword research spreadsheet with the searchers inferred intent. Now we can start understanding the semantics behind our list of keywords, which will aid our analysis in choosing our target keywords and mapping these to different content targets in the purchase cycle.
You’ll end up with something that looks like this:
You may find that some of your terms could apply to multiple categories. In this case, you need to fully understand how your product and service offering will affect your tagging, as well as the general semantic nature of the term, before assigning user intent.
Due to the nature of a comparison / review site, I am particularly interested in terms which have been tagged in the ‘Commercial Investigation’ intent group, and even the ‘Informational’ group. These happens to be the majority of my terms.
You can further segment your list into categories that represents the “funnel” stage the keyword belongs to. For example, brand terms may be attributed to “lower funnel” and research or “buy” related terms to “upper funnel”. For now, I am simply going to further segment my list into ‘micro-categories’, so that I can identify any potential holes in my keyword research from a product category perspective. As a product comparison site, this is going to make up the bulk of my site architecture.
Analysing your keywords for the highest growth potential and maximum impact is naturally a pretty manual process. It requires human input to review keywords according to certain criteria, and from here we can assign an aggregate priority score, effectively choosing our “Best” keywords. The criteria we are going to evaluate are:
For this criteria we are simply going to use the data collected from Google’s Keyword Planner. As experienced SEO’s, we all know that this can be inaccurate to say the least, however it gives you a good comparison of search volumes between different keywords.
While assessing competition, we need to estimate how difficult it will be to rank each page of our site for a given keyword. We are going to look at competition from a number of different angles.
Things I am going to pay close attention to:
This particular report gives me a great indication of the commercial interest of this keyword, as well as some useful contextual suggestions for search terms closely related to mine.
This is a great tactic to get an immediate sense of the organic competitiveness of the term without having to actually trawl through the SERPs. We can see straight away that the majority of the top 10 results for ‘best cat food’ are informational based sites, with one entry from an actual retailer. While this is the kind of result I want to see, as this will be the nature of my site, it does indicate that competition is going to be stiff.
We can also assess the level of difficulty for seizing competitor’s rankings with a particular keyword. This ‘Keyword Difficulty’ score will take into account how powerful the other pages are, how many links they have, and how powerful these links are etc.
Now we can start assigning numerical scores (from 0 to 1) to the organic ‘Competition’ for each keyword based on these three factors. I then add the score into my keyword matrix spreadsheet for easier analysis later on.
As we all know, all SERPs are not created equal. Different queries will provide different results, some with lots of ads, some with a shopping feed, and some with the top result containing site links etc. Ranking #1 on each of these results pages is going to result in a different click-through-rate (CTR), so we need to individually assess the competition and opportunity for each keyword.
We also need to understand the opportunity for our own business. Regardless of whether we are ranking #1 or not, different keywords and going to have a different value of our business, and will largely be based on the types of conversions.
For this phase, I assigned a score from 1-5, depending on how valuable the keyword is for the business. A lower score indicates that the term was more general, and may not be targeting our main user persona. A higher score indicates that there is a high level of commercial or purchase intent, and that the user may be at the point of converting directly from the SERPs.
From these three “scores” we can use a very simple evaluation model, which will help us understand the potential SEO value of each keyword in terms of cost vs revenue.
This is by no means a perfect equation, however is does help assess the low hanging fruit, and the growth potential of certain keywords:
(Search Volume ) x (Opportunity) x (1- (Competition)) = SEO Priority Score
The process of evaluating keywords is both an art and a science. There are a lot of different ways you can evaluate your keywords for SEO opportunity, with most SEO’s having their own process that works for them. My keyword research process is one that I have refined over time, but also one which is tweaked depending on the nature of the business / topic I am analysing. The idea behind this process is to collect as much data as possible, in order to start making informed decisions about your keyword targeting and topical optimisation efforts.
I hope you have found the article useful. Please do share your thoughts in the comments, as well as any tips you might have for improving the process.