Twelve questions to ask a web designer, and what a good answer sounds like

The twelve questions below are not the point. The point is what the answers sound like. You can find a list of questions to ask a web designer on a hundred agency blogs; almost none of them tell you how to tell a good answer from a fluent one, which leaves you finishing a sixty-minute call with a page of notes and no more clarity than you started with.

So every question here carries three things: why it matters, what a good answer sounds like, and what should make you pause. Some of the answers we would give ourselves are in here. So are the questions we find uncomfortable.

The twelve questions

1. 01

   Who exactly will do the work, and who will I talk to each week?

   The most common disappointment in this industry is meeting a senior person in the sales call and getting a junior (or a subcontractor on another continent) for the actual project. That is not automatically bad. It is only bad when nobody told you.

   A good answer sounds like

   Names, roles and a straight answer about subcontracting. "I do the design, my developer Sam does the build, and you talk to me every Thursday." An offer to introduce you before you sign is the strongest possible version of this.

   Be careful if you hear

   "Our team handles it." A refusal to name anyone, or an evasive answer about whether the work is subcontracted. If they will not tell you who is doing it, you cannot check whether they have done it before.

2. 02

   What happens before anyone opens a design tool?

   Design that starts before anyone has asked who your customers are and what makes them hesitate is decoration. This is the question that separates studios who solve a business problem from studios who make things that look nice.

   A good answer sounds like

   A named phase with named outputs: interviews, a look at your analytics and enquiry history, a competitive read, and something written down (a brief, a sitemap, a set of priorities) that you sign off before design starts.

   Be careful if you hear

   Mockups offered in the first week, or unprompted in the pitch. Speculative design before discovery is a sales tactic. It also means the first thing you judge them on is a guess about your business.

3. 03

   Can I see a site you built that has been live for two years?

   Portfolios are launch-day photographs. This question asks a different thing: did the work survive contact with a real business? Has the client been able to maintain it? Is it still fast? Did anyone answer the phone in year two?

   A good answer sounds like

   A link, without hesitation, and ideally with the sentence "here is what has changed since: the client has added forty pages themselves." Better still: an offer to introduce you to that client.

   Be careful if you hear

   Only recent work. A portfolio where everything launched in the last eight months either means the studio is new (fine, if disclosed) or that older work has been quietly removed, which is worth understanding.

   While you are there, open the two-year-old site on your phone on cellular data and time it. You will learn more in fifteen seconds than in the next half hour of conversation.

4. 04

   What is in scope, what is out, and how are extras priced?

   Scope ambiguity is the number one source of friction between clients and studios, and it almost never surfaces in week one. It surfaces in week six, when you ask for something reasonable and get a change order.

   A good answer sounds like

   A written scope in page templates rather than pages, an explicit list of what is excluded, and a rate or a flat-fee process for anything additional. "Anything outside this gets quoted flat before we start it, and you approve it in writing."

   Be careful if you hear

   "We are flexible, we will figure it out as we go." This sounds generous and reliably ends in an awkward conversation. Flexibility without a baseline is not flexibility, it is an undefined contract.

5. 05

   How many revision rounds, and what counts as one?

   The definition matters more than the number. If one email with one comment burns a round, two rounds is stingy. If a round is a consolidated set of feedback at a named checkpoint, two rounds is plenty and keeps the project moving.

   A good answer sounds like

   A number, tied to checkpoints, with a definition. "Two rounds: one after the direction is presented, one before launch. Collect your feedback, send it as one list, and we work through all of it."

   Be careful if you hear

   "Unlimited revisions." It is the most oversold promise in this industry. Nobody works for free, so it is priced in somewhere, and a project with no definition of done drifts until one side gives up.

6. 06

   When we are done, what exactly do I own?

   The question that costs the most to get wrong. Businesses discover the answer years later, at the worst possible moment, usually when they want to leave.

   A good answer sounds like

   "All of it. The domain is registered in your name, you get the source code and the design files, and every account we open is yours." Some studios put this in the contract without being asked, which is the strongest signal you will get all call.

   Be careful if you hear

   The domain registered to the agency "for convenience". Design files withheld or charged for. Code you cannot take elsewhere. A proprietary platform you can only stay on. Any of these means leaving costs you the site.

   Ask the follow-up too: "If we part ways next year, what do I walk away with, and how long would it take you to hand it over?" The pause before the answer tells you most of what you need to know.

7. 07

   Who is writing the words?

   More website projects miss their launch date because of unwritten copy than for any other reason. It is nobody's favourite job and it is very easy for both sides to assume the other one has it.

   A good answer sounds like

   An unambiguous assignment with a date. "You are writing it and it is due in week three, or we write it and it costs this much." Either is fine. Being decided is the whole thing.

   Be careful if you hear

   Silence, or "you can just send us your content". That sentence is how a four-week project becomes a four-month one, and the delay ends up looking like your fault because technically it is.

8. 08

   What performance are you building to?

   Speed is not an aesthetic preference, it is revenue. Conversion rates fall roughly from 3% at a one-second load to under 0.5% at five seconds, and around 53% of mobile visits are abandoned after three seconds. Vodafone A/B tested two visually identical pages and found a 31% improvement in load time produced 8% more sales.

   A good answer sounds like

   A named target, measured on real visits. "Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, Cumulative Layout Shift under 0.1, at the 75th percentile, and we will show you the report before launch."

   Be careful if you hear

   "It will be fast, we optimise everything." Also worth noticing: a portfolio of heavy, animation-loaded sites and no measurements anywhere. Beautiful and slow is a choice somebody made on your behalf.

9. 09

   What accessibility standard are you building to?

   This is now both an ethical and a legal question. Plaintiffs filed 4,928 web accessibility lawsuits in 2025, and roughly two-thirds of the businesses sued had revenue under $25 million. Meanwhile the 2026 WebAIM Million found detectable failures on 95.9% of the top million home pages, averaging 56 errors per page, so the odds that a studio has simply never thought about it are high.

   A good answer sounds like

   A named standard and a method. "WCAG 2.1 AA as the floor, 2.2 AA as the target. We check colour contrast, keyboard navigation, form labels and alt text as we build, and we run the site through an audit before launch."

   Be careful if you hear

   "We can add an accessibility widget." Overlay widgets have repeatedly failed to protect businesses in court and are actively disliked by many disabled users. An overlay offered as the answer means accessibility was never in the build.

10. 10

    What happens to my Google rankings when we launch?

    If you have an existing site that gets found, a redesign can quietly destroy months of traffic. Every URL that changes without a redirect is a page Google has to rediscover and re-earn.

    A good answer sounds like

    A migration plan, named as such. "We map every existing URL to its new address, set 301 redirects, keep the pages that are earning traffic, submit a new sitemap, and watch Search Console for four weeks after launch."

    Be careful if you hear

    "The new site will be better optimised, so it will rank higher." That is a hope, not a plan. Ranking better is a goal; not losing what you already have is an engineering task, and they are not the same job.

11. 11

    What can I change myself, and what will I have to come back for?

    There is a right answer here and it is not "everything". A site where you can edit every pixel is a site you can break. A site where you cannot change your own opening hours is a site that will hold you hostage in small, annoying ways for years.

    A good answer sounds like

    A clear line, drawn deliberately. "You edit text, images, prices, hours, staff and posts. Layout and structure come to us, because that is what keeps it from drifting." Plus a handover session and something written down.

    Be careful if you hear

    "You can do everything yourself" with no training offered, or its opposite: every small change routed through a support ticket and billed. One of those is neglect and the other is a business model.

12. 12

    What happens after launch, and what does it cost on day 31?

    The end of a project is the moment the incentives change. Ask what the arrangement is before you are in it, not after.

    A good answer sounds like

    A defined support window with a defined scope, and honest pricing after it. "Thirty days of bug fixes included, and anything broken is on us. After that it is hourly, or a retainer if you want ongoing work, and the retainer is month to month."

    Be careful if you hear

    A mandatory long retainer attached to a site that will not change, or no post-launch answer at all. Also worth asking: "What is your typical response time when something breaks on a Saturday?" Nobody rehearses that one.

The red flags that should end a conversation

Everything above is a judgement call. These are not. Any one of them is enough to walk away, and none of them get better later.

• A price before any discovery. A number produced before anyone understands the job is a number that will change, and the change will not be downward.
• Manufactured urgency. A discount that expires Friday, or a "slot" that will be gone. Studios with a real pipeline do not need to rush you into a decision about the next three years.
• Refusal to name who does the work. There is no legitimate reason for this.
• No references, or only very recent ones. Ask again. If it is still a no, that is the answer.
• Ownership retained by them: domain, code, or design files. This is the one that costs the most to unwind.
• Guaranteed number-one rankings. Nobody can promise this. Anyone who does is either lying or does not know enough to know they are.
• Only aesthetics in the whole conversation. If nobody has mentioned speed, mobile, or what you want visitors to actually do, you are buying a picture of a website.
• They never ask about your customers. The best signal in the entire process is who is asking the questions.

What a good studio will ask you

A one-way interview is a bad sign. If you get through an hour and they have not asked you anything difficult, they are selling rather than scoping. Expect some version of these.

• What do you want someone to do on the site: call, book, buy, apply, visit?
• How do customers find you today, and which of those channels actually produces work?
• What do people ask you before they buy, every single time?
• Who else were they considering, and why did they pick you?
• Who signs this off, and are they in the room?
• What has to be true in a year for this to have been worth the money?
• What is the budget, honestly, and what were you hoping it would be?

How to run the call

1. Send three or four of these questions ahead of time. You are hiring for thoughtfulness, not improvisation, and the ones who prepare are the ones who will prepare for your project too.
2. Ask the ownership question early, while everyone is still relaxed. The answer changes tone if it comes after a pitch.
3. Take notes on specificity, not enthusiasm. Enthusiasm is free.
4. Ask the same three questions to every studio you talk to, so you have something you can actually compare.
5. Ask for the two-year-old reference site and open it on your phone before the call ends.
6. Sleep on it. Anyone who makes that difficult has told you something.

If you are still working out what any of this should cost, the companion piece has the numbers: what a small business website actually costs in Austin. And if you are not yet sure you need a new site at all, run the ten-point check first. A fair number of "we need a redesign" conversations turn out to be three fixable problems.

Sources

• UsableNet: 2025 Year-End Digital Accessibility Lawsuit Report, for the 2025 filing count and defendant profile.
• WebAIM Million, 2026, for the 95.9% failure rate and 56.1 errors per page.
• web.dev: The business impact of Core Web Vitals, for the Vodafone A/B test.
• Portent: Core Web Vitals impact on traffic and conversions, for load time against conversion rate.