How to choose a web design company

Every studio shows its best six projects, photographed well, with the client's logo attached. You cannot tell from a screenshot whether the site launched on time, whether it converts, whether the client is still speaking to them, or whether the designer who made it still works there.

You also cannot tell who did it. Subcontracting is routine and rarely disclosed — the work in the portfolio may have been produced by a freelancer who is no longer available, or by an offshore team you will never meet.

So use the portfolio for one thing only: has this studio solved a problem shaped like yours? An agency with ten restaurant sites and no e-commerce is a poor fit for a catalogue, however good the restaurants look. Then stop looking at pictures and start asking questions.

These are the ones that cost people the most and get asked the least. Ask them before you discuss design, because a wrong answer here makes everything else irrelevant.

• —In whose name will the domain be registered, and when do I get the login? The only acceptable answer is your business and immediately. A domain held by an agency is leverage over you, and we have been called in to unpick this more than once.
• —Whose hosting account will the site live in? Same test. 'We host it for you' is fine as a service and not fine as the only arrangement — you should hold credentials to the account you are paying for.
• —Who owns the code, and will I get the repository? You should. Ask what happens to it if you stop working together.
• —What exactly do I receive at handover? Push for a list: accounts, credentials, DNS records, source code, editing documentation. If the list is vague, the handover will be too.

These separate a studio with a process from one that will improvise on your budget.

• —Who writes the copy? If the answer is you, understand that you have just taken on the hardest part of the project, and that 'client to provide content' is the single most common reason websites sit unfinished for months.
• —How many individually designed templates, as opposed to pages? Twenty pages from three repeating templates is a fraction of the work of eight bespoke ones. Quotes that look wildly different are usually answering this question differently.
• —What happens when I want a change after approval? Ask for the revision process and what triggers a fee. 'Two rounds included' plus an expensive third round is a business model, and you should know about it now.
• —Who will I actually talk to, and did they do the work? Ask whether the people in the pitch are the people building it. A layer between you and the work is a layer where intent gets lost.

A launch is the start of the site's life, and most conversations stop before covering any of this.

• —What performance target does the site ship against, and how will you show me? A studio that works to a budget will name a number — a Lighthouse score, a load time on mobile — and can demonstrate it. One that has never thought about it will talk about how fast their sites feel.
• —Is analytics configured before launch or after? Before is the correct answer. Sites that get instrumented later have a blank first month exactly when the most useful data is available.
• —What is included in maintenance, and what does it cost? Get patching, backups, monitoring and content updates itemised. 'We'll look after it' is not a scope.
• —Can I edit it myself, and what will you show me? Ask for a walkthrough recorded against your own site rather than a generic manual.

Not warning signs to weigh up. Reasons to stop.

• —A guaranteed Google ranking. Nobody can promise a position, and anyone who does is either misleading you or planning something that will be corrected — with your domain carrying the consequence.
• —A price quoted before anyone asked what your business does. That number is either padded to cover the unknown or about to grow.
• —Reluctance to name who does the work. There is no good reason for this to be evasive.
• —Their own site is slow, or their contact form does not work. This is the one piece of their work you can evaluate completely. Test it on your phone before the first call.

Line up the totals last. First line up the scope: template count, who writes copy, whether strategy is included, performance target, revision terms, and the ownership answers from the top of this article.

You will usually find the quotes were never comparable. The cheap one excluded copywriting and strategy; the expensive one included a content system you may not need. Once the scopes are aligned, the price difference becomes a real decision instead of a guess.

One more test, and it is the most reliable of the lot: which studio told you something you did not want to hear? A studio that pushes back during the sales process — on your timeline, your budget, your idea for the homepage — will push back later, when it matters. One that agrees with everything is selling you a website, not a result.