How to Choose a Software Development Partner: 9 Questions That Reveal Everything

You will buy custom software a handful of times in your career. The firms selling it do this every week, and that asymmetry is why so many engagements go wrong in ways that were visible before signing. The good news is that a small number of questions reliably separate firms that will do the work from firms that will sell it. Here they are, with the answers worth hearing.

Questions 1 to 3: proof and process

"Show me three projects like ours and tell me what they achieved." A portfolio of screenshots is decoration. You want outcomes, and ideally a client you can call. Firms that cannot name a single verifiable client are asking for a great deal of trust in exchange for none.

"Walk me through your process and what you need from us." The answer must contain phases, decision points and client responsibilities. If the process is effectively "send us the requirements, we will send you the software," expect surprises at the end, because there were no checkpoints in the middle to catch drift.

"What happens when we are late with something you need?" This is the question nobody asks, and the answer is revealing. Every project is delayed by client-side inputs at some point. A mature firm has a plan for it. A defensive answer here predicts a defensive relationship.

Questions 4 to 6: money and ownership

"What exactly is in the price, and what is explicitly not?" The out-of-scope list matters more than the total. Its absence is the single best predictor of surprise invoices.

"Who owns the code, the repositories and the accounts?" You should, from day one and in writing. Ask specifically where the code lives during development. If the answer is the supplier's infrastructure, handover is a negotiation rather than a formality.

"What does this cost after launch?" Hosting, maintenance, support and future changes. A firm that has not thought about your total cost of ownership is quoting to win a bid, not to keep a client.

Questions 7 to 9: the people and the truth

"Can I speak with the engineers who would actually work on this?" Fifteen unscripted minutes with the team tells you more than any proposal document. If the only available contact is sales, you are buying a layer of translation between you and the work.

"What would you advise us not to build?" This is the best question on the list. Any firm can list what it will happily charge you for. A partner who identifies a feature you do not need, or suggests an off-the-shelf tool for part of the problem, is demonstrating the judgment you are actually hiring.

"What went wrong on your last project, and what did you do?" Every firm has one. The ones worth hiring answer specifically and without blaming the client. Anyone claiming a flawless record is either new or not being straight with you.

Red flags worth ending a conversation over

A guaranteed first position on Google, or any guarantee about a platform the supplier does not control. Pressure to sign this week for a discount that expires. A price quoted with no questions asked about your business. Refusal to sign your NDA or to work with your contract. A portfolio with no verifiable names. And the quiet one: an estimate that is dramatically below every other bid, which usually means the scope was read differently and the difference will arrive as change orders.

Two practical notes on comparison. First, budget the same amount of attention for the exit terms as for the kickoff, since that is where cost hides. Second, prefer a small paid engagement to a long evaluation: how a trial project works tells you more than three months of proposals.

We are happy to be measured against this list. Ask us all nine on the first call, and expect an honest answer to the awkward ones.