Questions to ask a web designer before you sign

Owners evaluate web designers on their work samples. That is understandable and it is close to useless. A portfolio is a selection, chosen by the seller, and none of it shows what the client went through to get there.

The useful questions are boring. They are about ownership, responsibility, and what happens when things go wrong. A good vendor answers them immediately and is glad you asked.

Which questions actually matter?

Ten, and you can ask them in one email.

Who will the domain be registered to? The only correct answer is your business, in an account you control.

What do I keep if we stop working together, and how fast? Files, content, analytics, and profile ownership, delivered in a usable format, within a stated number of days.

Who writes the copy? If the answer is you, that changes your timeline. If the answer is them, ask to see writing they have done for a business like yours.

What is not included? Photography, copy, hosting, maintenance, and post-launch changes are the usual gaps. Get the exclusions in writing, because that list is where the second invoice comes from.

How many revision rounds, and what counts as one? A round should be a consolidated list, not an individual change.

What do you need from me, and by when? A vendor who cannot answer this has not thought about your part of the schedule, which is where projects actually run long.

Who hosts it, and can I log in? Both parts matter.

What happens in the first 90 days after launch? If the answer is nothing, price that in.

How will we know if it worked? Anyone who cannot describe what they will measure is selling appearance.

Can I talk to a client who left you? The reaction to this question is more informative than the answer.

A vendor who is uncomfortable with these questions is telling you exactly what the next year will feel like.

How should I read their case studies?

Sceptically, and there is peer reviewed work that tells you why.

Kohavi, Deng, Longbotham and Xu, writing in KDD 2014 from thousands of controlled experiments at Microsoft, Amazon, Booking.com and LinkedIn, devote a rule to exactly this. On published web case studies, verbatim: "The quality varies... Was it peer reviewed? Was it properly run? Were there outliers? Was the p-value low (we've seen tests published where the p-value was > 0.05, which is normally considered not statistically significant)." Their conclusion is blunter still: "What works in one domain may not work in another."

They name examples. One is the widely repeated advice to use the word "free" in your copy, attributed to Neil Patel, which the authors cite as a result that does not travel between domains. A result from somebody else's website is a fact about somebody else's website. It is not a forecast for yours, and a designer presenting it as one has either not read the source or is counting on you not to.

So do not ask whether they have case studies. Ask three questions of the one they show you. What was the starting number. Over what period. What else changed at the same time. A vendor who answers all three has done real work. A vendor who shows you a percentage with no baseline is showing you a decoration, and a multiplier without a starting number is the single most common tell in this category.

What answers should worry me?

Four patterns, and each is common enough to name.

Vagueness about ownership. "We handle all of that for you" is not an answer. It is the answer that becomes a problem in two years, and it is the reason to be specific about what you own before you sign anything.

A quote with no scope. A single number with no page count, no content responsibility, and no exclusions is not a quote, it is a starting position. If you cannot see what drives the price, you cannot compare it to anything, which is most of the difficulty in working out what a small business website actually costs.

Guarantees about rankings, or a proprietary score. Nobody controls search results. Google says so itself in its guidance on generative AI features: "Be wary of third-party tools that promise ranking success or claim to use 'internal' Google metrics. No third-party tool has access to our internal ranking or AI systems." The same document tells you to prioritise ordinary SEO over what it calls AEO and GEO hacks. And when Google launched its Search generative AI reports in Search Console in June 2026, the dimensions were impressions, pages, countries, devices and dates. Clicks, click-through rate, position and the user's prompt are explicitly absent. If a proposal shows your rank inside an AI answer, it is reading a source that does not exist.

A design-first pitch. If the entire conversation is about how the site will look and none of it is about who your customers are or what you want them to do, you are buying decoration. The brief matters more than the mockups, and the brand decisions that come before a website are the part of the brief a designer cannot invent for you.

How do I check whether the quote covers my actual site?

What about accessibility, and who is telling me the truth?

Ask whether it is built in. Then be careful with the compliance claims attached to the answer.

The technical standard itself is public and specific. The World Wide Web Consortium sets a contrast ratio of 4.5 to 1 for normal text and 3 to 1 for large text at level AA, and WCAG 2.2 adds a minimum target size of 24 by 24 CSS pixels at AA. A designer can build to that at no extra cost during a build, and retrofitting later is where it gets expensive.

The legal framing is where vendors overreach. The Department of Justice published a web accessibility rule in 2024 adopting WCAG 2.1 Level AA, and that rule binds state and local government entities under Title II. It is not a rule for private businesses, and the Department's separate guidance for businesses under Title III names no technical standard. Anyone selling you a build on the claim that a named federal standard legally binds your restaurant is going past what the documents say. I am not your attorney and this is not legal advice. Ask for the requirement in writing, then ask your attorney what applies to you.

What should I bring to the conversation?

Enough that the quote can be real rather than a guess.

Your service list, your location list, and a rough page count. Your current traffic and inquiry numbers. Your logins, or at least the knowledge of who holds them. Whatever photography and copy already exists.

Bringing this changes the conversation immediately. You get a scoped number instead of a range, and you find out quickly whether the vendor is listening or reciting a package. What to prepare before a web project starts is the same list, and it is worth assembling before you talk to anyone rather than after you hire them.

Also bring the one thing you want the site to do. Booked appointments, quote requests, calls, catering inquiries. A designer given a single priority builds toward it. A designer given "a professional presence" builds something that looks professional and asks for nothing.

What should I ask about after launch?

The part almost nobody negotiates, and the part that determines whether the site stays useful.

Ask what maintenance covers, what it costs, and whether it is required to keep hosting. Ask how small edits get requested and what the response time is. Ask whether tracking will be set up and who verifies it works.

Then ask how you will see what visitors actually do. Not a monthly slide of traffic going up, but something that shows behavior. Watching real visits is cheap and immediately useful, and what session recordings show in the first thirty minutes is usually more actionable than a quarter of reporting.

One question that sits outside the website and belongs in the same conversation: does the vendor understand your Google Business Profile, which for most local businesses produces more customers than the site does. If they cannot explain how to pick a primary category when several fit, they are a web shop, not a local marketing partner. That is fine, as long as you know which one you are hiring.

What to do this week

Send the ten questions by email to everyone you are considering, in one message, before any meeting. Email answers are specific in a way sales calls are not, and you get a written record.

Run the page count arithmetic and put your number in the same email. It reframes every reply you get.

Compare the replies on speed and directness. The vendor who answers in a day with plain answers is showing you their operating standard.

Then ask the ownership question again, separately, and make sure the contract language matches the email answer. Contracts and sales conversations disagree more often than they should.

Then decide what you are actually buying: repairs to a site you keep, or a new site. Those are different vendors sometimes, and always different budgets. Fixing fundamentals on an existing site is plumbing work. A full build is the house.

Questions about a proposal you received?

Email me at eric@seod.com with the proposal or quote, exactly as it was sent to you. I will tell you which of the ten questions it already answers, which it dodges, and what I would ask before signing. I will do this whether or not the vendor is a competitor of mine, and I will tell you when a proposal looks good.

More on hiring and scoping in the web design and branding library.