A practical guide to keyword research in 2026 - covering search intent, AI-era SEO, long-tail strategy, and tools that help you find terms that actually convert.
Keyword research used to be simple. You opened a spreadsheet, typed your product category into a tool, sorted by monthly search volume, and aimed for the terms with the best numbers. That approach still works - partially. But in 2026, the search landscape has shifted enough that relying on volume alone is a reliable way to invest months of effort into content that gets swallowed by AI Overviews before a single human ever clicks.
This guide covers how to approach keyword research properly. Not the version that worked in 2019, and not the panicked "AI is killing SEO" version either. The practical, grounded approach that helps you find terms worth targeting.
Before we get into the how, a word on the why.
Google's AI Overviews, Perplexity, and ChatGPT Search have changed how people consume search results. For some queries - especially simple factual ones - users get their answer directly on the results page and never click through. That is genuinely a concern for content marketers.
But here is what the data actually shows. According to research from Semrush published in late 2025, long-tail informational queries with genuine depth - the kind that require examples, comparison, and context - still drive substantial click-through rates even when an AI Overview is present. The terms that got crushed were thin, single-answer queries where a paragraph could satisfy the question entirely.
The implication is clear. Keyword research now has to account not just for "is anyone searching this?" but "is there room for a real piece of content here, or will an AI summary eat this alive?"
"The future of SEO is content that AI systems want to cite, not content that competes with them." - Lily Ray, VP of SEO Strategy at Amsive
That changes which keywords you target. It does not eliminate the need to target keywords at all.
Every search query falls into one of four categories. Getting this right is more important than almost anything else in keyword research.
Informational - The user wants to learn something. "How does compound interest work?" These queries can drive huge traffic but rarely convert directly. They build trust and authority over time.
Navigational - The user is trying to reach a specific site or page. "Athenic login" or "ASOS returns portal." These are only worth targeting if people are looking specifically for you.
Commercial - The user is researching before buying. "Best CRM for small businesses" or "Notion vs Monday." These convert well and are highly competitive.
Transactional - The user is ready to act. "Buy standing desk UK" or "sign up Shopify trial." High conversion, high competition.
Most content strategies over-index on informational content (because it is easier to write) and neglect commercial intent queries where the real business value lives. A balanced keyword strategy includes all four types, mapped to different stages of your funnel.
Let's walk through how a Shopify store selling sustainable footwear might approach keyword research. This makes it concrete.
Start with the obvious seed keywords: "sustainable trainers," "eco-friendly trainers UK," "vegan running shoes." Drop these into a tool like Ahrefs, Semrush, or the free Google Keyword Planner.
From "sustainable trainers UK" you might see:
Now ask: what is the intent here? Someone searching "sustainable trainers UK" is probably in the commercial or transactional phase. They are not asking what sustainability means - they want to buy something. That is a product category page, not a blog post.
Meanwhile, "are sustainable trainers actually sustainable?" has lower volume (maybe 300 searches/month) but is pure informational intent. It deserves a well-researched article. And crucially, it is the kind of nuanced question that an AI Overview cannot fully satisfy - it requires brand comparisons, certification analysis, and genuine editorial voice.
This is the distinction that separates good keyword research from great keyword research. Not just finding terms, but understanding what content format those terms demand.
| Keyword Type | Example | Avg. Volume | Difficulty | Conversion Rate | Best Content Format |
|---|---|---|---|---|---|
| Short-tail / Head | "trainers" | 100,000+ | Very High | Low | Category page |
| Mid-tail | "sustainable trainers UK" | 1,000-10,000 | Medium-High | Medium | Category + landing page |
| Long-tail informational | "how to tell if trainers are sustainable" | 100-1,000 | Low-Medium | Low (brand building) | Blog post / guide |
| Long-tail commercial | "best sustainable trainers under £100" | 200-2,000 | Medium | High | Roundup / comparison |
| Question-based | "are vegan trainers better for the environment" | 50-500 | Low | Low-Medium | FAQ / article |
| Local | "sustainable trainer shops London" | 100-500 | Low | High | Local landing page |
The key takeaway from this table: high volume does not equal high value. A 200-search/month commercial keyword with low difficulty and high conversion potential is often worth more to a small business than a 50,000-search term where you will never crack page one.
Start with what you know. Write down every product, service, topic, and question your customers ask. Talk to your sales team. Read your own customer support emails. These are gold.
Add your competitors' URL into a tool like Ahrefs Site Explorer to see what terms they rank for. Not to copy them - to identify gaps.
Take your seed list and run it through a tool. Free options include Google Keyword Planner (requires a Google Ads account, but you do not need to spend anything) and Google Search Console if your site already exists. Paid tools like Semrush, Ahrefs, and Moz give you far more data.
Look at:
For each candidate keyword, actually Google it. Look at the first page of results. Are they product pages, blog posts, videos, forums? The format of existing results tells you what Google thinks the intent is. Match your content format to that.
If the top results for a keyword are all detailed 3,000-word guides, a thin 500-word article will not rank no matter how well-optimised it is.
Keyword difficulty tools are helpful but imperfect. A score of 40 on one tool might mean something different on another. What matters is: can you realistically rank here given your domain authority, the quality of content you can produce, and the time you have?
New sites should focus almost exclusively on low-difficulty, long-tail keywords for the first six to twelve months. It is not glamorous advice. But it works.
Organise your chosen keywords into clusters. Group related terms around a central "pillar" topic. For our trainer example, "sustainable trainers" is the pillar. Supporting cluster content includes: "how to care for sustainable trainers," "sustainable trainer brands ranked," "what certifications to look for in eco footwear."
This cluster approach helps with internal linking and signals to Google that your site has genuine topical authority - not just a single well-ranked page.
| Tool | Cost | Best For | Limitations |
|---|---|---|---|
| Google Keyword Planner | Free | Initial volume data, CPC | Ranges not exact numbers, requires Ads account |
| Google Search Console | Free | Seeing what you already rank for | Only works if your site has traffic |
| Ubersuggest | Free / £29/mo | Beginners, content ideas | Limited data depth |
| Semrush | From £108/mo | Full keyword + competitor analysis | Expensive for small businesses |
| Ahrefs | From £99/mo | Backlink + keyword research | Steep learning curve |
| AnswerThePublic | Free / £99/mo | Question-based keyword discovery | No difficulty/volume data |
| Google Trends | Free | Seasonal trends, rising topics | No absolute volume figures |
For most small businesses, a combination of Google Search Console (free), Google Trends (free), and one paid tool on the lowest tier gives you 80% of what you need.
Here is the part most keyword guides are not fully addressing yet.
Generative Engine Optimisation - or GEO - is the practice of optimising content to appear within AI-generated responses, not just in the ten blue links. Tools like Perplexity, ChatGPT Search, and Google's AI Overviews draw from indexed web content. But they favour certain types of content.
What gets cited:
What this means for keyword research: question-based queries ("how to", "what is", "why does") have become more strategically important, not less. Even if the click-through rate for these queries has fallen, appearing within an AI Overview still builds brand visibility and trust.
It also means that thin content targeting high-volume head terms is now doubly risky. You are unlikely to outrank established sites anyway, and even if you do, an AI Overview may absorb the query.
Athenic's AI SEO Engine automates much of the keyword discovery and analysis process. It surfaces high-opportunity keywords based on your existing domain, identifies content gaps versus competitors, and maps keywords to the right content formats automatically.
Rather than spending hours in spreadsheets, you describe your business and your goals. The platform does the heavy lifting - competitor gap analysis, difficulty scoring, intent classification, cluster mapping - and surfaces a prioritised list of opportunities. You stay in control of the editorial decisions.
If you are currently managing your SEO manually and spending significant time on keyword research, it is worth exploring what a structured AI-assisted approach can do for your output. You can read more about how AI is changing SEO strategy.
What makes a good keyword to target?
A good keyword has three qualities: there is genuine search volume (enough people are looking for it to matter), you have a realistic chance of ranking for it given your current domain authority, and it aligns with a stage in your customer's buying journey. Volume alone is not enough. A 200-search/month keyword that converts buyers is usually more valuable than a 20,000-search/month keyword that you cannot rank for.
How often should I update my keyword research?
Quarterly reviews are a sensible minimum for most businesses. Certain industries - anything tied to consumer trends, technology, or regulation - warrant monthly checks. Google Trends is useful for spotting seasonal spikes. Search Console data tells you which terms are losing click-through rate, which is often the earliest signal that a keyword is being absorbed by AI Overviews or competitors.
Is free keyword research good enough for a small business?
Often, yes - at least to start. Google Search Console, Google Keyword Planner, and AnswerThePublic give you a workable foundation. The biggest gap with free tools is competitor analysis: knowing what terms your rivals rank for requires a paid tool. If SEO is a meaningful channel for your business, the investment in one paid tool tier usually pays for itself quickly.
What is keyword difficulty and should I trust the scores?
Keyword difficulty is a proprietary score (typically 0-100) that estimates how hard it is to rank for a term on the first page of Google. Different tools calculate it differently - Ahrefs weights backlink profiles heavily, Semrush factors in a broader range of signals. Treat the scores as directional guides, not absolute truth. A KD of 35 in one tool might be 50 in another. The real test is: look at who is currently ranking, and honestly assess whether you can produce content that competes with them.
How has AI search changed which keywords are worth targeting?
AI search has accelerated a shift that was already underway. Simple, single-answer queries are increasingly satisfied on the results page itself. The keywords most worth investing in are those requiring depth, nuance, comparison, or genuine expertise - content that AI systems want to cite rather than summarise away. Question-based and commercial-intent keywords with real complexity are the sweet spot for 2026.
Keyword research is not the most exciting part of running a business. But it is the foundation that everything else - your content strategy, your SEO, your organic acquisition - is built on. Get it right and the returns compound over time. Get it wrong and you spend months producing content that nobody finds.
Take the time to do it properly. Then revisit it regularly. The search landscape will keep evolving, and your keyword strategy needs to evolve with it.