“Keywords are dead” is a good headline. It is a poor strategy. The move away from mechanical exact-match SEO towards better understanding of intent, context, entities, customer questions and conversational search is real. The leap from there to “stop researching keywords” is not.
The useful question is not whether keywords or intent win. They are describing different parts of the same problem.
A keyword or search query tells us how demand is being expressed. Intent helps us understand why that demand exists. Customer situations explain what has happened around the search. Decision factors tell us what somebody needs before choosing. Content, brand, search, AI, reviews, social, email and the website then become the touchpoints through which the organisation can earn the next action.
Wordstell us how people express demand
Intentexplains what they are trying to achieve
Evidencehelps them judge whether to believe the answer
Brandreduces dependence on being rediscovered every time
The WLW position
Do not build an SEO strategy which is only a spreadsheet of keywords. Do not throw the spreadsheet away either. Use keyword and query research as one evidence source inside a wider commercial, customer, content, brand and journey strategy.
Do not build an SEO strategy which is only a spreadsheet of keywords. Do not throw the spreadsheet away either. Use keyword and query research as one evidence source inside a wider commercial, customer, content, brand and journey strategy.
The short version
Keywords still matter
Google’s Search Essentials still says to use the words people use to look for your content. Its SEO Starter Guide still tells site owners to anticipate the different search terms people may use.
Exact-match obsession matters less
Google’s language systems can understand synonyms, related concepts and relevance even when the page does not exactly repeat the query.
Intent adds meaning
“Best running shoes for beginners” and “Hoka vs Brooks” both concern running shoes, but they express different needs and different stages of a decision.
AI does not abolish search foundations
Google says its generative AI Search features remain rooted in core Search ranking and quality systems and retrieve material from the Search index.
Clicks are less dependable
Maps, Shopping, snippets, video, local packs, AI Overviews and AI Mode all compete for attention which once flowed more directly towards conventional organic listings.
Brand becomes more valuable
The strongest position is not being found once for a generic phrase. It is becoming known, remembered, searched for directly and encountered through several useful routes.
The argument in one line
Modern search strategy should expand the chain, not cut the first link off it.
1Identify commercially useful demand
2Understand intent and customer situation
3Build the right page, proof and journey
4Develop brand and multiple discovery routes
Are keywords dead?
No. The statement confuses the decline of an old optimisation habit with the disappearance of the underlying demand signal.
Google itself still tells site owners to use words people would use to find their content and to place those words in useful locations such as the title, main heading, alt text and link text.
Its current SEO Starter Guide goes further. It tells site owners to think about the words different users might search for and explicitly uses the word keywords when explaining how experienced and inexperienced users may phrase the same need differently.
That does not mean Google wants pages mechanically written around an exact phrase. It means the language people use remains useful evidence.
The old model
Treat the keyword almost as an instruction to the page.
17×Repeat phrase
1Exact target
ThinIntent
WeakJourney
RankMain KPI
The stronger model
Use the query as evidence of demand, then understand what sits behind it.
QueryLanguage
IntentMeaning
NeedSituation
ProofConfidence
ValueOutcome
Keywords are not keyword stuffing
This distinction should be obvious, but a lot of online debate relies on pretending the two are the same thing.
Keyword stuffing is the old behaviour of repeatedly forcing a phrase into copy because somebody believes frequency itself will produce rankings. Google explicitly warns against it.
Keyword research is different. It can tell you:
- what people call a product, service or problem
- how demand changes between novice and expert audiences
- which subjects deserve commercial pages
- which questions need supporting content
- how informational and transactional demand differ
- where language used internally does not match the market
- which categories, locations, specifications or use cases matter
- where a company may be missing commercially useful demand
Google already understands more than exact words.
The fact that Google can understand related language is an argument against crude repetition. It is not an argument against researching how customers express demand.
The fact that Google can understand related language is an argument against crude repetition. It is not an argument against researching how customers express demand.
Intent is the meaning behind the query
A sophisticated strategy should absolutely go beyond a list of keywords. The important next question is what the person is trying to accomplish.
QueryWhat words or question did the person use?
IntentWhat are they trying to understand, find, solve or choose?
SituationWhat has happened to create the need now?
DecisionWhat factors and proof will determine suitability?
ActionWhat useful commercial or organisational step should follow?
Consider these searches:
One category. Several different needs.
best running shoes for beginners
Hoka vs Brooks
how long do running shoes last
running shoes for bad knees
do expensive running shoes last longer
best running shoes for marathon training
Those are not merely six strings of text. They reveal different levels of knowledge, concerns, use cases and decision stages.
But the query is still useful. It is the observable evidence from which we start interpreting the need.
A better sequence
Keyword → query → intent → customer situation → decision factors → page or content → proof → conversion route.
Keyword → query → intent → customer situation → decision factors → page or content → proof → conversion route.
Customer moments are useful, but not new search physics
Terms such as “research moment”, “buying moment”, “doubt moment” or “comparison moment” can be useful because they force a strategy to think about the circumstances around demand.
They are not a replacement for keyword and query research. They are another way of organising customer intent and the journey around it.
What happened?
A problem appeared, a requirement changed, an existing supplier disappointed, a product failed or a new opportunity emerged.
What is wanted?
The customer wants to learn, find an option, solve the problem, compare alternatives, reduce risk or make a decision.
What stops action?
Price, uncertainty, specification, availability, compatibility, risk, evidence, experience, timescale or another decision factor.
Those layers make keyword research more commercially useful. They do not make it unnecessary.
Commercial pages still have a job
One of the risks in the current debate is replacing an overly narrow keyword strategy with an equally narrow “answer every question” strategy.
Businesses still need clear commercial pages.
A boiler company should still have useful pages around boiler repair, installation, replacement, servicing and the manufacturers or systems it genuinely supports. A riding holiday business should still have pages which make destinations, experience levels, types of holiday and relevant commercial choices easy to find.
Supporting content should strengthen those routes rather than compete with them or replace them.
Commercial demand
Identify the service, product, location, category or outcome people actively seek.
Identify the service, product, location, category or outcome people actively seek.
Commercial page
Create the strongest useful destination for that demand.
Create the strongest useful destination for that demand.
Supporting questions
Answer the needs, doubts, comparisons and situations around the purchase.
Answer the needs, doubts, comparisons and situations around the purchase.
Conversion route
Connect useful information to the next appropriate action.
Connect useful information to the next appropriate action.
The boiler example shows how the pieces fit together
Suppose a boiler company wants commercially useful visibility for:
boiler repair
boiler installation
boiler replacement
boiler servicing
Vaillant boiler repair
Those are valuable commercial concepts. They should map to useful service pages where the business genuinely provides the service.
Now add the questions which appear earlier in the problem:
Why is my boiler making a banging noise?
A problem-led information need which can establish expertise and explain when an engineer may be required.
Why does my boiler keep losing pressure?
A diagnostic-style customer question which can connect symptoms to safe next steps without pretending every cause can be identified online.
Why is my boiler whistling?
A search which may begin informationally but can develop into a service requirement.
What does a flashing red light mean?
A highly specific situation which can lead naturally towards manufacturer guidance, troubleshooting and repair.
The useful information answers the immediate question. The journey can then explain when professional help becomes appropriate and link into the relevant service and local coverage.
Nothing revolutionary happened
The business connected informational demand to commercial relevance without turning the information page into an advert.
AskThe customer expresses a problem
HelpThe content answers the question properly
ProveExperience and evidence build confidence
RouteThe right service becomes available when needed
AI search makes the strategy broader, not obsolete
The arrival of AI Overviews, AI Mode and other generative search experiences makes this distinction more important, not less.
Google’s current guidance says SEO best practices remain relevant for generative AI Search because its AI features are rooted in core Search ranking and quality systems. Google describes using retrieval augmented generation to retrieve relevant pages from the Search index and query fan-out to run related searches around the original request.
So the new layer does not abolish search infrastructure. It uses it.
Google’s position is unusually clear.
AEO and GEO may be useful industry labels, but Google says optimising for its generative AI search experiences is still optimisation for Search. Strong technical foundations and useful, distinctive content remain central.
AEO and GEO may be useful industry labels, but Google says optimising for its generative AI search experiences is still optimisation for Search. Strong technical foundations and useful, distinctive content remain central.
Query fan-out actually strengthens the need to understand demand
If an AI system expands one question into several related searches, it becomes even more important for a business to understand the subject, the adjacent customer questions and the decision journey around it.
It does not mean creating a separate page for every possible fan-out query. Google explicitly warns against mass-producing pages for every query variation simply to manipulate search or generative AI responses.
Weak AI response
Create hundreds of near-duplicate pages because an AI system might ask hundreds of related questions.
Stronger response
Understand the commercially useful subject properly, build clear pages and publish distinctive evidence which answers the important needs around it.
The real change is distribution
Where the “keywords are dead” argument does contain an important truth is that a search no longer guarantees the kind of conventional organic click businesses once expected.
Google has been adding its own answer formats and vertical experiences for years:
Featured snippets
Maps
Shopping
People Also Ask
Video
Knowledge panels
Local packs
Flights
Hotels
AI Overviews
AI Mode
The SERP has changed. The amount of attention available to a conventional blue link has changed. Some searches can now be satisfied without the user making a normal organic visit.
That is a distribution problem.
It does not prove that keywords are dead. It proves that a business should not treat Google organic clicks as its entire acquisition strategy.
Do not confuse a changing SERP with the disappearance of customer demand.
The interface through which Google presents an answer can change dramatically while the underlying need, language, problem and decision still exist.
The interface through which Google presents an answer can change dramatically while the underlying need, language, problem and decision still exist.
Google Ads provides a useful reality check
Organic Search and Google Ads are different systems, so paid-search behaviour should never be presented as proof of how Google’s organic ranking algorithm works.
It is still a useful check against the sweeping claim that Google no longer cares about search terms or keywords.
Google’s current responsive search ad guidance tells advertisers to use relevant keywords in headlines and descriptions. Ad Strength evaluates, among other things, how well creative assets relate to keywords in the ad group. Google also continues to offer keyword insertion and uses AI-powered matching to expand from existing keywords, creatives and landing pages towards relevant search queries.
Google’s own paid search system still connects several layers
AI has expanded matching. It has not made the relationship between query, keyword, message and landing page irrelevant.
QueryWhat the person actually searched
KeywordA targeting and relevance signal
MessageThe ad adapts towards customer relevance
PageThe destination must still satisfy the need
Ranking for a keyword and understanding a topic are different
A modern site should not treat every phrase as though it needs a separate exact-match page.
It still needs to understand what each important page is for.
A simple coffee example
The commercial architecture and the supporting subject knowledge do different jobs. Use the tabs to see the distinction.
The point is not the exact phrase count. The point is knowing what the page represents and which commercially useful demand belongs there.
Supporting material expands the useful subject coverage around the commercial range without requiring one page for every phrase.
Keyword and query research can help determine what pages are needed, what they should be called, how people describe the subject, what questions need answering and where informational demand differs from transactional demand.
The stronger evolution is therefore not “stop using keywords”. It is “stop treating keyword targeting as the whole strategy”.
Brand is the bigger strategic prize
This is where the argument should move beyond SEO.
A business whose entire acquisition model depends on being rediscovered for a generic Google query remains exposed to every change in rankings, SERP design, paid competition, AI answers and competitor activity.
A stronger business is increasingly known before the search.
Be discoveredSearch, AI, PR, social, referral, advertising or another encounter creates awareness.
Be usefulThe organisation answers a real need rather than merely chasing visibility.
Be rememberedDistinctive brand and repeated useful encounters build recognition.
Be searchedMore people begin looking for the organisation by name.
Be chosenProof, relevance and experience turn awareness into action.
Brand search is still search demand. Direct visits are still digital behaviour. Referrals are still acquisition. Email still brings people back. Social and video can create discovery. PR and independent coverage can create reassurance. AI can create another route into the same commercial system.
The objective should be reduced unnecessary dependence on any single route.
The strategic shift is bigger than Keywords → Moments.
A more useful sequence is: Demand → query → intent → content → proof → brand → multiple acquisition routes → conversion → retention.
A more useful sequence is: Demand → query → intent → content → proof → brand → multiple acquisition routes → conversion → retention.
A modern search and content model
The job is not to pick one fashionable concept and declare the previous one dead. It is to connect the layers which now matter.
Commercial demand
What products, services, problems or outcomes create value?
What products, services, problems or outcomes create value?
Keywords and queries
How do people express that demand in the real market?
How do people express that demand in the real market?
Intent and situations
What are they trying to understand, find, solve or choose and why now?
What are they trying to understand, find, solve or choose and why now?
Decision factors
What determines suitability, confidence and risk?
What determines suitability, confidence and risk?
Architecture
Which commercial pages and supporting content does the organisation genuinely need?
Which commercial pages and supporting content does the organisation genuinely need?
Proof
What results, experience, research, reviews and independent evidence earn confidence?
What results, experience, research, reviews and independent evidence earn confidence?
Discovery
How do search, AI, PR, social, email, referral and other channels work together?
How do search, AI, PR, social, email, referral and other channels work together?
Commercial outcome
What behaviour, lead, sale, booking, application or other useful action should improve?
What behaviour, lead, sale, booking, application or other useful action should improve?
Completeness of research does not require completeness of production
A serious strategy may research hundreds or thousands of queries. It does not follow that the business needs hundreds or thousands of pages.
The research is there to understand the market. Production should be limited to work with a clear role.
Research broadly
Understand the full demand landscape, the customer questions, alternatives, situations, decision factors, competitors and emerging themes.
Build selectively
Create only the pages, content, proof and activity which have a clear commercial or organisational role.
What might AI systems pick up from a stronger strategy?
The goal should not be to write for a machine. But clearer, more explicit information also gives retrieval and generative systems more useful evidence to work with.
Clear relationships
What the organisation does, who it helps, where it operates, which situations it solves and how products, services and expertise relate.
Decision evidence
Price, specification, experience, location, compatibility, availability, risk, proof, results and other factors customers use to judge suitability.
Distinctive knowledge
Original research, first-hand experience, named specialists, real cases, useful comparisons and material which cannot be reproduced by simply summarising everybody else.
Google’s 2026 generative AI guidance is particularly relevant here. It recommends unique, useful, non-commodity content and warns against producing a separate page for every possible query variation.
So the practical implication of query fan-out is not “generate more pages”. It is “understand the topic and customer need well enough to create useful coverage”.
What should actually be measured?
Ranking positions are useful evidence. They are not the business outcome.
| Layer | Useful measures | What it tells you |
|---|---|---|
| Demand | Search volume, query groups, trends, branded search | What people appear to want and how the market language changes |
| Visibility | Impressions, rankings, Search Console clicks, AI or generative visibility where available | Whether the organisation is being surfaced for useful demand |
| Engagement | Landing page behaviour, repeat visits, content progression, calls, form starts | Whether the experience is useful after discovery |
| Conversion | Enquiries, quotes, bookings, sales, applications, subscriptions | Whether visibility is creating the required action |
| Commercial value | Qualified pipeline, revenue, margin, lifetime value, cost per acquisition | Whether the activity creates meaningful value rather than activity for its own sake |
| Brand resilience | Direct traffic, branded search, returning users, referral, email, social and PR-led discovery | Whether acquisition is becoming less dependent on one algorithm or channel |
Traffic can fall while value improves.
A changing SERP means clicks alone need context. Fewer but more qualified visits may be commercially stronger than a large volume of weak informational traffic.
A changing SERP means clicks alone need context. Fewer but more qualified visits may be commercially stronger than a large volume of weak informational traffic.
What businesses should do now
Keep keyword and query research
Demote it from “the strategy” to a core research input.
- Use it to understand market language and commercially useful demand
- Map important query groups to the right commercial pages
- Protect pages which already rank, convert or carry authority
- Do not replace useful evidence because a new terminology becomes fashionable
Add the customer situation and decision
Interpret why the query exists rather than treating the words as the complete brief.
- What created the need?
- What is the person trying to achieve?
- What alternatives are being considered?
- What concerns could stop action?
- What proof determines trust and suitability?
Do not chase a new absolute
“Keywords are dead” is no more useful than the old claim that one exact keyword should control every page.
- Do not create one page for every query variation
- Do not confuse AI visibility with abandoning SEO
- Do not measure success only through organic clicks
- Do not let channel tactics replace commercial strategy
Search changed. The job became bigger.
WLW connects search demand, customer questions, content, AI visibility, proof, brand, website journeys, campaigns and conversion around the commercial problem rather than treating SEO as an isolated list of rankings.
The question is not “keywords or moments?” It is which evidence helps us understand where demand and value actually sit, then what should happen next.
Get my full online presence auditFrequently asked questions
Are keywords dead in SEO?
No. Google still recommends using words people use to look for content and tells site owners to anticipate the different search terms users may use. What has declined is the value of crude exact-match optimisation and keyword stuffing.
Does Google still use keywords?
The word “keyword” can mean several different things. Google does not use the old meta keywords tag and warns against keyword stuffing. However, Google still discusses the words and search terms users enter, and its Search Essentials recommends using that language in prominent, descriptive parts of a page.
What is the difference between a keyword and search intent?
A keyword or query is the language used to express demand. Search intent is the purpose behind it: what the person is trying to learn, find, solve or choose. Strong strategy uses both rather than pretending one replaces the other.
What are customer moments?
Customer moments are situations around demand, such as research, doubt, comparison, urgency or purchase. They can be useful for understanding the customer journey, but they are not a replacement for keyword, query or market research.
Does AI Mode make SEO obsolete?
No. Google’s current guidance says SEO best practices remain relevant for generative AI Search because AI Overviews and AI Mode are rooted in core Search ranking and quality systems and retrieve relevant material from the Search index.
What is query fan-out?
Query fan-out is Google’s term for generating several related searches around an original user request so its AI systems can retrieve additional relevant information. It does not mean publishers should create a separate page for every possible variation.
Should every keyword have its own page?
No. A page should have a clear useful subject and role. Google’s language systems can understand relevance without an exact query match, and Google warns against mass-producing pages simply to target every query variation.
Are organic clicks becoming less dependable?
For many search types, yes. Search results now contain maps, shopping, video, snippets, local results, AI Overviews and AI Mode. Businesses should therefore measure the value of visits and develop several acquisition routes rather than relying only on conventional organic clicks.
Why does brand matter to search strategy?
A stronger brand creates branded search, direct visits, referrals, repeat traffic and recognition across several channels. It reduces unnecessary dependence on winning the same generic search battle every time a customer enters the market.
What should replace an old-fashioned keyword strategy?
Nothing should simply replace it. Expand it. Start with commercial demand and keyword or query evidence, interpret intent and customer situations, map the right pages, add proof and authority, build brand across several touchpoints, then measure conversion and commercial outcomes.
Sources and evidence basis
- Google Search Essentials. Current guidance includes using words people would use to look for content in prominent page locations.
- Google SEO Starter Guide. Current guidance discusses anticipating readers’ search terms, different keywords and Google’s ability to understand related queries without exact wording.
- Google’s guide to optimising for generative AI features on Google Search. SEO remains foundational, Google describes retrieval augmented generation and query fan-out and recommends unique, useful, non-commodity content.
- Google Ads: Ad Strength for responsive search ads. Current documentation discusses the relationship between creative assets and keywords in ad groups.
- Google Ads: Create effective responsive search ads. Google recommends adding popular keywords to relevant headlines.
- Google Ads: About responsive search ads. Responsive ads adapt messages towards potential customers’ search terms.
Evidence boundary: Google Ads and organic Search are different systems. Paid search guidance is used here only as evidence against the sweeping claim that search terms and keyword relevance have ceased to matter to Google, not as proof of how organic ranking works.
