Rescuing a stalled software project: what to do when your developer goes quiet

Qasim Ali · Senior Software Engineer (Java, AI), CS2 Technologies
Published · 6 min read

It usually starts small. A deadline slips, then another. Replies take three days instead of one. Eventually the agency or freelancer you hired stops answering, and you’re left holding a half-built product you’ve already paid a good part of.

The first instinct is to find someone new and start over, or to send a sharp email and hope the old team comes back. Both can wait a few days. What you do first decides how much of the money you’ve spent you get to keep.

Here’s the order we’d work in if you brought a stalled project to us: lock down access, get the code, find out what it really is, and only then decide whether to finish it or replace it. You don’t need a technical background to start.

1. Secure your accounts before you do anything else

Make a list of every account the project touches: the domain registrar, the hosting or cloud account, the code repository, the payment processor, email and DNS, analytics, app store listings, and any outside service that uses an API key. Next to each one, write down whose name it’s in and who can log in.

If any of those sit under the agency’s account, ask in writing for ownership to be moved to you, and keep the tone calm. A polite, specific request (“please add me as owner of the AWS account and the GitHub organization”) gets a faster yes than a threat. Once the domain and the hosting are in your hands, the worst case is a rebuild. Without them, things can get a lot harder.

For decision-makers: Today, not next week: list every account the project uses and whose name it’s in.

2. Get the code, with its history

A zip file of the latest version is better than nothing, but it isn’t the whole project. Ask for the full repository with its commit history, the database structure and a recent backup, the list of settings and secrets the app needs to run (shared through a password manager, never by email), and whatever notes exist on how to build and deploy it.

The history matters more than people expect. It shows what changed, when, and often why. A new team reading it can tell in an afternoon whether the work was steady or rushed at the end. Without it, every odd line of code is a small mystery that costs time to solve.

For decision-makers: Ask for the full repository, a database backup and the deploy notes. A zip of the last version leaves out most of what a new team needs.

3. Find out what you actually have

Before anyone promises to finish the job, someone independent should look at it. The questions are plain ones. Does the code build on a clean machine? Does it run? Which features work end to end, which are half done, and which exist only in a design file? Is anything live, and if so, is it the same code you were handed?

Compare the answers with your original scope or contract. That gives you an honest picture of how close to done you are, which is often different from the one you were told. It also gives the next team a fair starting point, so their quote is based on what exists rather than on guesses.

For decision-makers: Pay for a short written assessment at a fixed fee before you commit to a full rescue. It costs little next to a wrong call.

4. Rescue or rewrite: let the assessment decide

Finishing usually makes sense when the code builds, it’s written in a mainstream framework other developers know, the data model fits how your business works, and the main user flows mostly run. In that case, completing the work is usually faster than starting again, and you keep the parts that were done well.

A rewrite earns its place when nobody can get the project to build, when it sits on versions that no longer get security updates, or when what was built no longer matches what you need. There’s a middle path that often works, too: keep the design, the content and the data, and rebuild the layers underneath that don’t hold up.

Be a little careful when a new team recommends a full rewrite on day one. Developers tend to prefer their own code. Ask them to put their reasons in writing, tied to what the assessment found.

For decision-makers: If a team says “rewrite,” ask which specific findings make finishing the existing code the worse deal.

5. Choose the next team for how they hand things over

Skills matter, but you’ve just learned they weren’t the only risk. This time, ask how the work is set up. Will the repository, domain and cloud account be in your name from the first day? Will you see working software at the end of each phase, or only at the end of the project? What happens to the code if you stop halfway?

Good answers are specific. At CS2, work runs in dated phases, and each phase is signed off in writing before it begins. Repositories, domains and data are set up in the client’s name from day one. We also run our own software products in production, so you can look at live systems the same team operates instead of taking our word for it.

For decision-makers: Ask every candidate team where your code lives while they work. “In your repository, from day one” is the answer you want.

6. Write the lessons into the new contract

A few clauses would have made the last situation much easier, so put them in this time. Code goes into your repository as the work happens, not at the end. Payments are tied to milestones you can click through and accept. Documentation, including how to deploy, is a deliverable with its own sign-off. And a short exit clause spells out what gets handed over, and how fast, if either side ends the work.

None of this signals distrust. A team that already works this way will sign it without a second look. If someone pushes back hard on you owning your own code, you’ve learned something useful before paying a cent.

For decision-makers: Tie every payment to something you can see working, and put the handover terms in the contract from the start.

The short version

A stalled project feels like a loss, but you can often save more of it than it seems on the bad day, as long as you move in the right order. Secure the accounts, get the full code, get an honest written assessment, and only then choose between finishing and rebuilding. When you hire the next team, weigh ownership and visibility as heavily as skill.

Qasim Ali

Senior Software Engineer (Java, AI), CS2 Technologies

Qasim has 18 years of backend experience, including ten years on the backend platform of Vodafone Kabel Deutschland, and works on CS2’s own products as well as client builds. If your project has stalled, the CS2 team can look at what you have and tell you plainly whether it’s worth finishing.