Defining your ideal customer before you have any customers
The usual advice is to start from your current customers. With none, most people invent a persona. There is a third route, and it takes a week.
Every guide says the same thing: to define your ideal customer, look at your best customers. It is excellent advice, and useless to the person who needs it most.
With nothing sold yet, two options appear to remain. Wait, and prospect blind in the meantime. Or invent a persona, with a first name, an age and hobbies, describing very precisely someone who does not exist.
An invented persona is not a hypothesis, it is decoration. A hypothesis is phrased so it can be proven wrong.
What replaces customers
What you lack is customers, not conversations. And it is the conversation that produces the material, not the transaction. Five to ten exchanges with people living the problem are worth more than a positioning workshop, and cost a week.
The difficulty is not getting those conversations, it is not turning them into demos. Someone willing to describe their situation stops the moment you start selling.
| What you do not have | What replaces it |
|---|---|
| Signed customers | Five to ten conversations about the situation |
| A sales history | What they already tried and abandoned |
| Objections heard | The reasons they did nothing about it |
| A customer vocabulary | Their words, written down as spoken |
The second row is the most underrated. What someone already tried, and why they stopped, says more about their problem than any description. It is also what your offer will be compared against.
Where to start, concretely
1. Write a one-sentence hypothesis
Not a profile sheet, a sentence, and one that could turn out false. Independents who sell over the phone lose deals because they never read their conversations back. That can be disproved in five conversations, and that is all that is asked of it.
2. Find ten people who could disprove it
Not ten kind people, ten concerned ones. Your network already holds some, and it is the only place you will get an answer quickly. Ask for a twenty-minute conversation about how they work, with nothing proposed.
3. Listen for the situation, not the opinion
Would you pay for this? is worthless, everyone says yes out of politeness. How do you do it today? and tell me about the last time are worth everything, because they make people describe facts.
4. Record the words, not your summary
This is the point almost nobody holds. Rewording in your own language destroys exactly what you came for. Keep the sentences as spoken; they will serve later in your offer and your messages.
What separates a hypothesis from a persona
The invented persona
- A first name, an age, a photo
- Hobbies that serve no purpose
- Not one line that could turn out false
- Written once, never reopened
The testable hypothesis
- A situation, a trigger, a consequence
- Criteria settled by yes or no
- Lines that can be disproved
- Corrected after every conversation
The difference is the last point. A persona is a deliverable, a hypothesis is a temporary state. If your avatar has not changed after ten conversations, it is not that it was right, it is that you never confronted it.
When to stop listening
There is no magic number, there is a signal: when you can predict what the person is about to say before they say it, you have listened enough. It often arrives around the eighth conversation, sometimes earlier.
Conversely, if every exchange still teaches you something new after fifteen, that is a bad sign: you are probably talking to people who are too different, and your hypothesis is too broad to be tested.
And the day you sign your first customers, none of this work becomes obsolete. It becomes comparable: you hold your hypothesis against what the people who actually paid describe, and the gap between the two is the most valuable information of your first six months.
Where Meidly helps
You do not need a history to begin. From the very first recorded conversation we pull out the situation described, what has already been tried, the objections and the exact phrasings. Your avatar builds itself during the research phase, and gets corrected with every new call instead of being written once.
See how the client avatar works→In short
Do not wait for customers, and do not invent a persona. Write a sentence that could be wrong, take it to ten concerned people, make them describe facts rather than opinions, and keep their words. In a week you have more than a workshop would give you, and above all something that can be corrected.
Frequently asked questions
How do you define your ideal customer with no customers?
By replacing customers with conversations. Write a one-sentence hypothesis that could turn out false, take it to five or ten concerned people, and make them describe how they work rather than give an opinion on your idea.
How many conversations does validating an avatar take?
There is no fixed number, there is a signal: when you predict what the person is about to say before they say it, you have listened enough. It often lands around the eighth. If everything is still new after fifteen, your hypothesis is too broad.
Is a marketing persona useful at the start?
Rarely. An invented persona with a first name, an age and hobbies contains no line that could turn out false, so it can teach you nothing. A hypothesis phrased to be disproved fills the same role and actually progresses.
What questions should you ask in those conversations?
Questions that make people describe facts: how do you do it today, tell me about the last time, what have you already tried and why did you stop. Avoid would you pay for this, which everyone answers yes to out of politeness.
What do you do with this work once the first customers sign?
Compare it. Hold your starting hypothesis against what the people who actually paid describe. The gap between the two is the most useful information of your first months, and it points precisely at what you had misunderstood.