You cannot evaluate a dev agency by reading their code. You are not supposed to. That is like evaluating a surgeon by watching them operate. You would not know what you were looking at.
But non-technical founders hire dev agencies all the time, and many of them make the decision based on the wrong signals: a polished website, a confident sales call, or a price that seems reasonable. Those signals tell you almost nothing about whether the agency will actually ship your product.
Here are seven criteria you can assess without writing a single line of code. For each one, I will give you the specific green flag that indicates a trustworthy agency and the red flag that should make you pause.
1. Portfolio depth: real products or just mockups?
What to look for: Ask the agency to show you live products they have built. Not screenshots. Not Figma designs. Products you can open in a browser or download from an app store and actually use.
Green flag: They can point you to 3-5 live products with real users, and they can explain what the product does, who uses it, and what their role was in building it.
Red flag: Their portfolio is entirely screenshots, mockups, or “confidential projects” they cannot show you. Some agencies do work under NDA, but if every single project is hidden, you have no way to verify their claims.
How to verify: Open the live product. Click through it. Does it work? Does it load quickly? Does it feel like a finished product or an abandoned prototype? You do not need technical skills to evaluate user experience as a user.
2. Contract structure: fixed-price or hourly?
What to look for: How the agency structures payment and what happens when the project takes longer than expected.
Green flag: Fixed-price contracts with a defined scope, explicit deliverables, and a clear statement of what happens if the project exceeds the timeline. The agency absorbs overruns on scope they agreed to.
Red flag: Hourly billing with estimated ranges like “200-400 hours.” That range means you could pay double the lower estimate and the agency has no financial incentive to be efficient. If you have been through a failed outsourced project before, this is likely the billing structure that burned you.
How to verify: Ask directly: “If the project takes 50% longer than you estimated, who pays for the extra time?” The answer to that question tells you everything about where the financial risk sits.
3. Demo milestones: when do you see working software?
What to look for: A specific date, early in the project, when you will see a working version of your product running in a browser or on a device.
Green flag: A mandatory demo by the midpoint of the project. At Buldtech, we do a Day-10 demo in a 21-day sprint. That means the founder sees and interacts with working software before more than half the budget is spent.
Red flag: “You’ll see the final product at delivery.” If the first time you see working software is the last day of the project, you have had zero opportunity to catch problems, give feedback, or redirect the build.
How to verify: Ask for it in the contract. A verbal promise of “regular check-ins” is not the same as a contractual obligation to produce a working demo by a specific date.
4. Code ownership: who owns what you paid for?
What to look for: An explicit clause in the contract stating that all code, designs, and intellectual property belong to you upon payment.
Green flag: The contract includes an IP assignment clause, and the agency sets up the project in a code repository (GitHub, GitLab) that you own from day one. You can see commits happening in real time, even if you cannot read the code.
Red flag: The agency works in their own repository and promises to “hand over the code at the end.” This gives them leverage if the relationship goes badly, and it means you cannot bring in another developer to review the work mid-project.
How to verify: Ask them to create a repository under your GitHub account before work begins. If they push back on this, ask why.
5. Post-launch support: what happens on day 31?
What to look for: A defined support period after launch, with clear terms about what is covered.
Green flag: A 30-day (minimum) post-launch bug-fix window included in the original contract price. Bugs are defined as features that were in the agreed scope but do not work correctly. New features are a separate engagement.
Red flag: “We can discuss a maintenance contract after launch.” This means there is no guarantee of support when you need it most: the first month when real users are finding real problems.
How to verify: Read the contract. If it does not mention post-launch support, it does not exist.
6. Communication cadence: how often do you hear from them?
What to look for: A structured communication rhythm with defined artifacts, not just ad-hoc Slack messages.
Green flag: Weekly (or more frequent) scheduled calls with a consistent format: what was done, what was decided, what is blocked, what is next. Written summaries after each call. A shared project board (Trello, Linear, Notion) that you can check anytime.
Red flag: “We’ll keep you posted.” No scheduled cadence. Updates arrive sporadically and only when you chase them. You find yourself sending “Hey, any updates?” messages more than once a week.
How to verify: During the sales process, pay attention to response times and follow-through. If they are slow and disorganized before you have paid them, it does not improve after.
7. Client references: will they connect you with past clients?
What to look for: Access to 2-3 past clients who will answer your questions honestly.
Green flag: The agency proactively offers references and gives you direct contact information so you can have a private conversation without the agency present.
Red flag: “We can share some testimonials.” Written testimonials are curated marketing material. They are not the same as a real conversation where you can ask uncomfortable questions.
Questions to ask references: How closely did the final product match what was agreed? Did the project finish on time and on budget? Would you hire them again? What was the biggest surprise, positive or negative?
The scorecard in practice
Rate each agency on all seven criteria using a simple 0/1/2 scale: 0 for red flag, 1 for mixed signals, 2 for green flag. Any agency scoring below 10 out of 14 should be reconsidered. Any agency with a 0 on code ownership or contract structure should be disqualified regardless of their total score.
This is not about finding a perfect agency. It is about filtering out the structural risks that cause projects to fail for reasons that have nothing to do with code quality.
If you want to see how these seven criteria look in practice, Buldtech operates on fixed-price contracts with Day-10 demos, code ownership from day one, and a 30-day bug guarantee. Book a free scope call to see the structure firsthand.
Your next step
Download the MVP Scope Clarity Checklist, which includes a printable version of this 7-point scorecard. Use it to evaluate every agency you talk to, and compare scores side by side before making your decision.