There’s something about keywords (or let’s just call it “prompts” nowadays) that I am deeply passionate about. In my opinion, keyword modelling might be one of the most underused opportunities in SEO.
I have spent years working with keyword databases containing tens of thousands, and sometimes hundreds of thousands, of keywords.
And one thing I have learnt is this:
A keyword list is not an SEO strategy.
A spreadsheet can tell you what people search for.
It cannot, by itself, tell you how those searches relate to one another, what the searcher is actually trying to achieve, which topics your business should own, where the commercial opportunities sit, or how search behaviour is changing.
That is where keyword modelling becomes much more interesting.
For me, keyword modelling is about moving from:
“What keywords should we rank for?”
to:
“What does the entire search landscape around our market look like, and where should we establish authority?”
That is a much bigger question.
And with Google increasingly incorporating AI into Search, I believe it is becoming increasingly important.
SEO used to be heavily keyword-centric
Traditional keyword research has generally followed a familiar process.
Find a keyword.
Look at search volume.
Check the competition.
Assign the keyword to a URL.
Create or optimise the page.
Track the ranking.
Repeat.
There is nothing inherently wrong with this process. In fact, understanding search demand remains fundamental to SEO.
Google itself still recommends using the words people use to search for your content and making important content discoverable and crawlable.
The problem is that the search environment is becoming much more sophisticated than a list of individual queries suggests.
A person rarely has one isolated search.
They have a problem.
And that problem creates a series of related questions.
For example, someone searching for:
“business insurance”
could subsequently search for:
- business insurance for small companies
- professional indemnity insurance
- public liability insurance
- how much business insurance do I need?
- business insurance cost
- business insurance providers
- business insurance comparison
- cheap business insurance
- insurance for contractors
- insurance for consultants
- what does professional indemnity cover?
- business insurance exclusions
- business insurance claims
- best business insurance for a limited company
These are not simply 15 keywords.
They are connected components of a much larger search model.
And that is where the SEO opportunity starts.
What is keyword modelling?
I think about keyword modelling as the process of mapping the relationships between searches.
Instead of treating every keyword as an isolated data point, we can model it around dimensions such as:
Entity → Topic → Intent → Need → Journey → Commercial value
For example:
Entity: business insurance
Topics: liability, professional indemnity, costs, providers, claims, exclusions
Intent: informational, commercial investigation, transactional
Need: understand, compare, evaluate, purchase
Journey: problem → research → consideration → decision
Commercial value: low → medium → high
This changes the question we ask.
Instead of asking:
“How many searches does this keyword have?”
we start asking:
“What role does this keyword play within the market?”
Search volume can hide SEO gold
One of the reasons I like keyword modelling is because search volume can be surprisingly misleading.
A keyword with 100,000 monthly searches might be commercially weak.
A keyword with 500 searches might be incredibly valuable.
And some of the most valuable opportunities may not even have reliable search volume data….
Think about emerging terminology.
New products.
New technology.
New regulations.
New consumer behaviour.
New questions being generated by AI.
There may be little historical search data because the behaviour has not existed long enough to generate it.
Yet these searches can represent the future of the market.
This is one of the reasons I am increasingly interested in search behaviour rather than simply keyword volume.
The biggest shift: from keywords to relationships
Google’s AI search experiences make this particularly interesting.
Google states that AI Overviews and AI Mode can use query fan-out, where an original search can lead to multiple related searches across different subtopics and data sources.
That is significant.
A user might ask:
“What is the best accounting software for a small business?”
The system may need to understand related concepts around:
- pricing
- integrations
- payroll
- invoicing
- tax
- ease of use
- accounting standards
- alternatives
- competitors
- business size
- industry
- implementation
- user experience
The search is no longer necessarily one query leading to one result.
It can become an interconnected research process.
That means businesses need to think about the interconnectedness of their content too.
This is where keyword modelling becomes SEO gold
Imagine your competitors are still working from a conventional keyword list.
They identify:
“accounting software”
“best accounting software”
“accounting software for small business”
They create three pages.
You build a model containing:
Core entity
Accounting software
Commercial themes
- accounting software pricing
- software comparison
- accounting software providers
- best accounting platforms
- alternatives
- free vs paid software
Use cases
- freelancers
- agencies
- ecommerce
- contractors
- startups
- SMEs
Functional topics
- invoicing
- payroll
- expenses
- VAT
- tax
- reporting
- bank feeds
- integrations
Evaluation criteria
- price
- ease of use
- scalability
- support
- integrations
- security
- automation
Decision-stage questions
- which software is best?
- which platform should I choose?
- what does it cost?
- what are the alternatives?
- how does X compare with Y?
Now you are not simply targeting keywords.
You are modelling the market’s information architecture.
That is far more powerful.
Keyword modelling can reveal content gaps competitors don’t see
This is probably one of the biggest practical benefits.
Most businesses analyse competitors by looking at their rankings.
I prefer to ask:
What parts of the search model does each competitor own?
Competitor A may dominate informational searches.
Competitor B may dominate transactional searches.
Competitor C might have excellent comparison content.
Competitor D might have built strong topical authority around a specific use case.
The market can therefore be fragmented.
And that creates opportunities.
You might discover that nobody has strong visibility across the entire journey.
That is potentially far more valuable than finding one isolated keyword with a large search volume.
Keyword modelling can also improve site architecture
One of the biggest mistakes I see is designing websites around organisational structures rather than search structures.
Businesses think about:
Products
Services
About us
Industries
Resources
But customers think differently.
They think:
I have this problem.
What are my options?
How much does it cost?
Which option is right for me?
What are the risks?
Who provides this?
How does it compare?
Can I trust them?
Keyword modelling can help bridge the gap.
Once you understand the relationships between the topics, you can make better decisions about:
- category structures
- landing pages
- supporting content
- internal linking
- information architecture
- navigation
- content hubs
- comparison pages
- product pages
- FAQs
- commercial journeys
The result is a website that reflects how the market searches, rather than simply how the business is organised internally.
It also changes how I think about content
I don’t believe the answer to every keyword opportunity is:
Create another article.
In fact, that can be a very bad strategy.
Google’s current guidance explicitly warns against creating large quantities of pages simply to target variations of searches or manipulate generative AI visibility. Google also emphasises unique, valuable, non-commodity content rather than producing large amounts of interchangeable material.
The model should therefore help you decide:
What content actually needs to exist?
Sometimes the answer is one authoritative page.
Sometimes it is a commercial landing page.
Sometimes it is a comparison.
Sometimes it is a tool.
Sometimes it is an original research report.
Sometimes it is a product page supported by several highly relevant resources.
And sometimes the keyword does not justify a new URL at all.
That is an important distinction.
Keyword modelling should reduce unnecessary content, not create more of it.
The future is not about ranking for every keyword
This is probably the biggest mindset shift I would recommend to businesses.
You do not need to rank for every variation of every keyword.
You need to become highly relevant to the underlying market.
That means understanding:
What does the customer care about?
What problems are they trying to solve?
What entities and concepts are connected?
What questions emerge throughout the journey?
Which areas influence the purchase decision?
Where does our expertise genuinely differentiate us?
That is much closer to how modern search works.
Google’s own guidance now says its systems can understand synonyms and general meaning, meaning businesses do not need to create pages simply to capture every possible long-tail variation.
So how can businesses stay ahead?
This is where I think SEO teams need to become much more strategic.
1. Stop treating the keyword list as the final output
Your keyword research should produce a model of the search landscape, not simply a spreadsheet.
Group keywords by:
- entity
- topic
- intent
- customer need
- journey stage
- product/service
- audience
- geography
- commercial value
The more structured the model, the more useful it becomes.
2. Build around entities and topics
Start with your core commercial entities.
Then map the ecosystem around them.
For every important entity, ask:
What does someone need to know before buying it?
What do they compare it against?
What alternatives exist?
What problems does it solve?
What problems does it create?
What terminology is associated with it?
What adjacent topics influence the decision?
This will often uncover opportunities that traditional keyword research misses.
3. Model the customer journey
Don’t separate keywords into “informational” and “commercial” and stop there.
Go deeper.
Someone can move through:
Problem → Education → Research → Comparison → Validation → Decision → Purchase → Post-purchase
Every stage produces different searches.
A business that only targets the final transactional query is giving competitors a huge opportunity to influence the buyer earlier.
4. Look for relationship gaps
Ask:
Which concepts are connected but poorly connected on our website?
You may have excellent pages about individual products, but very little connecting:
- products to use cases
- services to industries
- problems to solutions
- solutions to outcomes
- comparisons to decisions
This is where internal linking, information architecture and content strategy become much more valuable.
5. Identify emerging search behaviour early
Do not build your entire keyword strategy around historical data.
Monitor:
- new terminology
- customer questions
- sales conversations
- support tickets
- Reddit and forums
- industry publications
- competitors
- emerging technologies
- regulatory changes
- Google Search behaviour
- AI search responses
The businesses that identify new demand before everyone else can build authority before the keyword becomes competitive.
That is genuine SEO advantage.
6. Connect keyword modelling to AI search
This is becoming increasingly important.
Google says AI Overviews and AI Mode can retrieve supporting pages through related searches and can surface a wider and more diverse set of useful links for complex queries.
So I would increasingly test questions such as:
What pages are being retrieved?
What pages are being cited?
Which brands are being recommended?
Which entities are being associated with the topic?
Which competitors appear across multiple related searches?
Which parts of our search model are invisible in AI answers?
That gives you a much richer understanding of digital visibility than traditional rankings alone.
The businesses that win will model entire regional markets, not just keywords
This is ultimately why I call keyword modelling SEO gold.
The value is not in having 500,000 keywords.
The value is in understanding what those 500,000 keywords collectively tell you about a market.
A good keyword model can reveal:
Demand.
Intent.
Customer problems.
Commercial opportunities.
Competitor strengths.
Content gaps.
Topic relationships.
Information architecture opportunities.
Emerging trends.
And increasingly:
AI search opportunities.
That is much more valuable than a spreadsheet full of search volumes.
My approach to keyword modelling
Personally, I would start with the market, not the keyword tool.
Understand the business.
Understand the customer.
Understand the products.
Understand the competitors.
Understand the commercial priorities.
Then build the search model around that reality.
I want to know:
What does this market search for?
But also:
Why do they search for it?
What do they search for next?
What concepts are connected?
What influences the decision?
Where are competitors strongest?
Where are they weak?
What is changing?
And where can the business establish authority before the market catches up?
That is when keyword research stops being a reporting exercise and starts becoming market intelligence.
And I think that is where the real value of SEO sits.
Final thought
SEO has spent a long time trying to answer:
“What keyword should we rank for?”
I think the more important question now is:
“What search ecosystem do we need to own?”
The businesses that understand that distinction will have a significant advantage.
Because the future of search will not be built around isolated keywords.
It will be built around relationships between searches, entities, topics, problems, products, brands and decisions.
And the better you can model those relationships, the better you can decide where your business needs to be visible.
That’s the SEO gold.
