← All field notes

How to Choose a Custom Software Development Partner

The best development partner is not the one with the longest technology list. Look for the questions they ask, the access you retain, and the evidence they use to call work finished.

Custom SoftwareVendor SelectionProject Planning

Choosing a software partner is partly a technical decision, but most failed relationships do not begin with the wrong programming language. They begin with unclear ownership, optimistic estimates, hidden subcontracting, long gaps in communication, or a launch process that nobody has rehearsed.

You do not need to become a developer to evaluate a developer. You need to pay attention to how the team handles uncertainty, explains tradeoffs, and protects your ability to operate the result after the contract ends.

Listen to the discovery questions

A serious first conversation should cover users, current work, painful exceptions, existing systems, data, access, success, budget constraints, and what can wait. Be cautious when a precise solution and deadline appear before those facts have been discussed.

Good discovery is not endless consulting. The partner should be able to say which questions must be answered now and what concrete output the investigation will produce.

Ask what you will own

Clarify ownership of custom code, design files, content, domains, repositories, hosting, analytics, store accounts, and third-party subscriptions. Understand any pre-existing libraries or licensed components that remain the developer’s or a vendor’s property.

Whenever practical, important accounts should be created under your organization with the partner granted appropriate access. That makes a future handoff an operating task rather than a negotiation.

Read the estimate for uncertainty

A useful proposal explains scope, assumptions, exclusions, responsibilities, milestones, review rounds, and how changes are handled. A range may be more honest than a fixed number when the existing system has not been inspected.

Ask what could move the price and who decides when that happens. The answer reveals whether the partner manages uncertainty or simply invoices it later.

Look at how progress becomes visible

You should see working pieces during the project, with enough context to give useful feedback. Ask how decisions, risks, test results, and approvals are recorded. Status meetings alone do not prove that the product is moving.

Notice how the team communicates bad news. Early, specific reporting is far more valuable than confident updates followed by a surprise delay.

Define done before the final week

Agree on supported devices, important workflows, accessibility, security, performance, data migration, analytics, documentation, training, backup, rollback, and post-launch support. Make sure production verification is included rather than assumed.

References and portfolios can show taste and experience. The stronger signal is an operating approach you can understand: clear boundaries, reviewable work, controlled releases, and a clean way for either side to hand the system over.