Most advice about choosing a web agency is a list of qualities: creative, rigorous, responsive. None of it is checkable before you have signed, which makes it useless at the moment you actually need it.
There is a harder starting point. According to the Standish Group CHAOS 2025 report, 67 % of digital projects fail partially or totally, and the first cause identified is not technology or budget. It is the choice of provider. So the useful question is not what makes a good agency, it is which checks you can run before committing.
1. Specialisation you can verify, not claim
Every agency says it masters modern technology. Ask which stack, why that one for your project, and what the trade-off is. An agency that cannot name a downside to its own recommendation is applying a habit rather than a reasoning.
Ask also about your sector. B2B problems are not e-commerce problems and neither resembles a SaaS product. Someone who has never sold what you sell will discover your constraints at your expense.
2. References you can actually call
A portfolio shows the best day, with the best client, sometimes years ago and sometimes by someone who has since left. It says nothing about what happens when a deadline slips.
Ask for two references on projects comparable to yours, and call them. One question does most of the work: what went wrong, and how did they handle it? Every project has a bad moment. The answer tells you far more than a showcase.
3. What the price covers, and what it hides
Ask what the quote excludes before asking what it includes. Content writing, photography, training, redirects from the old site and support after launch are the four lines where proposals diverge without anyone noticing.
Then look for the missing recurring costs. Maintenance, hosting and updates typically represent 15 to 25 % of the project cost per year. A budget that ignores them is not cheaper, it is incomplete.
On payment terms, be wary of anyone asking for the full amount upfront. A staged schedule tied to deliverables protects both parties and is the norm.
4. What happens after launch
This is the criterion people skip and then regret. Around 75 % of a website's value is created after it goes live, through optimisation, fixes, new features, content updates and continuous search work.
An agency that disappears at launch has sold you an asset that starts depreciating immediately. Ask what month six looks like, concretely, and whether monthly support contracts exist.
5. Size, and the responsiveness test
Freelancers and small studios of one to ten people bring proximity and speed, but can run short on capacity or on specific skills. Mid-sized agencies of ten to fifty often strike the best balance: enough resource for an ambitious project, small enough that yours genuinely matters.
There is a free test available to you right now. Time how long the reply to your first message takes. If it is five days before the project has even started, it will not improve once it has.
6. The four questions to ask before signing
What is your year-on-year client retention rate? Good agencies keep 70 to 85 % of their clients, and a low number says more than any case study.
Who owns the code, the domain, the hosting account and the analytics at the end? If leaving means starting over, the quote is not the real price.
What exactly happens in month six?
And who specifically will work on this project? The people in the pitch are not always the people who build.
Start with your own numbers
Before you brief anyone, it helps to know where your current site stands. Our free audit tool measures it in seconds, with no signup and no email address: tags, indexability, mobile, structured data, response time.
With those figures, a first conversation stops being a sales pitch and becomes a discussion about specific problems. If you want ours, describe your project and we reply in writing, by email, with no meeting.
