How To Build A Startup
From picking an idea to building a company worth millions from someone who's done it twice
“How do you get a startup idea?”
“How do you learn to code and set up servers?”
“How do you do marketing and sales to get first customers?”
Many people want to found their own startup— but don’t know where to start.
In this short guide, we’ll answer every detail you can imagine...
From big concepts to the small details of setting up your servers and what to say to close your first deals.
And the best part is— building a startup will cost you nothing but your time.
1. How Do You Get An Idea? (Day 1)
Everyone loves to talk about startup “ideas”.
But, customers don’t pay for “ideas”— they pay for problems you solve for them.
So find a problem you’re capable of solving.
First ask yourself: “Do I have any problems I pay, or would pay, to solve?”
Then, ask your friends: “Are there any problems you pay, or would pay, to solve?”
Still have no good answers? Use another method...
Find software you already pay for or know about, and find out what market they aren’t serving.
Then make a better solution for a small specific market!
A great way to do this is by visiting review sites like G2, Capterra, or Reddit and seeing complaints people have about specific companies.
An example of this is Toast.
Stripe already had payment software and payment systems, but Toast made a system specifically for restaurants.
Toast is now worth over 12 Billion dollars.
Now you should have an idea/problem to start with!
Let’s say for example you speak Chinese and know a lot about writing.
Most current AI blog writing startups do not work with Chinese.
So, you can be the AI blog writing startup specifically for Chinese companies!
Don’t worry if your initial idea isn’t that good, while building your startup you’re bound to run into more problems yourself.
Okay, now that you have your idea… it’s time to start building!
2. Designing Your Product (Week 1)
Now that you have your idea, it’s time to spend 100+ hours building your product!
Sike! Just kidding!
Your time is valuable, so don’t spend 100+ hours building a product that no one may care about.
Instead, build a beautiful landing page in an hour with Unicorn Platform, Framer, or Magic if you are technical.
Then design a mock-up in under 20 hours with Figma, or if you’re technical use a UI library like Shadcn!
It’s easy.
Go and download an amazing basic template for free from Figma’s community marketplace and change it up a little bit.
Make it beautiful and make navigation work.
Then record a quick demo of your “product” using Screen Studio, OBS, or Loom.
It will look like a real product… except you didn’t spend 100+ hours writing the code!
Lastly, link your landing page website to Stripe, and add a 3-day free or paid trial— but only after customers enter their credit card!
A good starting price is ~30% less than the industry leader, and try to choose an idea you can charge at least $39/month for.
In the 100+ hours it would’ve taken you to build a single product, you can design five ideas for absolutely free!
3. Proving Your Idea (Week 2)
So you have your website and your demo live!
Now comes the hardest part, getting people to visit your website, enter their credit card details, and click “start trial”.
Where do you find customers?
Find when people are complaining about the problem you solve and message them.
Platforms like Fiverr and Upwork can also be a goldmine for finding people who offer complimentary services.
Let’s build on our earlier example: You have an AI startup that generates Chinese blogs.
Go find people who build websites for Chinese companies and message them:
“Hey ____ 👋 saw you create websites for Chinese companies. ____ helps Chinese companies generate blog posts with AI! Would you be interested in referring your clients? We would pay you for every person you refer! $$$ ”
Then send your website link and demo video.
Want to go more direct?
Find potential customers doing manual research and then use Seamless.ai, or Apollo.io to find the contact information of 1,000+ potential customers.
Then build a cold outreach sequence using tools like Reply.io, Close, Clay, or Salesloft
Email potential customers, message them on LinkedIn, and most importantly— call them.
Using these strategies, you will get 100+ responses within a week.
People will either sign up and enter their credit card details— or they won’t.
If they don’t sign up, it is for one of 3 reasons:
They don’t actually have the problem.
They aren’t willing to pay to solve the problem.
They don’t think your product is good enough to try and pay for.
Do not quit until you get 100 responses.
After getting 100 responses, you will know the flaws in your thinking.
Either people have signed up or they haven’t.
If they don’t have the problem— start again with a new idea!
If they aren’t willing to pay to solve the problem— spend 2 hours making your website and demo better by “adding” new features, or restart with a new idea!
If they don’t think your product is good enough to try and pay for, spend 2 hours making your website and demo better by “adding” new features!
Most of your ideas will not work.
So, don’t get stuck “adding” new features, or features you can’t build.
The right choice is usually to start over with a new problem.
Okay, now your first idea either worked out and people signed up, or you retried 2-10 times until people finally entered their credit card information and clicked “sign up”!
Congratulations!
But wait…. you don’t have a real product!
4. Building Your Product(Weeks 3 & 4)
Okay, now that you have an idea that people are interested in and willing to pay for— it’s time to spend two weeks building your product.
Set a 14-day deadline and stick to it— no matter what.
“But I don’t know how to code!”
Go to Vercel.com, open ChatGPT, and download Cursor.
Vercel is a company where you can download a template to build software, and then deploy that software to your website with one click for free.
They handle everything from the servers to uploading your code.
Don’t understand something? Ask ChatGPT, Cursor, or any AI chatbot.
Use Javascript or Typescript for your website and PostgreSQL for your database.
Javascript/Typescript are not beautiful or fast programming languages, but they are the most popular— think of them as the English of the programming world.
Everyone “speaks” and “writes” Javascript/Typescript because they are popular, and most “books”(websites) are written in them.
Use a template library like Chakra, Tremor, Daisy, or Shadcn… and make your code match your Figma design from your demo as much as possible.
For your backend functions, use external APIs or scrape websites with ScrapingAnt, Smart Proxy, or ScrapingBee.
For a database use Supabase or Xano!
But don’t strive for perfection— work fast and focus on 1-2 features.
It will be a two-week grind, but stay positive.
After two weeks have passed you will have a basic first version done!
We call this an MVP, or “minimum viable product”.
It’s not pretty— but it will work.
Now it’s time to get your first real paying customers.
5. Your First Customers (Week 5)
Your first customers will be hard— but the good news is you have done this already!
The only difference is you have a real product now, besides this difference, just follow the same playbook from step #3!
Grind out and don’t stop until you get 20+ signups (regardless of whether they end up paying after their trial ends).
Your goal here is to get customers on calls to get feedback.
When you speak to people your goal is not to talk their face off.
Make some small talk: “Tell me about yourself/your business”!
Then ask these 3 questions:
What made you sign up for a trial?
Why did(or didn’t!) you pay after your trial?
What do you wish our product did better or had?
Then shut up and listen while taking notes.
6. Implementing Feedback (Week 6)
Okay, now you’ve had 20+ signups and have spoken with 10+ customers…
It’s time to implement their feedback.
By now you have a long list of requests.
So look for trends in your requests list and use your intuition.
What is easy to build and takes a week to add?
Spend a week and add these small improvements.
Spend a week— no more, no less.
7. Repeat(Weeks 6 to 8)
There is no magic sauce, there is no silver bullet.
It’s time to reach out to people who have canceled their subscriptions and reach out to new potential customers again.
You will now spend every waking hour improving your app and getting revenue.
Get new customers. Talk to customers. Make small improvements. Repeat.
8. A Stable Customer Base (Months 2 to 4)
A major problem that you will encounter is churn.
Churn is just a fancy word for “What percent of customers cancel their subscription every month?”
At first, churn will be high, 10-20% of your customers will cancel each month.
This is normal.
You must discover: Why do people cancel?
Then fix your product accordingly.
There are two reasons why churn happens:
Your product needs improvement.
You are targeting the wrong customers.
Taking our earlier example…
Let’s say your AI Chinese blog startup has 20% of customers cancel each month!
In other words, the problem is not your product— it’s who you are selling to.
Try selling to blog writers, and your churn may be much lower.
You need to get your revenue to $2,000/month AND churn below 5%.
This phase is brutal and for complex products can take months— we call it in “The Loop”.
Once you have at least $2,000/month in subscriptions and get churn under 5% per month— it’s time to scale.
9. Automating Sales (Month 4-6)
Now you have $2,000/month in subscriptions and under 5% of your customers are canceling each month!
You’re doing great!
Until now, you’ve got all your customers by simply reaching out and speaking with them.
So should you try to scale with Google and Facebook ads? Right?
Hell no.
“If it isn’t broken, don’t fix it” is great advice.
What you are doing is working, you just need to automate it.
Use software such as Apollo or Seamless to find potential customers, MillionVerifier or Clay to confirm email addresses, and use reply.io or instantly.ai to automate your outreach!
If you want to really invest in a system that’s working, get Outreach or Salesloft.
Your goal is to do exactly what you’ve been doing— but use software to automate your processes.
You should now also be “marketing”.
Posting on LinkedIn, hitting Facebook Groups, launching on Twitter, etc.
Your goal shouldn’t be to go to $1M+ in subscriptions next month.
Your goal should be slow and steady progress on your metrics.
Increasing revenue, fewer cancelations, and happier customers.
This is also the stage where you can(and maybe should) raise a little money!
10. Hard Scaling (Months 6+)
Now you will have $10,000+ in monthly subscriptions!
You are doing amazing!
You’ve now probably raised or are raising big money.
You now know people love your product, now it’s just getting them to sign up.
“But how do I scale?” you may ask.
Always chase the answers to these two questions:
How can you expand the number of people willing to pay for your product?
How can you charge people more?
This will lead you to develop new sales strategies, features, and products!
At this stage, there is no easy answer for how to scale— it depends on your business.
You need to know:
Who are your customers?
Where are your customers?
How much do you charge customers?
Different companies will have different answers.
If you’re selling a low-ticket product, you should probably be doing marketing and ads.
If you’re selling a product that costs on the expensive side, you should probably be doing outreach and eventually hiring salespeople.
When you first reach this stage, you should consider changing from a “cheap product” to a premium product.
Add a killer feature, do a beautiful redesign, and update your pricing structure to get:
Higher paying customers
Lower churn customers
Better sales economics
Instead of “30% less than the industry leader”, aim for 2-5x their pricing.
Why?
Selling a product for $499/month isn’t much more work than selling it at $199/mo.
The difference in work is marginal for 2.5x the revenue.
You should make your product nicer and go premium as fast as possible.
But remember, always remain strict with your spending.
Conclusion & Disclamer
Building something new is extremely hard.
Honesty, you probably don’t have what it takes and should get a cushy tech job instead.
Or even better, work at early-stage startups!
Risk-adjusted, early employees make the most money.
As an early-employees:
You get less stress
You get paid a six-figure salary
You get equity that could be worth tens of millions
You get freedom. You are not tied down to one startup and can bet on 10+ startups
At a $1 billion-dollar startup, normally:
The two Founders make $200M each
The first 10 employees make $5+ million each
The next 10 employees make $2-3 million each
The next 10 employees make $1-2 million each
At each unicorn 30+ employees become millionaires, make a six-figure salary, and guarantee a $250k/year salary for any future role.
It’s way better to be the first 30 employees than to be a Founder.
But if you’re really set on being a Founder, know what you’re getting yourself into.
You will probably be a loser for years— possibly decades.
Otherwise, you will not make it.
All that said, Founding is not complicated.
Find a problem you’re passionate about, find people willing to pay for your solution, and you have a startup.
I hope this playbook helps you and wish you the best of luck conquering the world.
Shoot me an email— I’d love to help you out and maybe invest in you :)
emilianoguerrero22@gmail.com