How to Do Keyword Research for SEO

How to Do Keyword Research for SEO (Without Wasting Time on the Wrong Keywords)

By a digital marketer who’s watched AI Overviews quietly eat into organic clicks

Hey, if you’re doing SEO for your own site or that of a client, I’m going to show you how to do it in a practical manner. Not with a lot of theory. Rather, the kind that enables you to know exactly what to enter into a keyword tool the moment tomorrow morning arrives.

In this article

1. Why Keyword Research Is Still Important
2. Keyword Discovery With Seed Keywords
3. Local Keywords
4. Keyword Analysis and Targeting
5. Optimization and Tracking
6. Prompt Research
7. FAQs
8. Conclusion

1. Why Keyword Research Is Still Important Even Though AIOs Are Everywhere

Rankings can remain the same even if your traffic is actually falling – it’s not just in your mind, because AIOs (AI Overviews) are responding to the query before anyone gets to your website. This means that keyword research for SEO has not disappeared, it’s merely taken on a new role now: identifying keywords that still generate a click, not just one that secures a ranking.

In order to target any keyword you must meet three requirements at the same time—search volume, keyword difficulty, and search intent. If you fail to meet any one of them then the content you produce will either be stuff that no one searches for or content that achieves a high ranking but doesn’t result in conversions.

Get This Right Before Anything Else

• Check search volume, but don’t dismiss low numbers if intent is strong
• Match keyword difficulty to your domain authority and backlink equity, not your ambition
• Confirm search intent by checking what’s actually ranking on the SERP

Image Source

2. For keyword discovery, begin with seed keywords rather than making guesses.

All keyword strategies start with seed keywords, which are the general terms that describe the actual activities you carry out. For example, if your seed keyword is ‘hiking boots’, and you use it in a tool, you won’t merely obtain volume figures; you’ll get long-tail keywords such as ‘waterproof hiking boots for wide feet’ — keywords that involve lower competition, clearer buyer intent, and are more in line with the way people actually speak to AI tools when they ask questions rather than typing in fragments.

Keyword discovery involves first creating a broad list and then narrowing it down.

3. Do not omit local keywords

When you’re serving a particular city, using local keywords is not optional. ‘Digital marketing agency near me’ is quite different from ‘digital marketing agency’ since Google includes your Google Business Profile together with your content when responding to that query. A single term at the city level is still insufficient; you also need to include neighbourhood and service-specific variations.

4. The analysis of keywords and the process of targeting: Reducing the list

When you have a long list, keyword analysis is the stage at which you work out which items are worth going for. You should sort the keywords by volume and difficulty, then look directly at the SERP to see what format is achieving rankings and whether the term is producing an AIO. In such a case, it becomes more difficult to predict the CTR and click-through rate, since people might get their answer without having to click.

From that point, keyword targeting involves assigning the correct term to the appropriate page and keeping an eye out for keyword cannibalization, which is when two of your own pages are quietly competing for the same term and as a result split your ranking signal.

Where Most People Get This Wrong

• Mapping five pages to one keyword instead of spreading intent across your site
• Ignoring AIO presence when deciding whether a keyword is worth the effort
• Skipping the manual SERP check and trusting the tool’s difficulty score blindly

5. Optimization and Tracking: The Technical Aspect

Keyword optimisation is the point at which structure and content come together—by including your main keyword in both the title tag and the H1, incorporating it naturally into the meta description, and having it appear in at least one alt text. Structured data assists search engines, and is becoming increasingly useful to AI systems, in understanding your page well enough to boost your visibility with AI tools such as ChatGPT and Perplexity.

Previously, keyword tracking involved checking positions on a daily basis; it now also includes monitoring CTR trends, since a steady rank coupled with a decline in clicks generally indicates that an AIO has taken over that query.

6. People are overlooking the prompt research stage.

Traditional keyword research is based on short, typed queries whereas prompt research takes into account the way people actually speak to AI tools—using complete sentences together with accompanying context. For example, as soon as someone types ’email marketing tools’ into an AI chat rather than Google, it becomes ‘what’s the best email marketing tool for a small Shopify store’. It is by building up genuine topical authority on a given subject that you end up being cited.

Frequently Asked Questions

What does keyword research mean in the context of SEO?
It involves identifying and assessing the search terms that your audience uses, taking into account search volume, keyword difficulty, and search intent, in order that you may focus on terms which you can actually rank for and convert on.

Does the use of AIOs mean that keyword research is no longer important?
Not at all—it means the target has changed. Now you’re having to decide whether a keyword causes an AIO to be triggered and whether achieving a ranking for that keyword still results in a click.

What is the difference between keyword research and prompt research?
Keyword research involves mapping out short, typed search terms, while prompt research involves mapping the longer, more conversational questions that people pose to AI tools, questions which often indicate different content angles for the same topic.

How many keywords should a single page target?
A single main keyword, together with a number of closely related long-tail keywords and different expressions of search intent. Trying to cover one page with too many unrelated terms results in cannibalization rather than in ranking strength.

Conclusion

Before writing anything, carrying out keyword research for the purpose of SEO wasn’t a one-off thing; it’s a continuous process involving discovery, analysis, targeting, optimization, and tracking, with prompt research and AI visibility now being carried out as the first step in this process. If you get the basic elements right, then both Google and the AI tools based on it will see a reason to direct traffic to your website.

If you’re on your own with your keyword strategy and don’t know where to begin, then leave your seed keyword in the comments and I’ll give you the first three long-tail variations worth pursuing.

Further reading: Neil Patel’s guide to keyword research.