What Freelancing Taught Me About Building Software for Clients

One of the biggest reasons I’m able to bring value to my clients’ software products is that I started my career as a freelancer.
I worked with more than 15 teams across very different products and industries. I didn’t spend those years solving the same kind of problem over and over. And I feel the value of that experience most in the first discovery meeting, before we’ve written any code.
The same feature can hide two different problems
Every application is different. Some are heavy on backend and algorithms. Others live or die by how easy they are to use. On others, security shapes the architecture from day one. Even two products in the same industry can need very different things.
Take two online stores. When I’m buying a shirt, I need to find my size. When I’m buying brake pads, I need to know they fit my exact car. Both stores might ask for better product filters. But for the parts store, I’d first ask where the fit information comes from and what we tell the customer when we can’t confirm a match. A great checkout won’t help someone who still isn’t sure they’re buying the right part.
That question changes what we build first. We can start with one category and prove we can reliably match a part to a vehicle before building out a whole catalog around it. That’s a decision worth making before we design all the screens.
What will users need in three months?
While a client is explaining what they want the app to do today, I’m already thinking about what their users might need in three months. That doesn’t mean I want to build three months of features on day one. I want to know which decisions are easy to change later and which ones could force us to rebuild.
Say we’re launching a dashboard where the founder is the only person using it. We can probably leave a full permissions screen for later. But when the next step is inviting customers to manage their own accounts, we should decide who owns the data and who can see it before we build the first version. The screen can wait. Getting the ownership wrong can mean revisiting every place the app reads or writes customer data.
Experience across different products doesn’t mean I walk into a discovery meeting with all the answers. It means I have better questions to ask while the decisions are still cheap to change. By the end of that first meeting, I want us to know what to build first, what we need to prove, and what we can leave alone for now.

