You've reached the point where your current site can't keep up, or there isn't one at all, and you know it's time to commission a proper one. You search Google for 'web design agency,' and within three seconds you have thirty results that all look fine, all promise roughly the same things, and none of them tell you how to choose between them. This is a decision you make rarely and one that costs you, so it's worth making with a clear head.
In this article we go through everything that matters in the choice: how to first decide what kind of provider you actually need, where to look and how to build a shortlist, what to ask in the first call, what warning signs to spot in a proposal, and what you should walk away with at the end so you aren't left stuck. By the end you should be able to walk into a conversation with any agency without nerves and without relying on luck.
01.First: do you need an agency, a freelancer, or can you do it yourself?
Before you look at who, it's worth looking at what you actually need. There are three paths, and each has its moment.
An agency makes sense when the site is part of something larger: you need brand and web that speak the same language, several pages or touchpoints, someone who thinks about strategy rather than just executing. The real advantage of an agency isn't that it has more people, it's that it gives you a single point of responsibility and coherence from the logo to the last page.
A good freelancer fits when you have one clear, well-defined deliverable, for example a campaign page or a focused refresh, and you can hold the strategy and coordination yourself.
DIY, on a builder like Wix or WordPress, makes sense only in a few real situations: a solo operation, with a single channel, in an experimental phase, or on a zero budget for the moment. It's a decision in its own right, one we write about separately, but in short: a drag-and-drop builder gives you a layout, not a system and not a conversion strategy.
If your business has ambitions to grow and the site needs to bring in clients rather than just exist, you're already in the territory where a professional is worth it.
02.Where to look and how to build a shortlist
Once you know you want an agency, the good sources are roughly the same: a Google search with your city ('web design agency Bucharest,' 'Cluj,' 'Timisoara'), recommendations from people you trust, and the platforms where designers show their work, Behance, Dribbble, and LinkedIn. A direct recommendation is still the best signal, because it comes with a real experience behind it.
When you have a list of five or six names, cut it to three using a few simple criteria. Look at the portfolio, but look for case studies, not just pretty images: what problem the client had, what was done, what came out. Check whether they've worked with someone in your field or a close one. Read the reviews and, if you can find them, the testimonials with a real name behind them. And notice one subtle thing: do they ship systems, or just isolated logos and pages? We explain that difference below.
03.The questions to ask in the first call
The first call is where you find out whether the agency has a process or improvises. A good agency asks you first: about the business, your audience, what you want to achieve. In fact, the questions they should be asking you are the same ones we use in discovery, before any design. If it jumps straight to 'we'll build you a modern site' without asking anything, you have your answer.
Turn it around and come in with a short list of questions:
- What is your process, step by step? You want to hear discovery, strategy, design, development, revisions, launch. If the process starts straight at design, the foundation is missing.
- What deliverables do I receive at the end, exactly? Files, formats, style guide, platform access. Ask for the list in writing.
- How long does it take, and what are the stages? A timeline with clear stages shows they've done this before.
- How many revisions are included? And what happens if I want more.
- What does post-launch support look like? The first month is critical. You want to know who answers and for how long.
- Who owns the files and rights at the end? The right answer is: you.
- What platform do you build on, and why? And, importantly, can I manage the content myself after launch?
- Will you show me references or clients in my field? A serious agency lets you talk to a past client.
- Do you know the rules of my niche? If you run a medical, dental, or aesthetic practice, advertising rules matter.
- How do you measure whether the site worked? You want to hear about goals and metrics, not just 'it'll look good'.
Confident, concrete answers with examples are a good sign. Vague or irritated answers to the question about rights and files are a sign you should stop.
04.How to read a proposal without getting lost
After the call comes the offer. This is where you see the difference between an agency that owns its scope and one that leaves everything fuzzy so it can bill extra later. Read five things carefully.
Clear scope. How many pages, what features, what's included and, just as important, what isn't. A good offer explicitly lists what's left out.
Stages and dates. Milestones with dates, so you know where you are along the way and don't discover delays only at the end.
Payment terms. Payment usually happens in stages tied to milestones, a deposit up front and the rest on delivery or per milestone. It's fair for both sides and it protects you.
File rights. Look for the clause that says the source files, the domain, and the hosting are yours at the end. If it's missing, ask, and ask for it to be written in.
Post-launch support. What happens in the first month, what maintenance costs afterward, and who answers when something breaks.
If the offer covers all five clearly, you're looking at an agency that has done this before. If it skips half of them, the conflict three months out is already written into it.
05.What a healthy handoff looks like (and what's missing when it isn't)
The last filter, and the most often ignored, is what you actually receive at the end. Many projects close with an 'it's live' and a link, and the client discovers a month later that they can't change anything themselves and own nothing.
A healthy handoff means, at a minimum:
- The source files and full platform access, in your name.
- The exports you need in practice: the logo in vector, light and dark background variants, favicon, single-color version.
- A style guide, a short document with the brand usage rules: colors, typography, spacing, what's allowed and what isn't. Without it, the brand gets applied inconsistently from the first post made in a hurry.
- A short training or walkthrough video, so you can manage the content yourself.
The absence of these is a red flag in itself. A logo delivered as a plain JPG, with no system and no rules, carries limited value: it looks good on delivery day and degrades with every later use. A brand is a system, not a picture, and the handoff is the moment the system is handed to you, not just its image.
In short
Choosing an agency is, at heart, an exercise in clarity. You first decide what kind of provider you need, you look at who ships a process and a system rather than just pretty images, you walk into the first call with your questions, you read the offer carefully for scope and rights, and you make sure you leave with everything you need to be independent. The rest is the fit of chemistry and budget, which matter too, but come after you've checked the foundation.
At GOODGLYPH we work on exactly these principles: discovery before design, a single point of responsibility from brand to web, the source files and a usage guide in your hands at the end, and availability after launch. If you're getting ready to commission a site and want an honest conversation about what fits you, let's talk.
Frequently asked questions
It depends on complexity. For a site that's part of a brand system, with several pages and touchpoints, an agency gives you coherence and a single point of responsibility. For one clear, well-defined deliverable, a good freelancer can be enough. DIY only makes sense for a solo operation, with a single channel, in an experimental phase or on a zero budget for now. We cover the build-or-buy decision in a separate piece.



