How to Choose a Software Development Agency (and the Red Flags to Avoid)

Choosing the wrong development agency is one of the most expensive mistakes a founder can make — not just the money, but the months lost and the codebase you inherit that nobody can maintain. The hard part is that the signals founders instinctively judge on — a slick sales deck, a low quote, a fast "yes" — are exactly the ones that don't predict a good outcome. Here's what actually does, and the red flags worth walking away from.

How do you choose a software development agency?

Choose a software development agency on process, code ownership, communication, and relevant shipped work — not on price or sales polish. The best agencies let you talk to the engineers, give you your code in your own repository from day one, demo working software weekly, and can show you products like yours they've actually launched. If any of those four is missing, keep looking.

Price is the worst primary filter. The cheapest quote almost always becomes the most expensive project once you count rework, missed scope, and the rebuild you'll eventually pay someone else to do.

The criteria that actually matter

| Criterion | What good looks like | |---|---| | Process | Fixed sprints, weekly demos, a board you can see, clear scope per sprint | | Code ownership | Your GitHub, your accounts, your IP — from commit one | | Communication | Direct access to the people building, not only an account manager | | Relevant work | Shipped products in your space they can point you to (and reference) | | Team continuity | The same people start to finish, not a bait-and-switch after signing | | Handover | Documentation and a plan for life after the engagement ends |

Notice what's not on that list: company size, office location, or how impressive the pitch was. A two-person studio that ships and hands over clean code will serve you better than a 200-person shop that staffs you with juniors after the senior closed the deal.

Red flags to walk away from

Some of these are dealbreakers on their own:

• They won't give you the code as they build it. Code held hostage until final payment is a control tactic, not a process. Your IP lives in your account from day one or you walk.
• You can't talk to the engineers. If every technical question routes through a salesperson, you have no way to judge competence — or to course- correct during the build.
• The quote is dramatically lower than everyone else's. Either they've misunderstood the scope (you'll pay for it later) or they're planning to make it back on change requests.
• No real references. "We've built hundreds of apps" with nothing you can click on, and no client who'll take your call, is a portfolio of screenshots.
• They say yes to everything. A good partner pushes back, cuts scope, and tells you when an idea is a bad one. A vendor that never disagrees is optimising for the contract, not the outcome.
• Vague on what happens after launch. If there's no handover plan, you're buying a black box you can't maintain.

Questions to ask before you sign

Bring these to the second conversation, after the pitch:

1. "Whose GitHub does the code live in, and from when?" (Answer should be: yours, from day one.)
2. "Can I speak to the developer who'd actually work on this?"
3. "Show me two products you shipped in the last year that are live today."
4. "What happens if we disagree on scope mid-sprint?"
5. "What does handover look like when we're done?"
6. "What would you cut from my scope to ship faster?" (A good answer here is gold — it shows they think about your outcome, not just billable hours.)

The quality of the answers matters less than whether they answer specifically. Vague, rehearsed responses to concrete questions are themselves a signal.

Fixed-price or time-and-materials?

For a first engagement with a new agency, a fixed-price, fixed-scope sprint de-risks the relationship — you know the number and the deliverable before you commit. Time-and-materials makes sense later, once trust is established and the work is genuinely open-ended. We break down the trade-off in detail in fixed-price vs time-and-materials.

It's also worth being honest with yourself about whether an agency is even the right call versus hiring in-house — we cover that decision in agency vs in-house.

The bottom line

Judge a development agency on process, code ownership, communication, and relevant shipped work — never on price or polish alone. Walk away from anyone who hides the code, hides the engineers, lowballs the quote, or can't show you something live. Ask specific questions and listen for specific answers.

The right agency feels like a partner who's invested in your outcome and isn't afraid to tell you hard truths. The wrong one feels like a vendor optimising for the next invoice. The difference is usually obvious by the second conversation — if you know what to look for.

Want to see what that looks like in practice? Book a free scoping call — you'll talk to the people who'd build it, your code lives in your repo from day one, and we'll tell you honestly what to cut. Not sure on budget yet? Start with how much an MVP costs in the UK.