How to Open a Software House from Scratch

Opening a software house is deceptively simple. You bring two good developers together, close the first client through referral and that's it, you have a company. The difficult part is not starting. The difficult thing is not to go broke in the second year, when cash gets tight, the partner gets tired and contracts that seemed good turn out to be traps.

Most software houses that close do not close due to lack of technical competence. It closes due to business decisions that no one taught: wrong model, poorly designed company, pricing that doesn't cover costs, contract that transfers all the risk to you.

This is an honest map of what you need to decide before and during, written by founder for founder.

Decide on the model before deciding on the name

The first strategic decision is not the brand, it is the business model. There are three main paths, and they lead to completely different companies.

The body shop model, or allocation of professionals, is the easiest to start with. You allocate developers to the client and charge each person hourly or monthly. It generates quick and predictable cash flow, but has a low margin ceiling and turns you into a people reseller. It grows in volume, not in value.

The project model, or delivery under scope, is what most imagine when they think of software house. You assume responsibility for the delivery of a product. Higher margin, higher risk, and requires real project management. This is where differentiation lives, but also where most people get hurt by poor pricing.

The product model is when you use the service operation to finance the building of something of your own. It's almost every owner's dream, but mixing service and product without discipline usually kills them both. Decide consciously, because each model requires a different type of sale, team and cash flow.

Partnership is marriage, and most start without a contract

Most software houses are born from a partnership between friends, generally one more technical and one more commercial. This can be great or be the cause of the company's death, and the difference is in the conversations you have before you have money to fight.

Define, in writing, at least three things. Who decides what when there is disagreement, because consensus for everything stops the company. How what each person produces is divided, because equal division of society with unequal effort poisons the relationship. And what happens if one wants to leave, because leaving without clear rules freezes the business.

The partners' agreement is not distrust, it is hygiene. Members who refuse to put the rules on paper are warning you something. The societies that last are the ones that agreed on the worst-case scenario while still liking each other.

The first contracts define which company you will be

There is a huge temptation in the first few months: accepting any project because you need cash. I understand the temptation and it has a hidden cost. Your first clients define your portfolio, and your portfolio defines who comes next.

If your first three projects are cheap, poorly defined, and from different industries, you've just built a cheap generalist company without realizing it. The next customers will arrive similar to the ones you've already served, because that's what your social proof shows.

I'm not saying to turn down money when you're starting out. I'm saying to choose, as much as possible, projects that you want to show. A good case in the right sector is worth more than three forgettable projects, because it is what opens the next door.

Contract and legal are not bureaucracy, they are risk management

The area that most fails young software houses after cash is poorly done legal work. Not by process: by open scope formalized in a loose contract. The client asks for change after change, you deliver to maintain the relationship, and the project that was profitable turns into a loss.

Your contract needs, at a minimum, to clearly define scope, describe how scope changes are handled and billed, establish payment milestones that don't let you fund the customer, and define who owns the code and intellectual property. These four points solve most real pains.

A well-written scope change clause is not hostile to the customer, it is what protects the relationship. Without it, every change becomes a tense negotiation. With it, change is a predictable process with a defined price. It's worth paying a lawyer once to put together a good model and reuse it.

Caixa kills more companies than lack of customers

Profitable software house on paper breaks per box all the time. The reason is the mismatch: you pay salary every month, but receive it from the client at spaced milestones, sometimes late. Profit is opinion, cash is fact.

The rule of survival is simple to say and difficult to follow. Receive enough upfront to not finance the customer, prefer payment milestones linked to deliveries and not distant dates, and maintain a reserve that covers the payroll for a few months without new invoicing. Without this reserve, any customer who delays a payment puts you in a panic.

The most common mistake is to grow a team in anticipation of contracts that have not yet been signed. You hire for the demand you imagine, the demand takes time, and the sheet arrives on the right day. Grow a team behind a signed contract, not behind a promise. The predictability of how much it costs to hire a software house on the client's side arises precisely from this cash discipline on the side of those who deliver.

The mistakes that most bring down those who start

After seeing many operations born and some die, patterns of failure repeat themselves. It's worth knowing them so you don't repeat them.

The founder becomes the bottleneck for everything: selling, delivering, billing, hiring, and the company doesn't move forward when he stops. This is not dedication, it is a lack of process, and it has a physical limit. The second mistake is confusing revenue with health: a lot of revenue with tight margins and negative cash is a sick company disguised as a successful company.

The third is not having a positioning, accepting everything and competing only on price, which erodes margin until there is nothing left. And the fourth is to treat pricing as a guess, without understanding the real cost per hour, which includes non-billable time, taxes and risk.

Opening is easy. Building a business that outlives you is the real work, and it starts with those unglamorous model, partnership, contract and cash decisions.

If you are setting up your operation now, it is also worth reading about how to attract the right customers and how to price without making mistakes. The other articles in this series cover each of these fronts in detail.