A Beginner’s Guide to Keyword Research

If you’re learning about SEO, then you already know that keyword research is a cornerstone of your overall SEO strategy. In fact, when the layperson is asked to describe SEO, keywords are usually one of the first things they talk about! Yet even though it might be somewhat common knowledge that you should perform keyword research, it’s harder to find someone who can actually describe how to perform it.

There’s a reason for that knowledge gap: even in the SEO community, it’s difficult to find a consensus on a standard set of best practices or steps to take as you’re performing your keyword research. If you’re looking for how-to articles, you’re likely to stumble across five articles that are completely different from one another, each filtered through the lens of the writer or agency’s preferred tools and metrics. When it comes to keyword research, everyone has their own “secret sauce.”

This doesn’t speak to some massive failure in communication within the SEO community so much as the level of personalization that keyword research inspires and the grains of salt with which you should take any universalized advice about keyword research. So I hope you have your salt shakers ready, because we’re about to walk you through some universalized keyword research advice! This beginner’s guide to keyword research will provide you with one step-by-step approach to keyword research, and it will share some theory along the way so that you can use the guide as a jumping-off point to develop your own secret sauce.

What is Keyword Research?

Let’s start with the most fundamental question first. What is keyword research, and why is it so important? Keyword research is the process of finding the most relevant, traffic-driving and realistic search terms for your site (as a rule of thumb, 3-5 terms per page). We’ll break each of those three qualities down in just a second, but first, let’s clarify what we mean when we say “search terms.”

Let’s say your ecommerce site sells blinged out dog collars and other fancy dog clothing. Think about your audience, and then think about what they’re likely to search for when they need a fancy new dog collar for their chihuahua. Whatever answer just popped into your head is a search term. People use millions of them every year to find what they need on the internet, and your job is to target the right ones.

We’ll go into more detail about how to determine the right ones later, but for now, let’s break down the three broad traits of a good keyword:

Traffic-driving: Search volume is an important metric when you’re performing keyword research, and SEO strategists often use the rhetorical shorthand “high search volume” when they mean “traffic-driving.” Search volume is an important metric, but it’s not a complete one until you think about the keyword’s ability to drive clicks to your page. For example, a keyword might have high search volume but it leads to an answer that can be shown directly in the SERPs (Search Engine Results Pages), without necessitating a click-through to any website for further information. They keywords are usually short, specific answers to specific questions, like “how old is Tom Hanks?”
Realistic: Also known as “low competition.” Is the competition realistic for your keyword? “Dog collar” might provide you with the highest search volume, but how many other websites are competing for a broad keyword like that, and who are they? Good keyword research uses a variety of metrics to realistically assess the competition.
Relevant: So you’re looking for keywords for your “pink dog collars” page and you notice that “leather dog collars” has an ideal combination of search volume and competition. None of your pink dog collars are made from leather, but some of them are leather-like, so you can use that keyword, right? Wrong. While there may be some crossover between the audiences, someone looking for a leather dog collar is unlikely to be satisfied if the search results turn up pink plastic dog collars instead. And since Google’s goal is to provide the best search results, they won’t prioritize your page for that keyword.

In short, keyword research is a method of discovering which search terms you should care the most about for each page on your site. How we get there is just a tad bit subjective, but it leans on a whole lot of data.

What Keyword Research Is Not

Keyword research is not:

Just about getting traffic to your site. It’s also about looking at the conversion opportunities so that you can make sure your traffic is going to behave the way you want it to once it gets to your site. A keyword that pulls in a bunch of visitors who bounce quickly will be detrimental to your site as a whole, since user behavior is an ever more important metric for Google.
Only useful for SEO. If your keyword research is in-depth, you can glean valuable insights into who your target audience is, how they behave, what they’re looking for today and what they’re likely to look for tomorrow. It will show you what the seasonal trends are and what’s on the horizon. It can predict the products that will become your top-sellers and it can give you ideas that you hadn’t considered before. You’ll learn about the questions your audience asks, the traits they value, and you’ll discover new ideas as you build out your content strategy. Keyword research is a powerful market research tool, and to use it for SEO alone is to make incomplete use of a wealth of information.
An excuse to focus heavily on the success of specific keywords. Here’s where it gets a little confusing. When we perform keyword research, the goal is to select 3-5 keywords per page that have the best shot at bringing in the type of traffic you want. So it makes sense to track how well those keywords are performing, right? Yes and no. The power of those keywords isn’t in the individual keyword themselves; it’s about the “buckets” of keywords that they pull in. Less than 30% of the searches on the web are those “high-value” keywords that we look for when we’re performing our keyword research; the other 70% are long-tail keywords: highly specific and unique searches that are only conducted once or twice a month — or once or twice, ever.

That means measuring success by the rankings of individual keywords isn’t going to provide you with a very accurate picture of how well your keyword research is paying off. For every “bucket” keyword, there are thousands of long-tail searches that the keyword can assist in bringing in, and accurately measuring that has little to do with an individual keyword ranking. In fact, it’s possible for your “bucket” keyword’s ranking to be lackluster, but its longtail results make it one of the most valuable keywords on your site. We’ll cover measuring your results in more detail shortly.

Where to Start: Keyword Topics

So now that we have an overview of what keyword research is, let’s get into how to actually do it. Here’s where everyone’s process starts to look a little different. The one we’re going to outline looks like this:

Brainstorm keyword topics
Research related search terms/expand your list
Cut down your list
Give your list a final cut, ending with 3-5 search terms for your page

The first step is to come up with a list of general keyword topics that may be good ones for the page you’re working on. This is going to turn into your “seed list,” or the list you’ll be expanding into other keyword opportunities. Because it’s just a seed list, you won’t have to worry at this point about whether the keywords are realistic or not; in fact, the broadest keywords often make excellent starts for your seed list.

There are a few different wells of inspiration we can pull from as we’re brainstorming topic ideas:

See what you’re currently ranking for: if your website isn’t brand new and has a fair amount of content, there should be keywords it’s already ranking for. Paid tools like Ahrefs, Moz Pro, or Spyfu are great ways to check in on what’s already ranking; and for a free option, nothing beats Search Console. In fact, prior to starting any keyword research, set up both your Google Analytics and Search Console accounts if you haven’t yet (you can use the same login for both) - it’ll make measuring things much easier down the road!
You’ll want to look at the search terms that are going to the landing page you’re working on, specifically. These don’t just make great seed keywords; they’re also potentially important keywords not to “remove.” If you have one keyword that seems to be pulling a lot of heavy weight, then removing content and re-optimizing the page for a different keyword may do more harm than good. If the page is pulling in some specific long-tail searches, see if you can find a common theme and create a broader “bucket” keyword that encompasses those searches.
Although looking at the keywords for your landing page specifically is a priority, take a look at the rest of the site while you’re there and get an overview of the keywords that are ranking site-wide, and which pages are ranking for them. This can be hugely important in informing not just your keyword research but the architecture of your site: were you planning on using “jeweled dog collars” as a keyword for the page you’re working on, but you can already see that search term is bringing in traffic to a different page? Perhaps it’s time to rethink whether the page you’re working on should address that topic. If another page is appearing relevant for an unexpected search term, take a look at that flag and consider what it might mean for your site architecture.
Check out the competition: Sometimes, your competitors are ranking for keywords that you might not have considered adding to your strategy. Spyfu and Ahrefs are both great tools for checking out the competition; but a manual look at your competitor’s site can also provide you with a lot of information. Pay special attention to Title tags, meta descriptions, and any common themes in the content.
Think about the traits that your audience is most interested in: Adjectives can be a healthy way to seed a keyword list. Is your audience more likely to search for “sparkly dog collars,” “creative dog collars,” “high-end dog collars,” or beautiful dog collars?” While we’re brainstorming, you can add all of these topic ideas to your list.
Consider niches: As you build your list, you’ll want to think in terms of niche topic ideas, or “buckets.” These are the broad topics that your niche target audience is most interested in. It’s helpful to come up with as many of these bucket ideas as you can on your own, because when it’s time to expand your list, every tool out there is somewhat limited in its ability to generate ideas that come from other broad topic categories. Instead, they’ll show you keywords that are closely related to the ones you gave it, usually getting more long-tail and specific.

Expand Your List

Now for the fun part: taking those seed keywords and blowing them up into a huge list of related keywords! There are several ways to do this:

Google it: Topping the list, we have Google’s “autocomplete” feature. Google your keywords and check out the related search terms that Google spins up for you. These are excellent keywords to add to your list, because if Google is suggesting them, it means people are actually making those searches. Check out the above searches related to “fancy dog collars”.
It’s possible that you hadn’t considered including the size or sex of the dog as target keywords, but these are words people actually use to focus their search and they would be valuable keywords to explore. Do you sell fancy dog collars with bows? If not, this is an example of keyword research doubling as market research; perhaps it’s time to start carrying them.
Other autocomplete tools: For autocomplete on steroids, check out Ubersuggest or Answer The Public. These are sites that spin out every related search term for a given topic, rather than just the top ones. Answer The Public is especially helpful when you’re designing a content strategy, because it seeds each list with the basic “Who, what, where, when, why” questions to unearth what people are asking about your topic idea.
Paid keyword research tools: Moz Pro’s Keyword Explorer and Ahrefs are both valuable resources for generating keyword ideas. They take your seed keywords and find closely-related search terms that may have a much better shot at being competitive.
Free keyword research tools: If you’re not ready to commit to a paid tool, there are still some great free resources. In many ways, the Google AdWords Keyword Planner Tool is the closest you’re going to get to mimicking one of the paid keyword generators, but it comes with a major caveat. This tool is built for advertisers, not SEO specialists. That means the ideas it’s giving you are based off of ad performance, not SEO. The difference is often striking, which means the ideas it generates for you will be limited in their ability to show you the complete picture. Wordtracker’s Free Basic Keyword Demand tool does a better job of designing its suggestions for SEO, since it’s pulling keywords from the search results and not just from ad data.

Finally, Google Trends is a valuable tool in showing you the searches that may be missing from the other keyword generating tools due to fluctuations in seasonality. Even the best keyword generators are limited when it comes to showing how search volume changes over time, making Google Trends a powerful ally in capturing a complete picture of search behavior.
Try a couple of rounds: Ready for some idea-ception? Good, because a whole lot of SEO is about generating ideas from generated ideas. Using the above tools of your choice should have produced a lot of related keywords and at least a couple of new “seeds.” Take a look at your full list and create a new seed list to try, and then repeat the process as needed until you’re satisfied that you have a comprehensive list of keyword ideas.

You should now have a sizable list. Not all of the tools we showed you will provide you with the data you need (the autocomplete tools, for example), so grab your entire list and run it through Keyword Planner, Moz Pro, Majestic SEO or Ahrefs (professional SEOs usually use three or more tools of their choice to collect all the data they need, but beginners can lean on Keyword Planner. If you’re looking into paid tools, we really like Moz Pro for its ease of use and its incorporation of difficulty metrics).

Capture all of the data into a spreadsheet (you should have the option of exporting any list as a CSV, but if not, organize it by — at minimum — keyword, search volume and competition or difficulty). If you’re using Keyword Planner, make sure you’ve set the Match Type to [Exact] prior to exporting your data.

Cutting Down Your List: Phase 1

By the end of the “expansion” process, it’s not uncommon to end up with a list of hundreds of possible keyword ideas. So, it’s time to take a look at them and cut all this down. We’re going to cut in three phases: phase 1 is the “eyeballing phase;” phase 2 is the heart of your keyword research, the “quantitative data” phase and phase 3 is what we’ll call the “quality check” phase.

Onward to Phase 1! Depending on who you are, you’ll either find this step fun and easy or tedious and boring. It involves making your large list a bit more manageable prior to the “quantitative data” phase.

First, sort your spreadsheet by search volume and cut out any keywords with 0 monthly searches. “Data not available” or a simple “ -- “ usually denotes a search volume that is too low to calculate, so you can cut these out too. Just use your judgement to make sure it’s not a temporary bug in the tool; a broad keyword like “dog collars” should never have a search volume of 0, so if it’s showing “data not available,” then run your list through a different tool or try again later. This is why it’s helpful to use multiple tools; cross-referencing is important in keyword research, since it relies on tools that at worst can be buggy and at best are using educated guesses to generate their metrics.
Additionally, consider seasonality before you cut; if it’s in the middle of May and you’re looking at a search volume of 0 for a Christmas-related term, that’s because nobody is searching for Christmas stuff in May. Most tools average out the monthly search volume to avoid this problem, but that doesn’t always leave you with the most accurate metrics for the months that matter. Look into the tools you’ve decided to use and learn if or how they adjust for seasonality and other fluctuations.
For More Advanced Readers: Skilled SEO strategists will also use metrics like the amount of traffic your site receives and where the page is in your architecture (the home page? A category page? A specific product? A blog post?) to determine a different number to use as a parameter. Depending on the site and the page, anything under 100 or even 1,000 searches a month might be considered too low-value. For other sites and pages, 10 searches a month is a perfectly acceptable number (highly specific products, blog posts and resource articles are unlikely to be tied to low-volume queries, because those broader searches would have been captured by broader category pages). Similarly, if you’re working on a niche product page and you see a search volume of a million visits per month, it’s almost guaranteed to be too broad/difficult for the page, so skilled SEO strategists may decide to set an upper parameter too.
However, these parameters are often too difficult for beginning SEO strategists to set during Phase 1, because to a certain extent they rely on the experience of having seen a lot of websites to understand what a reasonable parameter is for a particular site, industry and page. So for beginners, we’re just focused on cutting out the 0’s and “no data availables” for now. And we’re only doing this to trim our list down to a more manageable size for this next part: eyeballing!
You should be looking at a slightly smaller list now. But there are decent odds that half of your list is comprised of unusable garbage words. These are words that will never be relevant to the page you’re optimizing, and you can tell right away without needing to take a further dive into the metrics. Rather than wasting time by bringing an unwieldy list into Phase 2, look through your list now and cut any words that don’t make sense.
We’d love to give you a metric to use for this, but when we say you’ll know these words when you see them, we mean it. You know your industry and what you sell better than anyone, and you’ll be able to spot the odd birds right away. For example, here’s just the tiniest sampling from the expanded keyword list for “fancy dog collars," as seen above.

Right away, we can cut “fancy martingale collars” and “gucci dog harness sale” because you don’t sell those brands. It’s likely that “pet leashes and harnesses” can get cut too, or if you sell them, bumped to a new seed list to use for a different category page. What about “diamond dog collars?” If you sell them, it might hit the mark, or it might be another keyword to reserve for a more specific page. If you don’t sell them, it can probably go.

If you’re on the fence, then either leave it for now or Google the keyword to see what comes up. For example, a quick Google search of “dog collars by design” returns websites that sell custom dog collars, so if you don’t sell custom dog collars, then the page you’re working on is unlikely to meet the needs for that search.

When you’re looking, you may find keywords that aren’t right for the specific page but they’d be great seed ideas for a different page. Feel free to start a new tab in your spreadsheet or create a new spreadsheet entirely for these keywords, along with the page for which they’d be good candidates.

It’s also not uncommon to run into keywords that would be a great fit for your site, but you don’t currently have a page that would be relevant for them. It’s lovely when this happens! We call it a “content gap,” and creating a new page on your site that can meet this need will boost your site’s user experience and cast your organic net even wider.

For More Advanced Readers: If you spot two words that are nearly identical (a misspelling or pluralized words), you can often combine the search volumes, average out the other metrics and turn those into a single keyword. This is dicey territory for beginners, though, because of how many factors influence whether it’s safe to consider keywords semantically identical, and because certain tools may have already done this work for you to an extent, creating an overlap in search volume between the two words and making it difficult to combine them.

Cutting Down Your List: Phase 2

We should now have a smaller, more manageable list of keyword ideas. So, it’s time to do the heavy lifting and dig into our data. For this part I’m going to wrap in metrics from Moz Keyword Explorer metrics because it’s hard to find a free substitute for them, but I’ll provide some additional options when we get to those parts. During this phase, you’re aiming to bring your list down to ten keywords or less, so you’re really doing the brunt of your work here.

Because you’re looking for an optimal combination of different metrics, there’s not a secret formula here, and you’ll get better at spotting optimal combinations as you get more familiar with keywords. Here’s what you’re looking for:

Search volume: If you’re going the freebie route, Keyword Planner’s data is good enough to use as a guideline. If you’re going the paid route, continue using the Keyword Planner data, while adding columns for the search volume metrics from your other tools. It’s ideal to compare at least three here, because each tool uses a different method of determining monthly search volume, which means before you make any cuts, you should compare metrics to see if the other tools concur.
If you’re using Keyword Planner, double check that you set the Match Type to [Exact] prior to exporting your data, and then look at the Local Monthly Searches.
As a general rule of thumb, you’re looking for a healthy search volume, but “healthy” is extremely subjective depending on the website and page. It’s best to just compare the search volume across all of the results you’re getting for that specific page, and aim for the ones near the top (without, of course, aiming for a search volume so high that it’s way too broad for the page or otherwise unrealistic).
Remember that the monthly search volume represents the total number of searches, which means every website ranking for that keyword is pulling its own share of the traffic. So, if a keyword has 350 monthly searches, even if you end up ranking #1 for that keyword it’s highly unlikely, if not impossible, that you’ll pull in 350 visitors from that term alone; in fact, the #1 ranking result can usually expect to get 30% of the clicks for any given search.
You’re also competing with paid ads, rich snippets and every other website that’s aiming for those 350 visitors — and even if you end up on page one, there are only so many pieces of the pie. This is another reason why keyword rankings can be a faulty metric to use when you’re measuring your results: the keyword you end up choosing as your best option may bring in a pretty low volume of traffic for that one term, but it can have a sizable pay-off with the “bucket” it creates.
Competition: Once you understand what your high-volume keywords are, you need to look at them in tandem with your competition metrics to find the keywords that hit that sweet spot between search volume and attainability. In an ideal world, you’d find the words with the optimal combination of high search volume and low competition. In reality, the emphasis that you’ll put on competition will vary according to the current traffic your site brings in, the position of the page in the site architecture and your page’s individual ability to be highly competitive for that search term.
In other words, don’t just automatically rule out high-competition terms. Before you cut your high-competition keywords, consider each one individually within the context of the page you’re working on. If it’s the home page, then a high-competition keyword (within reason) might not be the scariest thing in the world, especially if you believe in the ROI that the keyword would produce for your site. You’re going to have to work much harder to rank for a high-competition keyword, but it may be well worth the time investment.
If you’re using Keyword Planner, you can use its Competition metric as your guide. It’s not ideal because it only defines competition broadly — high, medium or low — and it measures against ad performance, not organic. You can also take a look at CPC (cost per click) and assume that higher-CPC words have higher competition. That said, an extremely high CPC term that you believe you’d have a reasonable shot at ranking for organically may produce a fantastic ROI because it could reduce your AdSpend!
If you’re using Moz Keyword Explorer, competition is a whole lot easier to measure. You’ll want to use the Keyword Difficulty and Keyword Opportunity Scores as guideposts, prioritizing Opportunity because it’s a “final” calculation of those other metrics. That said, you’ll need to take every Difficulty and Opportunity score with a grain of salt; people can do their best to quantify subjective and fluctuating metrics, but there’s no substitution for personally understanding your site and audience. Additionally, the Opportunity Score is based on the traffic opportunity of ranking for the keyword, not the business opportunity. A high-opportunity keyword that doesn’t convert wouldn’t be the ideal keyword for a product page.

Cutting Down Your List: Final Phase

And that brings us to the final phase of our keyword research, where we take a step back from all of our data and put on our subjective hats again. You should now have a list of ten-ish keyword options.

Start thinking critically about the ROI that each keyword would produce. The data on a specific keyword might look fantastic, but if it brings in a search audience that won’t ever become customers, then it’s not a worthwhile keyword. So, consider the conversion potential of each keyword, tailoring your definition of a conversion to the goal of the page. For example, a “conversion” on a product page might be a sale, but a “conversion” on a blog post might be moving to other pages on the site to learn more.
Google each keyword. More than any metric, this is the best way to get a tried-and-true look at your competition. Who’s ranking on the first page for that keyword? The second page? If it’s nothing but major stores, you’re going to have a really difficult time competing for the keyword. If it’s nothing but Wikipedia and informational posts and you’re working on a product page, then your prospective keyword’s search intent is off. If your search results pull up a featured snippet (a summary of the information your user is looking for), then they’re potentially receiving the answer to their question without ever needing to visit your site.
Pop on over to Google Trends. What’s the seasonality of your keyword? What’s the overall trend? Is it trending upwards or downwards? When decisions get difficult, Google Trends can be the tie-breaker.

Tools and data are extremely helpful in generating ideas and culling the list, but (so far) only human eyes can truly assess an opportunity. And that part’s all you! So if you thought we could reduce this entire process to a single data-driven system that works for every website every time, well, that won’t ever quite be the case with keyword research. In truth, keyword research is as intuition-based as it is data-based, and the “intuition” part develops with experience. But following this guide is a great way to dig into keyword research so you can start developing that eye. Good luck!