Service Agency vs Software Product Business

The grass is not greener on the other side of the hill. It is just expensive in a different way.

I have lived both sides. Agency and product. To actually build the product, I had to deliberately let 40 customers go. This is not theory from a book. It is a decision that hurt.

What you take away here:

• Why the agency margin keeps you trapped and the product pre-finances you.
• The three traps when switching to product.
• Why I build systems that stay instead of dependency.

My thesis: the agency sells hours and creates dependency. A product is a system that stays. I leave, the system stays.

🧨 Year by year instead of project by project

The agency business feels safe. Cash flow from the first project. But you are in the office at 11pm, working off support cases on Saturday, selling hours. If you succeed, you have to hire people. And you stay at a 0 to 20 percent margin. Tedious.

The product flips the math. Instead of project by project, you collect year by year. Recurring revenue that compounds. But it arrives late.

🛠️ The three traps when switching

1. Product instead of project management. In a project you build what the one customer wants. In a product, every special request is poison. A single checkbox that only one customer needs costs a fortune to maintain over the long product cycle. I often said yes when someone came with the carrot and 50,000 francs. It never paid off. The maintenance was too high every time.

2. Pre-financing. This is where most people miscalculate. A project brings maybe 100,000 francs plus 20,000 support per year. In a product the same customer pays around 40,000 a year, but you pre-finance the development. With the first customer there is a gap. With the second it is twice as big, with the third three times. The more customers, the bigger the gap. On top of that, CAC payback of 9 to 18 months. And I have never seen a true product-market fit in 6 months. It takes years.

3. Organizational structure. A good product team is 3 to 4 people working efficiently. That costs 300,000 francs a year and more. If everything hangs on one person, you have a lump risk. If they drop out, the knowledge is gone and the company is dead.

🤖 The master plan

Do not build from zero. SaaSify an existing project. Take the solution you already built for one customer and turn it into a reusable product. DigiTickets came out of a ticketing solution for an aquarium that way. Bexio, Paymash and Localina came out of agencies. The pattern repeats.

The worst version is being neither fish nor fowl. A year of product, then back to project work because it is too slow. That just burns money.

Today there is a lever that did not exist before. I call it Get Multiplayer: how people and AI agents work together. A hyperlean team, a few pros plus agents running around the clock. Three instead of thirty. That shrinks the 300,000-franc team and the deep valley of tears. The agents run on the Context Engine, your company's business and code context. It is exactly that context which walks out the door when an agency leaves.

🎢 Highs, lows, warning

✅ What works: a product scales, an agency does not. More customers do not need more hours. The people who could sell a product made more money in the end.

❌ What does not work: the Costa Rica fantasy. Open the notebook on the beach, answer two requests, go surfing again. No product runs like that. It is always a launch, always high performance required.

⚠️ Warning: a product company is like a zoo. You have to align your customers. Build a zoo for elephants and put mice in it, and you have the wrong zoo. And: as a product manager you have no friends. You say no to money. That is hard.

Why I still choose the product: the agency creates dependency, the knowledge stays outside. I build a system that stays installed in the company. The risk is bigger, the reward too. I leave, the system stays. If you want such a system built for you, take a look at Pedalix.