The $1 Android Game (2015, 18 years old)

I build an Android game app that has 500 downloads. It earns 1€ from advertising. The game looks extremely ugly and has performance issues.

Don’t be afraid of launching.

I was trying to perfect every aspect of the game. In reality, I was delaying the launch because I feared no one would download the app.

Many entrepreneurs, especially those who are more technical, make this mistake. It’s painful that your work gets ignored, it’s easier to keep coding and have this fake feeling that you are making progress.

In fact, there are plenty of users who are willing to try new stuff and give genuine feedback.

Commit to the project or kill it.

“Very likely it will succeed, but it has to be dramatically improved: adding new levels, etc”. This was the review of one of the users.

One could think of this as a great fuel of motivation, but for me, it had the opposite effect. Creating the game had been 50% of the time, the other 50% had been related to fixing endless issues making the app responsive to any device.

This project was no longer fun. Most importantly, I wasn't learning anything new. I wanted to move to something else.

Forex Bot Regret (2016, 19 years old)

Biggest regret as a side project. I discover Forex and lose dozens of hours looking at charts trying to find inexistent patterns.

I develop a trading indicator that gives buying and selling signals and a trading bot that executes orders on those signals. It has consistently negative performance.

In day trading the one making the real money is the broker.

The fact that everyone is selling trading bots makes clear that no one is making money trading and they are trying to make money selling software (why would anyone sell a money printing machine?).

The same logic applies to the guys selling $10k per day trading strategies.  

Develop an unfair advantage.

By working on personal projects in my free time, I developed a strong coding foundation that gave me an edge both at university and in the tech world, especially when dealing with non-technical business people.

The key is finding something you’re both good at and genuinely enjoy—whether that’s coding, public speaking, networking, etc.—and focusing on it intensely.

Over time, the countless hours you invest will create a skills gap between you and others, one that becomes increasingly difficult for them to close.

The $700 Instagram Exit (2018, 21 years old)

I grow an Instagram motivational account to 60k followers and sell it for 700$.

Follower Quality > Follower Quantity. 

I focused on growth and ended up with an audience I couldn’t truly define. I just knew that 90% of the followers lived in low-income countries and brands were not interested in promoting their products on my account.

Having a big follower base is pointless if you don’t feel connected to them. Without a clear ICP, I produced the same generic content as everyone else, and eventually, it felt like nobody was reading, so I lost the motivation to post

$1, $5, $500. Invest money in the business, or it won’t grow.

I spent two months posting daily and gain only 60 followers, while accounts with 100K+ followers are using paid growth services.

I decide to spend $15 monthly on such a service, a significant expense for a zero-income student. Over time, I start gaining traction, and eventually, my following grows by over 500 people per day.

Great 3rd party product + domain authority = Affiliate marketing works.

The growth service offers an affiliate program that is easy to sell because it works, and my 60k-follower account gives me the credibility to recommend it.

When people ask how I grew, I direct them to the program, earning around €30 in MRR. Still, with one hour of daily work translating to roughly $1/hour, it isn’t worth my time, so I decide to sell.

The Illegal Amazon Review Marketplace (2020, 23 years old)

My mum leaves a negative review on an Amazon product. Sellers decide to reimburse all the money and she can keep the product.

I come across a market for selling reviews through Facebook groups where Chinese sellers are trying to break into the European market. The process is simple: you buy a product, leave a positive review, and then get reimbursed, essentially getting the item for free.

I go for it and spend 2 months building a marketplace to buy reviews in exchange for free products.

At some point, I find out that review incentivization is illegal within Amazon policies. I get infinitely demotivated and stop the project.

Strike the right balance when doing idea assessment.

In this project, I have to admit I was too excited about the idea of building a marketplace and learning about WordPress.

Looking back, a minimum of research would have shown it was a no-go. Having said this, it’s easy to fall into the analysis paralysis mode and find endless reasons why an idea might fail.

There will always be red and orange flags. it’s about learning to differentiate between them.

Clear red flags: illegal (it will make everything complicated, ads services will not promote the service, most importantly, you can get into legal issues), or relying 100% on another service (more on this in the next project)

If there’s competition, it’s good, if they are making money it’s even better.

In the beginning, I was thrilled when I saw no competition for my “unique idea”. Later, I discovered the obvious reason.

More often than not, businesses fail due to a lack of product-market fit. Sure, someone has to be the first to enter a profitable market, but just like winning the lottery, the odds that you are the chosen one are low. If no one else is doing it, there's probably no market.

Copying a “Proven” Business Model (2020, 23 years old)

Tired of trying to invent something new, I copy the Instagram growth service I’d used years before. The original worked for Asian accounts, so I thought it would also work in Europe.

I build it, and get 20 users, but nobody pays. Turns out, Instagram has changed its algorithm, and the old “you comment on my post, I’ll comment on yours” trick doesn’t work anymore.

2 months after starting, a friend joins the project.

Consider future project synergies before selling

When I sold my old Instagram account, I didn’t realize it was like throwing away a resume. Now potential clients asked: “Why should I trust you to grow my account if yours has no followers?”. I feel like those guys selling trading algorithms.

At this point, I have no money to purchase an account, and growing a new one to a few thousand followers will take me months.

Had I never sold the account, I would have onboarded more users faster.

Fall in love with the problem and hate your MVP.

The more I manage to see automation progress the more I like coding, even if users are hardly using the product.

I get drown in developing features that are unnecessary, but that the original service has (telegram bot with multiple scenarios handling, affiliate portal, etc.). Passion for building completely blinds me: the solution is not working and I am too attached to it that it feels painful to throw it away.

When building an MVP, avoid spending too much time on it. The longer you invest, the harder it becomes to pivot or start over due to the sunk cost.

Do not build a business that depends 100% on another business, it is too risky.

Mr. Musk can increase Twitter on API pricing to $42,000 monthly without notice and Tik Tok can be banned in the US.

In our case, Instagram changed the algorithm from one day to the other, and the “comment per comment” strategy stopped working.

We had built a product that was not useful, and worse, now we had no idea how to grow an IG account.

NFT Marathon Medals: Hype Couldn’t Cross the Finish Line" (2021, 24 years old)

My flatmate invests his scholar loan into flipping NFTs. The way he speaks about the technology makes me very curious. I spent 3 days figuring out how NFTs work.

At that time I had completed my 3rd marathon and I felt disappointed with the medal received. I create a marketplace with WordPress and start selling medal NFTs to runners.

I convince 2 non-profit race organizers to sell NFTs. I sell 20 NFTs for 5€ each, still, only 3 people claim the NFT.

Market timing is crucial.

Despite the NFT hype in the tech world, the runner community does not understand them. I spend 95% of the meeting educating prospects about what the heck is blockchain and NFTs.

Society is not ready to adopt blockchain (it is not even prepared today). Race organizers do not know what they are selling, and runners do not know what they are buying.

I find myself guiding a 68-year-old man setting up a crypto wallet and telling him that if he loses his private key he will lose his medal. Of course, he loses the private key and accuses me of stealing his medal.

Hype is not enough to make a business. Not in the NFT era, not in the AI era. If there’s no real value, there’s no business in the long term (even if you have paying customers)

The 30-day rule in Fanatical Prospecting.

I read a great book about sales called “Fanatical Prospecting” by Jeb Blount. On it Jeb, exposes the 30-day rule in Fanatical Prospecting: any prospecting you do in a 30-day window will pay off over the next 90 days.

I send over 300 cold emails and land over 20 meetings between January and February, 2022. The 2 clients get closed in April and May.

While serving the clients, I can’t keep up with the outreach. I do not prospect in March or April, and by the end of May, I had to start my sales funnel from scratch. I kill the project.

How Co-Founder Misalignment Killed Our AI Portal (2023, 26 years old)

ChatGPT is released. My previous partner is bullish on the potential of AI in SMEs and we decide to launch a portal to help businesses identify relevant AI use cases within their industry. 

At some point we start having discrepancies about vision, what we want to become, and execution, how we will reach there. We decide to stop the project.

Today there’s still 1 daily user that gets registered in the platform.

NextJS + Vercel + Supabase: Great stack to create a SaaS MVP.

The initial learning curve is steep, but once you get familiarized with these tools you can leverage AI and increase your development speed 10x.

Suggestion: do not use AI to create a project using these frameworks unless you know how they work. AI can speed up development, but without deep understanding, you’ll drown in bugs that neither the AI nor you can fix.

SEO is king, but the future is SAO.

One of our users creates a use case on “Changing Song Lyrics with AI.” Not being our target use case, it brings 90% of our traffic.

I've always focused on paid acquisition channels like Google Ads for their immediacy and clear logic: more money = more leads. But this small signal changed my SEO view. A well-executed long-term SEO strategy can far outperform paid campaigns.

That said, the SEO world is undergoing a major shift with the demand increase of AI-generated responses.

Search Agent Optimization will be key: making your service the preferred choice for agents. Brands that establish relevance with agents will attract new customers and retain existing ones.

No Co-founder alignment, no business.

When starting a business define roles as clearly as possible. Person 1 will be responsible for A/B, Person 2 will be responsible for C/D. And make these responsibilities mutually exclusive.

In this project our biggest strength ended up killing us. Both founders had strong strategic skills and we were constantly arguing about decisions.

This is not about one person having all the power, but about making sure someone is responsible for making the final call in their area. Not every choice will be perfect but that’s okay.

The worst approach is trying to agree on every little decision. It kills energy, slows things down, and often leads to the same or worse results than if one person had just decided and moved forward. 

50% of 0$ is 0$.

Splitting equity 50/50 assumes either that both partners contribute equally (which is unlikely) or that they’re avoiding a tough conversation about fairness (which will come up later, especially in tough times).

Among others, the equity should reflect three key factors: Workload (who will carry the biggest long-term responsibility), Unique Value (whose skills are hardest to replace), and Crisis Management (when things go wrong, who is the go-to person).

Building an AI tool in 3 Weeks and getting ghosted (2024, 27 years old)

An SEO agency owner shares a recurring client pain point: multi-brand e-commerce teams are manually rewriting thousands of product descriptions from retailer catalogs for SEO.

I get connected with one of his customers to automate this. Clear problem, paying customer (I thought).

I buy a Next.js template with Stripe integration, and build the tool in 3 weeks. After reaching back to the customer, he no longer needs it.

Today the Stripe account still shows $0 in revenue.

Solve it manually first.

My biggest mistake was coding before doing. By manually executing a task you can discover bottlenecks and hidden edge cases.

Example 1: GPT-4 128k context window seemed ideal for bulk processing until I realized LLMs prioritize the start and end of inputs. When sending 50 descriptions, only the first three got properly rewritten while the rest were just copy-pasted.

Example 2: Splitting the workflow worked better. Extracting features with Mistral-7B and then rewriting with GPT-4 increased output quality compared to brute-forcing a single model.

This necessary trial-and-error process meant I kept rewriting code only to throw it away. Jumping straight into building a solution ended up costing more time than it saved.

Fall in love with your ICP or walk away.

I still believe there is a real problem here and that could probably be monetizable. The thing is that scaling demands strong ties to SEO agencies or energy to build them. I had neither.

Passion is important and I realized I didn’t enjoy working with SEO agencies. Looking back, I should have been honest with myself and admitted that I wasn’t motivated enough by this type of customer.

If the idea of another call with your target customers doesn’t make you excited, you’re building something that won’t last.

Use templates, no-code, and open-source for prototyping.

90% of MVPs don’t make it. Instead of building everything yourself, leverage templates and open-source tools.

The open-source ecosystem is growing fast, and chances are, someone has already built and shared something similar to what you need.

In my case, using a Next.js template saved me about four weeks of development only to hit the same dead end, but much faster.

Ignoring Code Perfection Doubled Our Traffic (2025, 28 years old)

I start using Twitter and discover the “Build in Public” trend. Solopreneurs are creating and growing businesses while sharing the journey on social media. One of them, John Rush, is creating niche directories.

I get fascinated by the traffic he is pulling into his websites without spending a dollar, just SEO. An ex-colleague, Nacho, is looking for a project and is very interested in AI Agents. We decide to launch a directory of AI agents

Less coding, more traction.

Being conscious of my biggest weakness, this time I decide to invest more time into content creation and directory promotion. I reduce coding to the bare minimum.

Every day I have to fight against myself not to code “indispensable features”. Surprisingly, the directory keeps gaining consistent traffic despite being far from perfect. Users don’t care about these bugs and typos that I once considered “unacceptable”.

Coding is seductive. There are always things that need to be improved. Fixing minor bugs is like dopamine shoots. They give short-term satisfaction at the expense of sacrificing other actions that have the potential to be more impactful in the long-term.

 Measure the impact of your actions and double down on what works.

Out of the many actions we do to promote the directory just a couple bring most of the traffic to the site. We set up an analytics system with PostHog and found wild imbalances:

• 1 Reddit post analyzing AI agent frameworks brought 10x more visitors than 20 promotional posts.

• 1 manually written blog post outperformed all AI-generated content combined.

• 1 listing accounts for 70% of the organic traffic.

Without tracking, we would continue investing time and resources into actions that yield little to no impact.

You have to start somewhere.

A common mistake among aspiring entrepreneurs is setting unrealistic short-term expectations and lacking persistence.

Everyone wants to create the next Facebook or Uber but often forgets to start with something manageable that can get off the ground.

By picking a more modest project, you gain valuable knowledge, build momentum, and develop the skill set needed to tackle bigger opportunities.

For us, the AI agents directory is much more than just a standalone site, it's a strategic project:

1. Discovering New Products: By managing the directory, we've encountered pain points other directory builders likely face, especially the two hours a day we spend on content curation and promotion.

   This firsthand experience not only drives us to solve our own problems but also reveals product opportunities that address real needs, like an AI solution for directory operations.

2. Gaining Domain Authority: By consistently publishing content and curating the best AI agents, we’re establishing ourselves as experts. This positioning will help us when launching future projects, such as the upcoming Agentic Project Podcast (spoiler alert).

3. Boosting Other Projects: The directory drives brand awareness and traffic that can be channeled into our personal brand and future products. It’s an engine we can leverage for growth across other initiatives.

Quitting My Job and Looking Ahead (2025, 28 years old)

The opportunity cost of staying in my full-time job outweighed the benefits. I decided to quit and focus on my AI projects full-time.

After more than a year of working with AI, I see huge potential in vertical AI agents, autonomous actors that can handle entire company operations like support, marketing, and sales.

This year, my focus is on helping SMEs integrate and automate workflows with AI. By 2026, I aim to have identified a clear niche and built the expertise needed to develop a network (or “mesh”) of specialized AI agents.

If your company wants to automate busywork and save time, reach out to us at Flowhog. We’d love to help.