Most software projects that go wrong don't fail because of the technology. They fail because of unclear scope, slow feedback, surprise invoices, or a developer who disappears after launch.
The best way to avoid that is to ask the right questions up front. Here are ten we'd ask if we were hiring someone ourselves.
1. Who will actually build it?
Will you talk to the people writing the code, or to an account manager who relays messages? Will the work be passed to subcontractors? Direct access to the builders makes everything faster and clearer.
2. How is the price set?
Hourly billing with no cap puts all the risk on you. Ask whether scope and price are agreed in writing before work starts, and what happens when something changes along the way.
3. When will I see working software?
You should see something you can click within weeks, not after months of design documents. Regular demos let you catch misunderstandings early, when they're cheap to fix.
4. Who owns the code, domain and accounts?
The answer should be "you, outright". Code in your repository, hosting and domain in your name, and no licence fees to keep using what you paid for.
5. What happens after launch?
Ask whether bug fixes are included, what support costs, and whether you're locked into a contract. A good partner offers ongoing help, but lets you leave whenever you want.
6. How do you handle changes?
Requirements always shift. You want a process where small changes are absorbed and bigger ones are flagged as a separate decision, not quietly added to the invoice.
7. Can I see similar work?
Ask for examples close to what you need, and ask what went wrong on one of them. Honest partners have a good answer to the second question. For example, our Kaladi Kitchens booking platform and the Minto Apartments sales-office app show two very different kinds of build.
8. How do you test and deploy?
Listen for automated tests, a staging environment, backups and monitoring. If the answer is "we test it before we hand it over", ask how.
9. What will it cost to run?
Hosting, third-party services and AI usage all have ongoing costs. You should get a realistic estimate before you commit, not after the first bill arrives.
10. When would you tell me not to build this?
This is the most revealing question. A partner who can explain when an off-the-shelf tool, an integration or a process change would serve you better is one who's looking after your interests.
Red flags
- Pressure to sign before scope is clear
- No working software for months
- Code and accounts held in the developer's name
- "We'll figure out the details as we go" with an open-ended hourly rate
- No answer to "what happens if you're unavailable?"
