How to Choose a Software Agency Without Getting Burned

Most software projects fail before the first line of code gets written. Not because the idea is bad. Not because the market doesn’t exist. But because the founder picked the wrong agency-and by the time they realised it, they’d already burned through $50,000 to $200,000 AUD and six months of runway.

This post is about not being that founder.

The First Red Flag: They Promise Speed Without Understanding Your Problem

Any agency that quotes you a timeline before asking real questions is already failing you. A good technical partner will want to know:

• What specific problem are you solving? (Not the feature list-the actual user pain point.)
• Who is paying for this, and how?
• What’s your runway? What’s the real deadline?
• What have you tried already, if anything?
• Who else is doing something similar, and why aren’t they winning?

If they’re talking about shipping an MVP in two weeks, they’re either lying or they’re not building anything real. That said, some agencies can move fast-but only because they’ve done the discovery work upfront and their process is tight. We ship MVPs live in 28 days, but that comes after a structured one-week discovery phase where we stress-test assumptions and validate the architecture with you, not in place of it.

The timeline should come after the diagnosis, not before.

Check Who’s Actually Doing the Work

This is the biggest hidden cost in software development: hidden offshore teams.

A lot of Australian agencies quote you a price, take your brief, and then spin up a team in India, the Philippines, or Eastern Europe to do the actual work. There’s nothing inherently wrong with that model-it works for some projects-but you need to know about it upfront. More importantly, you need to understand the time-zone delays, communication tax, and quality variance that come with it.

Ask directly: “Who writes the code? Where are they? Will I be talking to them, or to a project manager?” If they hedge or give you vague answers, walk.

The best signal is this: can you have a technical conversation with the person who’ll actually build it? Not a sales call. A real technical call where they push back on your idea, ask hard questions, and help you think through architecture decisions. If the agency won’t let you do that until you’ve signed and paid, that’s a sign the real builders are someone you won’t meet until you’re locked in.

Understand the Real Cost Structure

Software agency pricing comes in roughly three flavours:

1. Fixed price per project. You pay AUD $40,000 to $150,000 upfront for a defined scope. This sounds good until scope creeps, requirements shift, or the agency discovers a technical blocker mid-project. Read the contract carefully: what happens to cost if requirements change by 20%?
2. Time and materials (hourly or daily rate). You pay AUD $120 to $250 per hour, typically billed monthly. This is more flexible but riskier if the agency has no incentive to finish efficiently. Make sure there’s a cap or a clear milestone-based payment schedule.
3. Retainer + project work. You pay a monthly retainer (AUD $3,000 to $10,000) for ongoing support, then pay separately for new builds. Good for long-term partners; bad if you’re a one-off project.

What matters more than the pricing model is whether they’ll give you visibility into how the money translates to effort. If they can’t tell you “this feature will take roughly 40 hours of backend work plus 20 hours of frontend” and break it down, they haven’t thought through the work-which means the estimate is a guess.

Ask for a detailed breakdown. If they won’t provide one until you sign, that’s a negotiation red flag.

Look at What They’ve Actually Built

Check their portfolio, but don’t just look at the design. Ask harder questions:

• Can they show you the code? (Open-source contributions, GitHub repos, architecture diagrams.)
• Do they have case studies that talk about the technical decisions, not just the pretty screenshots?
• Have they built something in your vertical or similar problem space?
• Can you talk to a founder they’ve worked with?

A solid agency will have concrete examples of systems they’ve built: databases that scale, payment integrations that don’t drop transactions, AI systems that actually work in production (not just prototypes). If their portfolio is all brochure websites and they’re pitching you a machine learning product, that’s a mismatch.

Also ask: how many projects have they shipped? How many did they ship on time? How many clients came back for phase two? The answers tell you a lot about execution consistency, not just design taste.

Test Their Technical Depth

You don’t need to be a technologist to ask this, but you do need to ask it: what’s their actual technical stack, and can they explain why?

A real agency will have opinions. They’ll tell you: “We build React frontends and Node backends for consumer apps because the ecosystem is mature and we can hire for it easily. For data-heavy systems, we’d recommend Python. For something that needs to scale to millions of requests, we’d architect around serverless or Kubernetes depending on your ops capacity.”

They won’t say: “We can build anything in any tech stack.” That’s a warning. It usually means they don’t have strong opinions, which means they’ve been burned by technology choices before, which means the next client (you) is going to be their testing ground.

Ask them: “What’s a technology you’d refuse to use and why?” If they can’t answer that, they don’t have a technical philosophy-they have a services menu.

Watch for Scope Creep Before It Happens

The single most common reason software projects go over budget is that the scope wasn’t nailed down at the start. So ask your potential agency: how do you manage scope?

A good answer sounds like: “We define the MVP scope in week one. Anything outside that goes into phase two or the backlog. We track scope changes and show you the cost impact before we build. We do this because uncontrolled scope kills projects.”

A bad answer sounds like: “We’re flexible and can handle whatever you need.” That flexibility will become your problem once you’ve spent two months and AUD $80,000 and the project is only half done.

Get the scope agreement in writing. Not vague (“build a marketplace”). Specific (“three user roles, five core features, payment integration with Stripe, email notifications, admin dashboard”). Then ask: what happens if you want to add a sixth feature mid-build?

Does the Agency Care About Your Success?

The best signal is whether they’re willing to have a hard conversation with you before taking your money.

If your idea is half-baked, will they tell you? If your timeline is unrealistic, will they say so? If they think you should do something differently, will they push back or just nod and take the cheque?

A good agency will occasionally say “no, we don’t think we’re the right fit” or “we can build this, but we’d recommend starting with phase one first.” That’s expensive honesty, and it’s a sign they’re thinking about your outcome, not their utilisation rate.

If you’re ready to have that conversation with a team that thinks in systems and cares about outcomes, talk to Amora about your build.

The Actual Checklist

Before you sign with any agency:

• Have a real technical conversation with the person who’ll build it.
• Ask who’s doing the work and where they’re located.
• Get a detailed breakdown of hours/cost per feature or component.
• Review their portfolio and talk to a past client.
• Ask them to explain their technology choices with reasoning, not just tools.
• Define scope in writing and ask how scope changes are handled.
• Watch whether they push back on unrealistic timelines or scope creep.

The cheapest agency isn’t usually the worst. The worst agency is the one that nods along, takes your money, and ships you something that doesn’t work or takes twice as long as promised. Getting burned isn’t about cost-it’s about communication, clarity, and alignment on what winning looks like.

Choose a partner who’s willing to think hard before they start building.