How to Choose a Software Development Agency (Without Getting Burned)
Choosing a development agency is mostly about detecting how they behave when things go wrong — which is exactly what a sales conversation is designed to hide. Here is what to ask instead.
Ask who actually writes the code
Agencies frequently sell with senior people and deliver with junior ones, or subcontract without saying so.
Ask: who specifically will work on this, what else are they on, and will I talk to them directly? A studio that routes everything through an account manager adds a layer of translation between you and the people making decisions, and translation loses detail.
Ask what happens when it runs over
Every project of any size hits something unexpected. What matters is what happens next.
Ask: tell me about a project that went over. What caused it, who absorbed the cost, and what changed afterwards? A studio that claims this has never happened is either inexperienced or not being straight with you.
Ask about ownership before you sign
The two questions people ask too late:
- Do I own the source code outright when I have paid in full?
- What does it cost me to leave, and what do I walk away with?
Some agencies keep the code and licence it back. Some build on proprietary platforms you cannot export from. Neither is automatically wrong, but both must be a decision you made knowingly rather than discovered in year two.
Get the answer in writing. "Of course you own it" in a call is not a contract clause.
Ask how you will see progress
Weekly, working software beats a long silence and a big reveal. If the first time you see it is at the end, problems have had months to compound.
Ask: what will I see in week two, and how often after that? Vagueness here predicts vagueness throughout.
Look at what they built, not what they say
Portfolios are curated. Go further:
- Open their work on your phone. Does it load fast? Does it feel considered?
- Ask which parts they actually did. Design, build, both? On a team of how many?
- Ask for a reference you choose from their list, not the one they offer.
Watch how they handle disagreement
The strongest signal in the whole process: push back on something they proposed and see what happens.
A studio worth hiring will either explain their reasoning specifically, or change their mind for a good reason. One that instantly agrees with everything is not going to tell you when you are about to make an expensive mistake — and you are paying partly for exactly that.
The warning signs
- A quote with no breakdown. One number, no composition.
- No questions asked. Anyone who can price your project without understanding it is pricing a template.
- Pressure to decide now. Discounts that expire are a sales tactic, not a scoping method.
- No written scope. If it is not written down, it is not agreed.
- Everything is easy. Experienced builders volunteer risks early, because risks discovered late are what destroy timelines.
What a good process looks like
- A conversation about the business problem, not the feature list
- A written scope, with what is explicitly not included
- A quote you can read, broken into stages
- Weekly visible progress
- A launch plan that includes what happens afterwards
- A handover with code, access and documentation you own
One last test
Ask them to talk you out of something. "Is there anything in this brief you would not build?"
A studio that says yes, and explains why, is a studio thinking about your outcome rather than your budget. That is the whole thing you are trying to detect, and it is usually answerable in one question.
If you want to see how we answer these, our FAQ covers pricing, ownership, timelines and what happens after launch — in writing, before you ask.