Every agency portfolio looks good. That's what a portfolio is for. So choosing between three shortlisted firms on the strength of their case study pages is close to guessing, and it's why so many website projects end in frustration.
Better questions exist. Here are the ones we'd ask if we were buying rather than selling.
Ask who will actually do the work. In plenty of agencies the person who charmed you in the pitch is not the person who writes a line of your site. Ask to meet the developer and the writer. Ask how many projects they run at once. A small team with capacity beats a large one where you're the fourth priority.
Ask what happens when it goes wrong, because at some point something will. What's the process when a deadline slips or a requirement was misunderstood? An agency that has never had a difficult project either hasn't done many or isn't being straight with you. The useful answer is a specific story about a project that went sideways and what they changed afterwards.
Ask what you'll own at the end. Domain, hosting, source code, content, analytics accounts. Get it in the contract. Anyone reluctant to hand over the keys is describing how the relationship will feel in three years.
Ask what technology they build on and why. There's no single right answer, but there are wrong ones. A proprietary in-house platform means nobody else can maintain your site. The question behind the question is: can a competent developer who isn't you pick this up?
Ask how they measure success. If the answer is about design awards or "engagement", be careful. The right answer references enquiries, calls, sales — your commercial outcomes — and includes how those will be tracked from day one.
Ask about content. Websites are mostly words, and most projects are delayed because the words never arrive. Find out who is writing them. If the answer is "you'll provide the copy", price in the reality that you probably won't, or budget for a writer properly.
Ask what happens after launch. Who applies security updates, who fixes the contact form when it breaks, what's the response time, what does it cost? A website is a living thing. An agency with no answer here has sold you a project, not a partnership.
Ask for two references, ideally including a client who has left. Then actually phone them. Fifteen minutes on the phone tells you more than any pitch deck.
Now the red flags. A quote produced without a conversation, because a firm that hasn't asked about your business is selling you a template. Vagueness on price, followed by a stream of change requests once you're committed. Pressure and artificial deadlines. Guaranteed rankings, which nobody can honestly promise. Disparaging your current site in the first ten minutes rather than being curious about why it is the way it is. Design mockups before any discussion of what the site needs to achieve. And, most quietly damaging, an unwillingness to disagree with you: an agency that says yes to everything in the sales process will say yes to everything during the build, and you will end up with a website designed by committee.
Some good signs, for balance. They ask about your customers before your colour preferences. They tell you something you didn't want to hear. They're specific about numbers from past work. They set out the process, including your obligations and the points where a delay from you moves the date. They're comfortable being compared with competitors. And they can explain technical decisions in plain English, because clarity of explanation usually reflects clarity of thinking.
One last piece of advice: don't choose on price alone, but don't ignore it either. The cheapest quote is often the most expensive over three years, because the corners cut in month one become the rebuild in year two. What you're really buying is judgement — someone who will steer you away from the expensive mistake you were about to make. That's difficult to see in a portfolio, which is exactly why the questions matter more than the pictures.