Your Website Doesn’t Have ‘Users.’ It Has People.
A logical question during any optimisation process is:
Is this page performing?
And, just as importantly:
For whom is it performing?
Google Analytics, Search Console and other tools provide a deluge of useful activity data:
Queries. Impressions. Clicks. Engagement. Conversions.
It’s vital evidence.
But evidence about activity doesn’t automatically describe the person behind that activity, what they’re trying to accomplish, or why they searched in the first place.
And that’s the real opportunity.
Content Exists to Help Someone Achieve a Goal
Content shouldn’t exist simply to rank higher for a keyword.
It exists to help a person looking for a solution to a problem achieve a goal.
When we’re doing our job properly, helping that person also contributes towards the business outcome that keeps the organisation running.
This is why personas, when used as a tool for context rather than a marketing exercise, can become considerably more valuable to SEO and conversion optimisation than their reputation might suggest.
A Persona Isn’t ‘Carmen, 43, Who Likes Yoga’
Marketing personas can occasionally wander into caricature.
Carmen, 43. Likes yoga. Drinks cappuccino. Perhaps enjoys loud Maltese festa petards.
Wonderful.
What exactly do I optimise with that?
A useful AWE persona focuses on context:
- What is this person trying to accomplish?
- What problem brought them to this search?
- What are they worried about?
- What constraints might prevent them from progressing?
- What information do they need before making a decision?
Personas aren’t decorations. They are filters through which we interpret evidence.
And when they’re developed collaboratively with the business and grounded in actual customer knowledge, personas can help expose search needs that raw keyword data may not immediately reveal.
Same Broad Topic. Same Keyword. Very Different People.
Imagine an elderly-care website.
Search Console shows pages appearing for broad searches such as:
- elderly care Malta
- carer cost Malta
Without context, they’re search data.
But look at the very different people who could be sitting behind them.
Maria – The Anxious Family Decision-Maker
Maria needs help caring for an elderly parent.
She’s worried about trust, dignity, affordability and the terrifying responsibility of making a decision for someone she loves.
She needs reassurance and evidence before she’s prepared to talk to anyone.
Ġużeppi – The Indomitable Older Person Investigating His Own Care
Ġużeppi may have fallen, but he’s perfectly capable of deciding what happens in his own home.
He’s more tired of his family fussing over him than he is of getting old.
His goal isn’t simply receive care.
It’s:
REMAIN INDEPENDENT.
Language suggesting that somebody else is deciding his future could alienate him immediately.
Sarah – The Prospective Carer
Sarah is looking for employment.
Her search behaviour may look beautifully aligned with the site’s topic of ‘care’, while being commercially irrelevant to a page intended to attract families looking for care services.
Same broad topic.
Potentially even overlapping search language.
Very different people. Very different goals.
If we don’t understand those contexts, we risk misinterpreting the data from the start.
Data tells us what happened. Personas help us ask whether it happened with the people the page was actually designed to help.
Context Also Changes Keyword Research
This is where personas can move from being a passive filter to becoming an active part of search optimisation.
Traditional keyword research might begin with:
Topic → Keyword → Volume → Difficulty → Page
But a persona gives us another route:
Person → Need → Constraint → Search Intent → Query → Page
Maria might search for:
how to convince elderly parent to accept care
help caring for elderly parent Malta
Ġużeppi might express a similar broad topic very differently:
home care while keeping independence
care at home on my terms
These aren’t simply longer versions of elderly care Malta.
They’re expressions of the circumstances surrounding a person’s need.
That can help uncover valuable long-tail searches shaped by problems, anxieties and constraints rather than the terminology of the service itself.
And reassuringly, this isn’t just an AWE flight of fancy.
HubSpot has discussed precisely this connection between user intent and long-tail keyword discovery.
Its argument is that concentrating solely on keyword popularity can cause us to miss valuable long-tail searches; understanding why somebody is searching can expose different keyword opportunities.
HubSpot – Why Your Keyword Strategy Is Incomplete Without User Intent
More recently, Semrush has taken the relationship between personas and search considerably further with AI search.
Its Enterprise AI Optimisation tooling includes persona-based prompt generation, using characteristics such as goals, budget sensitivity, experience and decision factors to investigate how different audiences might frame their questions. Semrush describes the broader shift as moving beyond merely what people search towards understanding who is searching.
Semrush – Persona-Based Prompt Generation for AI Search
I didn’t arrive at persona-informed optimisation because HubSpot or Semrush told me to do it.
The thinking grew naturally from trying to connect search evidence, page purpose, customer needs, customer journeys and business outcomes.
But it’s reassuring to discover others independently applying similar principles.
It suggests this isn’t merely a fancy thought worth discussing over a nice cuppa tea.
There’s something worth investigating.
One Person Can Search Many Ways.
One Keyword Can Represent Many People.
There’s another consequence.
Discovering ten long-tail queries doesn’t mean we should create ten pages.
They may simply be ten linguistic expressions of the same underlying need.
Equally, one apparently simple keyword could conceal several very different intentions.
So perhaps the optimisation question shouldn’t simply be: ‘Which keywords should we rank for?’
It becomes: ‘Which people and needs should we be relevant to, how are those needs expressed in search, and which page should best satisfy them?’
That distinction becomes even more interesting as search grows increasingly conversational.
Google’s current guidance for generative AI search specifically warns against creating separate content for every possible variation of how people might search. Instead, it recommends concentrating on useful, satisfying content for visitors; Google’s systems are designed to understand page relevance even without an exact query-to-page wording match.
Google – Optimising for Generative AI Features in Search
Google’s broader people-first content guidance makes another particularly relevant point. It asks whether a site has an existing or intended audience that would find its content useful and whether somebody leaves having learned enough to help achieve their goal.
Google – Creating Helpful, Reliable, People-First Content
Google isn’t saying: ‘Create personas and you’ll rank higher.’
It isn’t.
But its people-first principles are remarkably compatible with persona-informed optimisation.
Both bring us back to a deceptively simple question: Who are we actually trying to help?
When AI Enters the Room
Most discussions about AI analysis seem obsessed with giving AI more data.
There is another possibility: Give it better context.
Without persona context, an AI audit might observe:
‘This page receives traffic for elderly-care queries.’
Useful.
Now provide the intended audience, page purpose, business objective, observed search evidence, and the needs and constraints established through the persona.
The diagnostic question becomes: Does this page appear to answer the questions, anxieties and decision needs of the people the business actually wants this page to help?
That’s a fundamentally different audit.
It shifts from AI finding patterns towards AI assisting human judgement.
And there’s an important boundary here.
A Persona Isn’t Evidence
We’re not asking AI to pretend to be Maria and magically tell us what Maria is thinking.
That would simply give a machine a fictional character and invite it to fill in the blanks.
The persona itself shouldn’t tell us what happened.
Search Console, Analytics, crawl evidence, page content, conversions and real business information provide evidence of what actually happened.
The persona provides context against which we evaluate that evidence.
And the persona itself should be grounded, as far as possible, in real business knowledge, customer experience, research and observed behaviour, rather than whatever demographic stereotypes an AI happens to manufacture.
That gives us something closer to: Site → Page → Evidence → Persona → Purpose → Business Outcome
AI can help us investigate the relationships between those things.
It doesn’t get to invent them.
AI can expand the hypothesis space. Evidence constrains it. Human judgement decides what deserves action.
From ‘Users’ Back to People
This is why personas can be valuable within AWE’s Site Intelligence audits and wider optimisation approach.
Site-level personas help establish: Who does this business exist to serve?
Page-level personas become more specific: Who does this particular page need to help?
What brought them here?
What are they trying to accomplish?
What might prevent them from progressing?
What might they search for before they arrive?
And what might they need next?
Search data can validate, contradict and refine our assumptions.
Analytics can show us what actually happened.
Business knowledge provides context.
AI can help us examine more of the relationships between them.
But neither the persona, the analytics nor AI should be allowed to impersonate the human being at the centre of the system.
AI doesn’t need another million data points nearly as badly as it needs to understand what we’re trying to achieve, for whom, and why.
Let’s understand people’s context.
Let’s use evidence to challenge our assumptions about them.
And when we use AI to help us think outside our own experiential and socially conditioned boxes, let’s give it clear context and boundaries around the business goal we’re actually optimising for.
Otherwise, we haven’t escaped the box.
We’ve simply asked AI to build us a bigger one.
Because a website doesn’t really have ‘users’.
It has people trying to get somewhere.
🔗 Follow AWE on LinkedIn: https://www.linkedin.com/company/awe-conversion
🔹 Follow the discussion on Facebook for daily AWE strategies: https://www.facebook.com/awe.conversion