Most founders who hire a software development company are not engineers. They are relying on a presentation, a portfolio, and gut feel — and the same gut feel that closed their last deal often fails them here. Software vendors have spent years optimizing for how they look in a sales process. Your job is to see past that.

You do not need to understand code to evaluate whether a software company is the right fit. What you need is a structured vetting process that tests their judgment, their honesty, and their track record with clients who looked like you. Here is that process.

Start with their discovery process, not their portfolio

The most revealing thing a software company shows you is not their case studies — it is how they approach the problem you bring them. A strong partner asks clarifying questions before proposing anything. They push back if your scope seems undercooked. They ask about your users, your business model, your success metrics. A company that immediately moves to pricing or timelines without understanding your problem first is optimizing for the sale, not the outcome.

Ask them: "What questions do you still have after our first conversation?" A vendor with no follow-up questions has not done the thinking required to scope your project honestly. Related: how to write a software project brief that gets accurate quotes — preparing a tight brief separates serious vendors from opportunists before the first meeting.

Check references the right way

Every agency gives you references who will say positive things — that is the point of curated references. What you want is to find clients they did not give you. Search LinkedIn for employees of their past clients and message them directly. Look for their work in ProductHunt, AppSumo, or app stores and read user reviews. Ask in founder communities whether anyone has worked with them. The references they control tell you what they want you to hear. The ones they don't control tell you the truth.

When you do speak with references, ask: "What would you change about how the engagement was structured?" and "Was the final scope what you agreed to at the start?" The answers reveal more than any testimonial.

Evaluate their technical communication, not their jargon

You cannot read code — but you can assess whether they can explain technical decisions in plain English. Ask them to walk you through a recent architectural decision on a past project and why they made it. If their explanation requires you to know what microservices, Kubernetes, or GraphQL means before you can understand the tradeoff, they have failed to communicate. Good engineers are excellent explainers. They make complexity legible.

Also assess how they handle uncertainty. When you ask about a technical risk you've identified, a strong vendor says "here's how we'd manage that" with specifics. A weak vendor says "don't worry, we've done this before." Reassurance without reasoning is a red flag.

Understand how they handle change — before anything changes

Every software project changes. Scope shifts. Priorities evolve. New information arrives after development starts. The question is not whether changes will happen — it is whether your contract and their process handles them in a way that protects you. Ask directly: "What happens when we discover mid-project that our original scope was wrong? Walk me through a real example." Their answer reveals their process maturity and their honesty about how projects actually unfold. For a deeper treatment of this, see how to prevent scope creep on a software project.

How TTGC approaches client vetting

At Through The Glass Creatives, Ravve Jay Prevendido — CEO and AI/dev engineer — approaches every new engagement with a two-week discovery sprint before any production work begins. The goal of that sprint is not to start building — it is to define exactly what's being built, in language both the client and the technical team can agree on. Clients who have worked with TTGC describe this as the point where the project stopped feeling risky. When the brief is right, the execution is faster and the relationship is smoother. No discovery, no engagement.

TTGC is also unusual in that its principals — Ravve and Mherie — are the people who do the work. There is no handoff to a junior team after the sale. That named accountability is worth something that a larger agency cannot offer.

The best software partner makes your project feel less risky, not more exciting. If a vendor's pitch is generating more energy than clarity, slow down.