Most "red flags in outsourcing" lists are written for the client. This one is written by the vendor, which means it includes the flags we watch for in clients too. A bad engagement is bad for both sides, and most of them are visible in week one.
Red flags in a development vendor
No live work. Screenshots, mockups and "under NDA" for everything. Some work is confidential; all of it is not. Ask for three things you can open in a browser.
A senior person sells, a junior person builds. Ask who will write the code and whether you will talk to them. If the answer changes after the contract is signed, you were sold a bench.
No questions. A vendor who quotes from a one-paragraph brief without asking anything is guessing. The quote will be wrong and the argument will come later.
Hourly with no cap and no scope. It is not always a scam, but it is always a risk transfer onto you. Ask for a fixed scope or a capped discovery sprint first.
The code lives in their repository. From the first commit, the repository should be in your organisation. If they resist, ask why.
Silence between milestones. Written updates should arrive on a schedule you agreed. If you are chasing for status in week two, week ten will be worse.
Every answer is "yes". Good vendors say "that will cost more", "that is a bad idea" and "not in this phase". If nothing is ever pushed back, nothing is being thought about.
Red flags in a client (what makes us walk away)
A budget that does not match the ambition, stated as a challenge. "Others quoted less" is fine; "build Uber for $2,000" is not a negotiation.
No decision-maker. If every answer has to go through a committee that meets fortnightly, the project will take a year and nobody will be happy. We ask on the first call who decides.
Refusing a trial scope. Clients who want to skip the small first piece of work and jump to a six-month contract are usually in a hurry for reasons that will become our problem.
Wanting hourly billing with no end date. It usually means the scope is unknown and nobody wants to admit it. A discovery sprint fixes that; open-ended hourly does not.
Treating a written scope as an insult. "Can't we just start?" We can, and then we will disagree about what "done" means.
What both sides should want
A short paid trial. A written scope. One decision-maker on each side. Code in the client's repository. Demos as the unit of progress. A clear way to stop. None of this is exotic; it is just rarely done in week one, which is when it matters.
We put our side of this in writing on the outsourcing page. If you are checking us out the way this post suggests, the case studies and the DevKey Technologies portfolio are where the live work is.