When I launched Polsia in December 2025, it was crickets.
Nobody knew the product existed.
Today, over 25,000 people pay for it.
But every stage between those two numbers needed a different playbook.
And if you run them in the wrong order, you waste time and money.
Here’s every step, in order:
↓↓↓
The idea
Before you validate anything, you need to know the idea is worth building.
My test is simple.
You have the idea, and the next morning you wake up still thinking it’s good.
Then for 5-7 days afterwards, you keep thinking it’s good.
If it passes that test, it’s probably an idea worth building.
But then you need to decide if you’re willing to dedicate the next 5 years to it.
Everyone’s answer is different.
The validation
You can validate an idea before you build a single thing.
Create a landing page with an email waitlist and a Typeform.
Then run Meta ads and send that traffic to the page.
What you’re watching for is whether a stranger who sees your ad will leave their email, finish the entire form, and tell you why they’re interested.
That gives you a lot of signal on whether you should build it.
So if you can’t get anyone to do that, move to the next idea.
This is how I validated Polsia.
My Typeform asked people how much autonomy they’d want from an AI, and how much they’d be willing to pay for it.
The responses were amazing.
The first 10 customers
I learned how to get first customers from Travis Kalanick at CloudKitchens.
I had built a service to help restaurants with their digital marketing, and I told Travis there were millions of restaurants who could use it.
Then I asked him if I could hire a salesperson.
He said. “Trust me, you don’t want to scale something that doesn’t work. You’re the salesperson. Go sign up 10 restaurants and prove it works.”
So I started calling restaurants myself.
Including ones already operating out of our kitchens.
A week later, I came back with 10 restaurants, and he let me hire a salesperson.
The lesson?
Your first 10 paying customers come from hustlin’.
The signal
So how many customers do you need before you’ve found product-market-fit?
1 paying customer is a good start.
10 means you’re on the right track.
Then once you’re at 100 active, happy, paying customers who stay for 2-3 months, you probably found product-market-fit.
But even that’s not true product-market-fit.
To me, you’ve found it when you can profitably grow a use case at scale, and whatever that looks like depends on your business model.
The product
Between 10-100 customers, focus on improving the product.
Knowing how to get customers is a good skill, but scaling something that doesn’t work is a waste of time and money.
Wait until you understand your customer and know you’re adding value.
That’s why I barely touched my pre-seed before I found product-market-fit.
Having capital doesn’t mean you should spend it.
It means you should reserve it for the moment you’re ready to scale.
The biggest product fix I made was onboarding.
The first version of Polsia was email-based.
You’d give it your email and get a daily update on the work done, which meant the wow moment was delayed until they checked their emails.
I realized people had to see the experience first.
So now, in the first few minutes of onboarding, Polsia researches your idea and starts building the company in front of you.
Only then does it ask you to start a free trial.
The faster your user hits the wow moment, the more users will pay.
The price
Charge enough to recoup your money at some point and be profitable.
When I priced Polsia, I added up what one customer cost me.
The AI runs on task every night, which came out to about $30 in AI costs.
On top of that, a web server with a database cost me $5 to $10 a month, and I embed API keys so customers never have to think about them.
That got me to roughly $50.
So I charged $49 a month plus 20% of the revenue their business made.
I roughly broke even on the subscription and made money when they made money.
You can also be smart about where your costs are heading.
I bet AI costs would come down, and once I cut my AI bill by running open-source models on rented GPUs, I could afford to offer a free version.
Some companies right now don’t charge anything at all.
But unless you’re really viral, I think that’s pretty irresponsible.
Because eventually you will have to charge and free users might leave.
The scale
Before product-market-fit, hustling on X and online is enough.
Once you have a business model and retention, hit every channel, all the time.
When I had a couple hundred users, I knew it was time to put fuel on the fire so I moved to SF and did my marketing stunt…
Having Polsia raise my $30 million round.
Nothing beats putting effort into marketing and product, it compounds.
The conclusion
A great product with no marketing gets you zero users, or very slow growth.
Great storytelling with a bad product gets you attention, but no growth.
And a great product with great storytelling will make you successful.
This process gets harder at every stage.
But if you’re consistent and follow this guide, 25,000 customers is inevitable.
Try Polsia here: polsia.com
