Questions to ask before hiring a software studio

Every non-technical founder walks into the first meeting with a studio carrying questions they're embarrassed to ask. What does this really cost? Who will code it? What if it runs late? Many people swallow the question so they don't look naive. And that swallowed question is exactly what becomes a problem three months later.

On September 23, 2026, the Tech for Non-Techies podcast, hosted by Sophia Matveeva, devoted an entire episode to these questions. The guest was Vuk Nikolic, who leads Hurricane Studio. An important disclosure: Hurricane sponsors the episode, the first sponsor in the podcast's history, as the host herself notes. Keep that in mind as you listen. Even so, the advice works for evaluating any vendor, including us.

What the episode recommends

The points that stood out most to me:

• Set aside at least 30% of the budget. For misunderstandings, requirements nobody said out loud and scope that grows.
• Scope matters more than hours. What drives cost is what you ask for, not the hourly rate.
• Start with the smallest product you can sell. And use discovery to size the team.
• Define what “done” means. Coded, tested, published, with marketing in the loop, or live?
• Be wary of anyone who says yes to everything. For Nikolic, accepting everything without questioning scope, timeline and budget is “the most expensive mistake you can make.”

Nikolic also shared cost ranges by product type. I won't repeat the numbers here, for a simple reason: they depend on scope, country and studio, and a loose number becomes the wrong anchor in the reader's head.

The questions you're embarrassed to ask

I combined what the episode suggests with what I've seen in more than ten years of meetings with founders. Bring this list printed out, if you need to.

On scope

1. What's left out of the first version? If the studio can't list it, it didn't cut anything. And scope with no cuts doesn't fit any date.
2. What happens when I ask for something new mid-project? The right answer has a process: assess it, show the impact on date and price, and you decide.

On price

3. How did you get to this number? A price given before understanding the problem is a guess. See how much a custom web system costs and what moves that number. In the episode, quoting an hourly rate before understanding the problem comes up as a red flag.
4. Is the price fixed or can it change? If it can change, ask under what conditions, in writing.
5. How much should I set aside for surprises? An honest studio won't say “nothing.” The episode's benchmark is 30%.

On who builds it

6. Who will build my product, and can I talk to that person? Having access only to a senior manager, with no contact with whoever codes, is another red flag Nikolic mentions.
7. How many projects do you run at the same time? A team split across many clients is a team missing from yours.
8. How do you use AI? A fair question in 2026. A good answer explains what AI does and who reviews it.

On timeline

9. Does the date go in the contract? A spoken date doesn't count.
10. What does “delivered” mean? Agree on the definition first. “Ready in the test environment” and “live with customers using it” are different dates.
11. When do I first see the product running? The sooner, the lower the risk.

On what comes after

12. Who looks after the product after launch? Live software needs someone watching it.
13. Do I have access to the project's code and tools? The episode recommends access to the repository and task board, without micromanaging.

The signal worth more than any answer

The episode's final advice is the one I endorse most: a good partner disagrees with you when they need to. In Nikolic's words, “if everything is yes, they're only there for the money.”

When a studio says “this doesn't fit the date, let's cut it,” it's protecting your money. When it says yes to everything, it's pushing the bill down the road.

How we do it at the Studio

We answer this list in the first conversation. The price is fixed after discovery, never before. The deadline is written into the contract. The code, files and access are in your name. We use Claude Code and Figma for design and build, and we explain what AI does and what we review.

And every system ships with an AI agent that runs daily, flags issues, suggests improvements and checks that the data makes sense. That's our answer to question 12.

If you're about to hire someone to build your custom system or SaaS, bring this list to the conversation with us. We make a point of answering every question.