What to Actually Ask Before Hiring an Outside Engineering Partner
Most vendor evaluations focus on portfolios and hourly rates. The questions that actually predict delivery quality are different — and most companies never ask them.
By the time a company calls us to fix a project another firm delivered, the pattern is almost always the same: the vendor selection process focused on the wrong signals. A polished portfolio, a compelling sales deck, a rate that fit the budget. None of those predict whether the team behind them can actually deliver a maintainable system on time.
We've been the second call — the team brought in to rescue or rebuild — often enough to know exactly where the original vetting process broke down. It's rarely a skills gap. It's a due diligence gap. The questions that would have surfaced the risk were never asked.
The Questions That Actually Predict Delivery
- 1Who owns the code, in writing, from day one? Not "at project completion" — from the first commit. Ambiguous IP and repository ownership is the single most common source of painful, expensive disputes at handover.
- 2What does your architecture decision record look like? A serious partner documents why they chose a database, a framework, a service boundary — not just what they built. If they can't produce this for a past project, they aren't going to produce it for yours.
- 3What happens if our point of contact leaves your company mid-project? If the answer is vague, you're relying on one person's memory instead of an institutional process.
- 4Can we talk to a client whose project ended, not one who is still actively engaged? A reference from an active client tells you about the sales relationship. A reference from a completed project tells you about the handover.
- 5What's your actual test coverage philosophy, and can you show it, not describe it? Ask for a sample of tests from a comparable past project, not a slide about their QA process.
The Red Flags We See Most Often
- A quote that's dramatically lower than every competing bid, with no explanation of where the margin is coming from — usually it's coming from documentation, testing, or senior oversight, all of which surface later as your problem.
- Reluctance to commit to a fixed architecture review milestone before full-speed development begins.
- A team structure where you never learn the names or seniority of the actual engineers writing your code, only the account manager.
- No willingness to walk you through how they'd approach the handover before the contract is signed — handover thinking that starts on day one produces a fundamentally different codebase than handover thinking that starts in the final sprint.
A partner who can't clearly explain how they'd hand the system back to you is not planning to build something you can operate without them. That dependency might be convenient for their business model. It's a liability for yours.
Where to Start
Before your next RFP, write down what you'd need to see, in writing, to be confident a stranger on your own team could pick up this codebase in two years and understand it without a single call to the original vendor. Then use that list — not the pitch deck — to score every proposal you receive.
As a custom software development company that's been on both sides of this — the original build and the rescue call — our view is simple: the vetting conversation you have before signing is worth more than any clause you add after.