Choosing keywords for SEO isn't just about finding a phrase with a lot of searches and repeating it several times in an article.
Well-done keyword research helps you understand how people express their needs, what questions they ask, and what kind of content they expect to find when they use a search engine.
That information can then be used to decide which topics to develop, which queries belong on the same page, and which content can provide a useful answer without relying on keyword density, repetition formulas, or supposed difficulty based purely on how long a query is.
If you're still building the foundation of the project, also check our guide on how to build a website step by step from scratch.
What is a keyword in SEO?
In SEO, "keyword" is usually used as an editorial reference for a topic, phrase, or set of searches we want to understand. A search query, on the other hand, is the actual expression a person types or uses when searching.
The same page can appear for many related queries even if you planned it around a single main topic. That's why it doesn't make sense to treat every linguistic variation as if it needed its own URL.
The priority is understanding the user's need
Before choosing a keyword, ask what the person is trying to accomplish. They may want to learn a concept, solve a problem, follow a tutorial, compare alternatives, find a tool, check a price, or take an action.
| Query | Likely need | Possible content |
|---|---|---|
| what is SEO | Understanding the concept | Introductory guide |
| SEO mistakes new website | Spotting problems | Mistakes and fixes |
| how to choose keywords | Learning a process | Step-by-step tutorial |
| keyword research tools | Finding tools | Comparison or guide |
This classification is just a guideline. It's also worth checking the actual search results before assuming which format best satisfies a need.
Step 1: define your site's focus
Not every query with demand should become an article. Google recommends creating content primarily for a real audience and keeping a clear purpose or focus for the site.
Before researching an idea, ask yourself:
- Does the topic fit my audience?
- Is it related to the site's purpose?
- Do I have enough knowledge, experience, or sources to develop it?
- Would the page be useful even if the visitor arrived directly and not from Google?
Step 2: build an initial list of needs and queries
Start from a broad topic and gather questions, common problems, technical terms, beginner-level phrasing, synonyms, and variations.
For example, within "SEO for new websites" you might find needs like:
- what is SEO;
- how to choose keywords;
- how to get organic traffic;
- how to check whether a page is indexed;
- what SEO mistakes a new site should avoid.
At this stage you don't need to pick a "winning keyword." First identify which questions are genuinely different from each other.
Step 3: review the search results
Search for some of the queries you're researching and look at what types of pages show up, what intent they seem to answer, what information keeps repeating, and what could be explained better.
The goal isn't to copy the structure of the top results. It's to understand the context of the search and decide whether you can create a useful, reliable, differentiated answer.
Results can change over time and vary by context, location, device, or other circumstances. A manual search is a useful signal, not a permanent snapshot of Google.
Step 4: treat demand data as estimates
Research tools can estimate the interest around certain terms. That information helps you compare opportunities, but it doesn't guarantee your page will get that number of visits.
Actual traffic will depend on whether the page appears, its position, the type of result, the country, the device, the intent, and the alternatives shown alongside it.
How to interpret Google Keyword Planner
Google Keyword Planner is designed mainly for planning Google Ads campaigns. It lets you discover terms, check estimated monthly searches, and analyze advertising data.
Its competition metric refers to the relative presence of advertisers for a keyword within Google Ads. It shouldn't be read as an official difficulty score for ranking organically in Google Search.
You can use Keyword Planner as an additional source of ideas and demand, but combine it with relevance, intent, and analysis of the organic results.
Step 5: use Google Trends to understand relative interest
Google Trends can help you observe changes in interest over time, compare terms, and spot differences between regions.
Its values are normalized on a scale from 0 to 100. So a value of 100 doesn't mean 100 searches; it represents the point of highest relative popularity within the selected comparison.
Trends is especially useful for spotting seasonality, shifts in interest, and geographic differences that a monthly average could hide.
Step 6: prioritize relevance over volume
A query with fewer searches can be far more valuable to your site than a popular topic that has nothing to do with its audience.
| Query | Hypothetical demand | Relevance for an SEO site |
|---|---|---|
| how to do SEO on a new website | Lower | Very high |
| viral entertainment topic | Much higher | Very low |
Choosing the second topic purely for volume can attract an audience that doesn't match the project's editorial purpose.
Step 7: don't turn long-tail into an automatic formula
The term long tail is usually used for more specific or less frequent queries. These searches can express very concrete needs, but a long phrase isn't automatically easy to rank for, low-competition, profitable, or right for a new site.
| Query | Level of specificity |
|---|---|
| SEO | Very general |
| SEO for websites | More specific |
| how to check the SEO of a new website | Very specific |
The more specific query may better express what the user needs, but you still need to check what results are competing for that need.
Step 8: group queries that share the same intent
You don't need to create a page for every variant of a phrase. For example:
- how to choose keywords;
- how to find keywords for SEO;
- how to research keywords for articles.
If they essentially represent the same need, one thorough guide can be more useful than three nearly duplicate pages.
Check existing content before opening another URL
If you already have a page that answers the question, consider updating it, expanding it, clearly differentiating the new intent, or consolidating overlapping content.
This helps keep an understandable editorial architecture and avoids creating multiple URLs competing to answer practically the same thing.
Step 9: pick a primary query as an editorial reference
You can select a main phrase to keep the page's topic clear, but there's no mandatory field in Google where you have to declare a "main keyword" for organic ranking.
Use it as a reference to check:
- whether the title correctly describes the topic;
- whether the introduction explains what the reader will learn;
- whether the subheadings cover the necessary questions;
- whether the page satisfies the need you're trying to answer.
There's no ideal keyword density
Google treats stuffing a page with keywords, or repeating words and phrases unnaturally to try to manipulate rankings, as spam. You don't need to hit a specific repetition percentage.
The main topic should show up naturally in the title, content, and other descriptive elements when it makes sense, but you don't need to force the same exact phrase into every H2, paragraph, alt text, or link.
Use related vocabulary when it helps explain the topic
A guide about keyword research will likely use terms like query, intent, demand, Search Console, and search results because they're needed to explain the process, not because Google requires some mandatory semantic list.
Step 10: use Search Console after publishing
Once Google starts showing your pages, Search Console's performance report can show you the queries, clicks, impressions, CTR, pages, countries, and devices tied to your site's performance in Google Search.
These real queries can reveal opportunities you hadn't anticipated. However, an impression for a phrase doesn't mean you should automatically insert it into the article.
Before changing the page, ask yourself:
- Is the query genuinely related?
- Does the user need information that's still missing?
- Would adding that explanation actually improve the content?
Prioritize opportunities with a simple matrix
You don't need to reduce everything to a single score. You can review each idea qualitatively:
| Criterion | Question |
|---|---|
| Relevance | Does it belong to the site's focus? |
| Intent | Do I understand what the user needs? |
| Demand | Are there reasonable signals of interest? |
| Capability | Can I create a useful, reliable answer? |
| Differentiation | Does it add something another page on the site doesn't cover? |
| Maintenance | Will I be able to keep the information up to date? |
A full keyword research example
Imagine you want to create content about traffic for new websites. You start from the broad topic website traffic and find queries like:
- how to get website traffic;
- how to get visits for a new site;
- how to increase organic traffic;
- traffic sources for a website.
The next step isn't to automatically create four articles. First you review the intent behind each query, the available results, what content you already have, and whether you can develop a sufficiently useful answer.
You might then conclude that several queries belong under one main guide, for example "How to Get Traffic for a New Website", using the others as questions or subsections within it.
Common mistakes when choosing keywords
- Choosing purely by volume: a popular query can fall outside the site's purpose.
- Confusing Google Ads competition with SEO difficulty: these are different concepts in different contexts.
- Assuming long-tail is always easy: length alone doesn't determine organic competition.
- Creating one URL per variant: group queries when they answer the same need.
- Repeating the keyword in every subheading: headings should describe the content naturally.
- Using a fixed density: there's no mandatory percentage.
- Chasing random topics for traffic: it can weaken the site's editorial focus.
- Copying the top results: analyze the search, but add your own value.
- Treating estimates as guaranteed traffic: no tool can promise visits.
- Ignoring Search Console after publishing: your own data can reveal things external tools couldn't.
Checklist for choosing a keyword
| Question | Check |
|---|---|
| Does the topic belong to the site's focus? | Verify |
| Do I understand what the user wants to find? | Verify |
| Have I reviewed the current results? | Verify |
| Are there reasonable signals of demand? | Verify |
| Do I already have a page that answers the same thing? | Verify |
| Can I provide a useful, reliable answer? | Verify |
| Am I confusing an Ads metric with SEO difficulty? | Verify |
| Does the title describe the page naturally? | Verify |
Frequently asked questions about SEO keywords
How many keywords should an article have?
There's no required number. You can use one primary query as a reference and naturally develop the related questions needed to cover the topic.
Should a keyword with zero volume be discarded?
Not automatically. Tools provide estimates, and some very specific queries may show little data. If the question matters to your audience, it can still be worth answering.
Does Google Trends show monthly volume?
No. Trends shows normalized relative interest, useful for comparing changes over time and by region.
Can I rank for queries I never researched?
Yes, that can happen. A page can appear for different queries related to its content; Search Console can help you discover them afterward.
Do I need a page for every keyword?
No. When several queries answer the same need, a single page can cover them thoroughly.
Do keywords guarantee traffic?
No. Research improves editorial planning, but no keyword guarantees indexing, ranking, clicks, or visits.
Sources and official resources
- Google Search Central – SEO Starter Guide
- Google Search Central – Helpful, people-first content
- Google Search Central – Spam policies and keyword stuffing
- Google Ads – Keyword Planner
- Google Ads – Keyword Planner metrics and forecasts
- Google Trends – How data is normalized and interpreted
- Google Search Console – Performance report
Conclusion
Choosing keywords for SEO isn't about finding the phrase with the highest volume and repeating it throughout a page.
Start with your site's audience and focus. Then identify needs, review search results, use demand tools with context, and group queries that genuinely represent the same intent.
Use a primary keyword as an editorial reference if it helps organize the content, but write naturally and avoid densities, repetition, or pages created solely for minor linguistic variations.
Once the site has data, round out your research with Search Console. Your real queries can help you correct, expand, or prioritize content based on what's actually happening, not just earlier estimates.
If you later want to explore how the project can generate revenue, check out the main ways to monetize a website once you've built a useful, coherent content foundation.
Editorial note
This content is for educational and informational purposes. Google's tools, systems, metrics, and features can change over time. Consult the current official documentation before making important SEO decisions.