For years, a lot of SEO advice trained people to think about content in a very technical way.
Put the keyword in the title.
Put the keyword in the first paragraph.
Put the keyword in an H2 heading.
Put the keyword in the image alt text.
Mention it a certain number of times.
Make sure the page is “optimised”.
The problem is that this way of thinking can easily turn writing into a box-ticking exercise.
Instead of asking, “Have I helped the reader?”, the question becomes, “Have I put the right words in the right places?”
And for a long time, that approach often worked.
That was part of the problem.
If putting the keyword in the correct places helped a page rank, there was less incentive to make the article genuinely better. Less incentive to add real experience. Less incentive to explain the issue more clearly. Less incentive to think about the person reading the page and what they actually needed.
The visible SEO technique became easier to measure than the usefulness of the article.
But in my opinion, that is exactly the habit content creators need to move away from now.
Keyword research is useful, but it is not the article
I am not saying keyword research has no value.
It clearly does.
Search terms, People Also Ask boxes, AI Overviews, autocomplete suggestions, forums and keyword tools can all help you understand what people are looking for.
They show you the language people use.
They show you the problems people are trying to solve.
They show you the questions people ask before they understand a subject properly.
That research can be very useful at the start of the writing process.
But once you know what people are looking for, the job is not to mechanically repeat the search phrase.
The job is to answer the question properly.
Naturally.
Clearly.
Usefully.
In a way that a real person would understand.
That is the difference between using search data as research and using keywords as a trick.
The old SEO question was too narrow
The old question was often:
Have I used the keyword enough?
But that is not really the best question.
A better question is:
If someone read this article cold, would they understand who it is for and what problem it solves?
That is a much more useful test.
Because a real reader does not look at an article and count how many times the same phrase appears.
They do not say:
“This phrase appears four times, so I now understand the purpose of the page.”
They read the meaning.
They notice the examples.
They notice the angle.
They notice the audience.
They notice whether the article seems to understand their problem.
They notice whether the page gives them confidence.
That is why I think a simple “cold reader test” could be a better way to review content than obsessing over keyword placement.
What is the cold reader test?
The cold reader test is simple.
After writing an article, give it to someone who has not seen your keyword research, your outline or your original idea.
Ask them three questions:
- Who do you think this article is trying to help?
- What problem, question or decision is this article helping them with?
- What might someone have typed into Google before finding this page?
That is it.
Their answers will tell you a lot.
If their answers match what you intended, your article is probably communicating clearly.
If their answers are very different from what you intended, the problem may not be your keyword placement.
The problem may be that the article itself is not clear enough.
Maybe the introduction does not set up the issue properly.
Maybe the article is too broad.
Maybe the examples are pointing in the wrong direction.
Maybe the headline promises one thing, but the body of the article delivers something else.
Maybe the article answers a general version of the question, but not the specific version your intended reader actually cares about.
That kind of feedback is much more useful than being told to add the keyword one more time.
Why this feels closer to how search works now
Search engines are much better at understanding meaning than they used to be.
They do not only rely on exact keyword matching in the old mechanical sense.
They can understand related questions.
They can understand different ways of asking similar things.
They can understand when two articles are essentially covering the same topic even if the wording is different.
That is why simply creating lots of slightly different articles for slightly different keyword variations feels risky.
If five articles are really answering the same question, just with small wording changes, then it is hard to argue that each one is adding something genuinely useful.
In the same way, if an article is only clear because the keyword has been repeated in obvious places, that is probably not a great sign.
The better aim is for the whole article to make its meaning clear.
The title should help.
The introduction should help.
The headings should help.
The examples should help.
The personal experience should help.
The comparisons, warnings and practical details should help.
The entire page should communicate what it is about.
Not because a keyword has been inserted into the right slots, but because the article is genuinely focused on helping a particular reader with a particular problem.
Ask what the article is really about
One of the most useful things about the cold reader test is that it forces you to look at the article from the outside.
As writers, we often know what we meant to say.
But readers only see what is actually on the page.
That is a big difference.
You may think you have written an article for over-50 men trying to get back into fitness.
But if a cold reader says, “This seems like a general article about exercise,” that tells you something important.
You may think you have written an article for beginner telescope buyers who are confused by technical terms.
But if the reader says, “This seems like a general astronomy article,” the page may not be specific enough.
You may think you have written an article for someone trying to build a helpful affiliate website.
But if the reader says, “This seems like a general SEO article,” perhaps the practical affiliate angle is not coming through clearly.
That does not mean the article is bad.
It means the signal is not strong enough yet.
And the answer is probably not to repeat the target phrase more often.
The answer is to improve the article.
Add the missing context.
Clarify who the reader is.
Use better examples.
Explain the real-life situation.
Make the problem more obvious.
Make the answer more useful.
You can use AI as a cold reader too
You can also use an AI model to perform a version of the cold reader test.
Paste in your article and ask:
Read this article cold. Who do you think it is for? What problem is it solving? What kind of searches might lead someone to this page?
Then look carefully at the answer.
If the AI gives you the same audience and purpose you intended, that is a good sign.
If it gives you a different answer, do not just ignore it.
Ask a follow-up question:
I intended this article to help [specific audience] with [specific problem]. Why does the article not clearly communicate that?
That can be a very useful editing process.
The AI might point out that your introduction is too vague.
It might say your examples are too general.
It might say you have not explained the reader’s situation clearly enough.
It might say the article spends too much time on background and not enough time solving the actual problem.
Used in that way, AI is not replacing your thinking.
It is acting like a mirror.
It is showing you what your article appears to be about from the outside.
This is not about ignoring SEO
The cold reader test is not an argument for ignoring SEO.
It is an argument for putting SEO in the right place.
Use search research to understand demand.
Use AI Overviews, search results, forums and keyword tools to understand what people are asking.
Use that information to choose useful topics and angles.
But when you write, write for the person.
Then, when the article is finished, test whether the meaning is clear.
The old approach was often:
Find keyword. Insert keyword. Publish.
A better approach is:
Find real demand. Write a genuinely helpful answer. Test whether a cold reader understands who it helps and why.
That is a very different mindset.
The best signal is clarity
In the end, the best way to signal what your content is about is not to force the same phrase into every “correct” place.
The best signal is clarity.
A clear title.
A clear introduction.
Clear examples.
Clear headings.
Clear explanations.
Clear practical advice.
A clear sense of who the article is helping.
That does not mean keywords never matter.
Of course the language of the reader matters.
But the keyword should serve the article.
The article should not serve the keyword.
That, I think, is one of the big shifts content creators need to make in the AI search world.
Search data can help you discover what people need.
But the writing still has to do the real work.
And perhaps one of the simplest ways to check that work is to ask a cold reader:
Who do you think this article helps?
If they can answer that clearly, you are probably on the right track.
If they cannot, the article may not need more optimisation.
It may need better communication.