Most content strategies fail before the first piece is published, at the moment you decide who you are writing for.
The usual answer is "for my potential clients". That answer lets you do nothing. It does not tell you which words to use, which subjects to cover, or which questions deserve a page of their own.
Here is the exercise that makes it concrete, in four steps.
Step one: list what has already happened
Take your last twelve clients. Not the ones you wish you had, the ones you actually had.
For each, write down four things: what they paid, how long the work took, how much you enjoyed doing it, and whether they came back or referred someone.
This table is uncomfortable to fill in, because it usually shows that the most demanding clients are not the best paying, and that the work you enjoy most is not always what you sell most. That is exactly the point.
Step two: rank rather than segment
The classic mistake is to cut your market into several profiles and try to speak to all of them.
The job is not to cut but to rank. Who do you want to sign first? Who next? Who would you accept if the calendar allows? And who, honestly, is no longer worth it?
The order matters more than the list. It determines which content you produce first, while your resources are limited and you have to choose.
Step three: describe the situation, not the person
An ideal client is not "a woman aged 35 to 50, management level". That description produces no content.
What produces content is the situation: where this person is when they start needing you, what event triggers the search, what deadline presses on them, what they fear when hiring someone like you, and who else weighs in on the decision.
A company looking for a provider because their internal lead just resigned does not have the same questions as a company looking because it is growing too fast. They can sit in the same sector at the same size and still need two completely different pieces of content.
Step four: translate into real questions
For each situation you identified, write the questions that person would ask out loud.
These need to be questions rather than keywords, phrased the way they would phrase them. The move to keywords comes afterwards, and it is mechanical. The hard part is recovering the natural phrasing, the one somebody actually types when they are alone at their screen and do not yet know your industry's vocabulary.
Your old correspondence is a goldmine for this: first-contact emails, messages before quotes, objections raised in meetings. People use their own words there, and those are the words they type into Google.
What the exercise gives you
At the end, you have a ranked list of profiles, a triggering situation for each, and a set of real questions phrased in your clients' language.
That is no longer an intuition about your market. It is an editorial plan, in which every question is a piece of content to produce, and the priority order tells you which to write first.
It is also, incidentally, the best way to discover that the client you have been targeting for two years is not the one paying your bills.