🤖 How well do you know your customer?
🤖 How well do you know your customer?
[FREE ISSUE] The strongest early signal of startup success is how well a founder knows their customer. So let's create a customer persona — and have our AI co-founder test how well we know them.
Hey friend 👋
If you don’t know your customer, you’re f*cked.
Knowing your customer is the first step in finding product-market fit, and it’s the single greatest early predictor of startup success.
Not to mention that 70% of buying experiences are based on how the customer feels they are being understood.
Unfortunately, you just can’t skip it and live to tell the tale.
Our tool of choice? Customer personas.
Let’s go 👇
Who’s your customer?
This is a slippery question, because it seems obvious.
It’s not.
If you’re a pre-seed founder, you’re looking for traction with two hands and a flashlight.
Yet if you ask one about who they’re trying to get traction with, you get vague, broad answers about massive markets filled with generic, bland customers.
Which is why I can predict with 90% accuracy something about your startup, even if we’ve never met:
Your customer isn’t specific enough.
Here’s a quick example. Let’s say you are serving high net worth men born in 1948 in the greater London area, who are interested in and own multiple dogs, and who take their vacations in the Swiss Alps.
That sounds awfully specific!
And yet… that describes both Ozzy Osbourne and King Charles.
What are the chances the same messages and value prop will appeal to both of them?
Not high!
You have to get even more specific than that.
Remember: startups begin and end with the customer
When doing anything innovative, the most important task is proving that the idea is desirable.
It doesn’t matter if your business is theoretically viable or or technically feasible if no one wants it.
And as we move toward defining a viable, feasible business model, we need to keep in mind that the customer sits at the centre. As we pivot and iterate around to change What, How, or Why, then the Who must necessarily change:
Changing how value is delivered? Probably a different customer shopping in that way.
Changing how or what money is charged? Probably a different customer that can afford that, or would want to pay in that fashion.
Changing what value you’re creating for the customer? Different customers care about different values.
Change anything, and the customer necessarily changes. That’s why we say we co-create the product with the customer.
The best tool for understanding your customers is a persona.
First, let’s talk about what we mean:
A customer persona is a semi-fictional representation of your ideal customer.
It’s about your ideal customer in your niche, because the only customer we care about is the specific one we want to convert right now — today.
It’s semi-fictional because, while it may not refer to a real person, it needs to be real enough that we can interrogate it as if it were a real person.
It looks something like this:
We give our persona a friendly name. Culturally, the startup community usually comes with with something alliterative and easy to remember:
Founder Fae
Restaurateur Ricky
Data Deb
Laundromat Larry
This is more important than it first seems, because one of our principal goals is to develop empathy, and the more real a persona feels when we talk about them, the easier it is to develop that empathy.
That’s just human nature.
Plus — friendly names just make personas easier to talk about.
We then define some demographic information about the person, give them a robust and believable backstory, and begin to understand their personality.
And then we dive deeper.
Much deeper.
We understand in depth what their interests are, who influences them, what their motivations and goals are, what they want, need, and fear, and much more.
If you want a more robust overview, I posted a deep dive on YouTube last year:
What tool should I use to create a persona?
My philosophy is always the same — the tool that actually gets used is the right tool for the job!
Honestly, it doesn’t matter… Notion, Google Docs, Figjam, etc — use what you’re already comfortable with.
But if you want useful starting point, you can copy the template we use at The Right Box:
It’s a FigJam file, which is like a digital whiteboard that you can co-edit with your team.
Use whatever you like, but...
Avoid these 3 common persona mistakes
Before we move on to the AI, let’s talk about three common mistakes that nearly every founder makes.
And they cost them tons of time and money.
Mistake 1. Make it about your product.
The first mistake nearly every founder makes is to create a persona through the lens of your product.
But this isn’t about your product — it’s about the people who have the problem you’re solving.
If you ignore the context in which their problem sits, you’ll get a distorted view of its urgency and severity, leading to the slow startup suicide, you’ll miss out on massive (and often unexpected) opportunities, and you’ll accidentally miss the things customers don’t want changed.
So create a persona as if you have no idea what product you’ll even be building.
Suppress your internal critic.
Capture everything:
You don’t know what’s relevant until you start hearing patterns from the customers you talk to.
Mistake 2. Thinking you are your customer.
If you’re starting a business or a startup within your field of expertise, don’t fall into this trap.
A lot of startups start this way. You’re working in a field, you encounter a problem, it sucks, you figure other people have it too, you decide to solve it, and you launch a startup.
And that’s fantastic!
But it can lead you to conclude that you know who the customer is.
You don’t.
If you really don’t think your customer is different from you, let me ask you this:
Why didn’t they launch a startup to try to solve this problem?
Because you’re different!
And you don’t know all the ways they are different from you until you start learning about them.
And sometimes, you find they don’t even exist. 😅
To understand who wants what you’re selling — and, at the same, what you can sell to whom you’re talking — you need to have a deep understanding of the person to whom you’re talking.
Mistake 3. Lack of specificity.
We keep talking about being specific, so... how specific should you get?
Honestly? You want to get so specific that it makes you uncomfortable.
And then you want to get two clicks more specific.
The riches are in the niches. Specific wins. You can always expand later.
To paraphrase Frederick the Great: to market to everyone is to market to no one.
If you don’t know the specifics, that’s ok! Make some educated guesses. You’re going to be wrong, and that’s ok — you’re just picking a starting hypothesis that you can prove wrong.
That’s it, bring out the AI!
This prompt takes X and provides Y.
Here's the prompt in action:
Ok get ready, because this prompt is pretty gnarly, and one of the longest I've written in a while. It could be shorter, but I didn't spend the effort, because I frankly liked the output I got. As Mark Twain quipped:
If I had more time, I would have written a shorter letter.
Anyway, as usual, this prompt should work in both free and paid ChatGPT, as well as in Claude and Gemini. My screenshots are from ChatGPT 4o.
NOTE: This is ordinarily where the paywall would kick in, with the prompts available only to paid to subscribers. But I wanted to give everyone a taste of the awesome power of AI, and hopefully inspire founders to write prompts of their own. If you find this valuable, perhaps you can share it with a friend?
Here it is (and it’s a doozy):
# ROLE
You are an experienced startup founder with a deep background in customer discovery and market research.
I am the founder of an early-stage venture-backable startup creating customer personas.
# TASK
I am going to give you a Name for my customer persona and a brief description of who they are, and you’re going to quiz me about the details I should know.
A customer persona is semi-fictional representation of my ideal customer. It does not reflect a real customer. It’s an avatar with the same level of specificity as a real person that I can use to help understand my target market and make better decisions.
I should understand on a deep and personal level various aspects of their lives. See the list of things I should know broad categories.
You will ask me one question at a time, and then pause for my answer. You will then evaluate the clarity of my response and the depth.
My answers should be:
- Precise, without giving generic or vague responses. I should know the customer better than they know themselves.
- Detailed, with as many of the who, how, what, when, where, and why questions as possible known.
- Falsifiable, in that I should be able to conduct an experiment or customer interview to find out if I’m right
- Specific to this one person, rather than general to a group of people
- Logically consistent with my prior answers
It is not your job to determine if what I am saying about the customer is correct. Your job is to ensure that my responses meet the above tests.
When I ask for another question, provide me another question. Vary the questions you ask both on their level of specificity and on the topic to which they pertain.
# EXAMPLES
If I were to give you a persona named Entrepreneur Ernie who is the founder of an early-stage B2B SaaS startup, you might ask questions like these:
- Who are the top 3 people Ernie listens to for advice on running a startup?
- What is the most important thing in Ernie’s life?
- What keeps Ernie up at night?
- What tools does Ernie use to keep himself and his startup work organized?
If you asked me this question:
> What keeps Ernie up at night?
And my response was:
> Not finding customers.
Your evaluation of my sophistication in knowing the Ernie persona might be:
> Startup founders often worry about not finding customers, but why is Ernie worried about that specifically? Is he worried about a sales problem or a product problem? How is he trying to solve that now?
The conversation would continue back and forth until I asked for the next question.
# THINGS I SHOULD KNOW
Demographics:
- Age
- Occupation
- Income
- Location
- Education
- Social status
- Biography & background
Psychographics:
- Interests
- Influences
- Intrinsic motivation & goals
- Needs and “jobs to be done”
- Wants
- Fears
- Problems, pains, & frustrations
- Priorities
Personality traits:
- Introvert vs extrovert
- Analytical vs creative
- Busy vs time rich
- Messy vs organized
- Independent vs team player
- Passive vs active
- Risk-averse vs risk-tolerant
# CUSTOMER
PERSONA NAME = {name, like Entrepreneur Ernie}
DESCRIPTION = {1 or 2 sentence description}Let’s take it out for a spin:
To kick the tires, I gave it this information:
PERSONA NAME = Restaurateur Ricky
DESCRIPTION = the owner of a local single-location fine dining restaurant
And my AI co-founder wasted no time!
Ooh! Interesting start.
Let’s try to be vague and see what happens:
Oh, dang — slapped in the face by the AI. For shame!
I decided to play along and give a better answer, which ChatGPT quickly accepted (along with an explanation of why) and moved on:
Pretty good stuff!
Later in the chat, it asked me another question:
Who are the top three people Ricky relies on for advice when it comes to running his fine dining restaurant, and what specific roles do they play in his decision-making process?
To further test things out, I told it I didn’t know the answer.
Honestly?
Your mileage may vary. I have no idea what your AI co-founder will ask you.
But I guarantee it will get you thinking, and that’s what matters.
And remember — you don’t know what’s important yet, so don’t be dismissive.
See you next week,
—jdm
PS: If you enjoyed this issue of Founding with AI, I’d be super grateful if you took a few seconds to share it with a founder friend. Referrals are a huge help in growing a newsletter — and a huge help in helping more founders realise the awesome power of AI!
And, as always, I’d love to hear your thoughts. If you have ideas how to make this newsletter even better, have suggestions on future topics, or even just want to share what you thought of this issue, just hit reply or leave a comment.