What is the website discovery process for? It's the stretch before design where the people building your site work out what it has to say, to which buyers, and what it has to prove, then write that down so design has something to be checked against. For an established industrial company, the questions that matter are the ones only you can answer: who actually signs off on a purchase, what wins a quote over another supplier, and what your sales team sends instead of the website. What you should get back is a short set of decisions, not a slide deck. If there's nothing in it you could disagree with, nothing was decided.
Most of what you'll read about discovery is written by agencies, for any kind of organization, or for software products. Their question lists could apply to a university or a bakery, and don't say which questions carry the weight for a manufacturer or distributor, or how to judge the result before the design invoice arrives.
What is a discovery phase actually deciding?
Three things, in this order. What the company is and who it's for, in words a buyer would use. Which pages the site needs, and what each one has to do. And what has to happen around the build: who writes what, which current pages and addresses survive, and how you'll know afterward whether it worked.
Everything else in discovery serves those three. Goals like "increase leads" or "improve the user experience" sound like decisions and aren't, because no design choice follows from them. Settle them early for the reason the redesign timeline gives: an open question at the start returns as a revision round, when it costs more to change.
Which website redesign discovery questions can only your company answer?
An agency can find out a lot without you. It can't find these. Expect to be asked them, and if nobody asks, raise them yourself.
Who buys. Who finds you first, who signs the purchase order, and who can stop it. For a contract manufacturer, that's often three people: an engineer who specifies, a buyer who compares, a quality manager who can rule you out. Answer from real accounts you can name, not from a persona.
What wins. Why a buyer picks you over another supplier quoting the same work, and why they sometimes don't. You and your estimator may give different answers, and that gap is worth collecting before anyone writes a headline.
What gets sent. What your sales team attaches to a quote instead of a link, and which pages, if any, customers forward to colleagues. A capabilities PDF that gets attached every week is a rough outline of a page the site is missing.
What's changed. Where the business has moved on and the site hasn't, and the work you want more of. The facts only your company can supply are a short list, already set out in what a redesign needs from you. Bring it written down.
You'll answer these better with records than from memory: a sample of real quote requests, the attachments your team sends most often, access to your Search Console and analytics, and the name of the one person who can approve the result.
What should the agency know before it asks you anything?
A fair amount, and the first meeting tells you whether they did the reading. Before discovery starts, an agency can read your whole site, see how it loads on a phone, look at what your competitors' sites claim, and, once you've granted access, see which pages already get search traffic and for what.
A few questions are reasonable later and a warning sign in the first hours:
- "Send us a few websites you like." Taste belongs after the message is settled, not before.
- "Which pages do you want?" The page list is something discovery should produce, not something you should bring.
- "Fill in this content questionnaire." As one input, fine. As the main way discovery happens, it's how the writing ends up on you.
What should you get back before design starts?
A short document you can read in one sitting. The format matters less than whether each part passes a test.
- Positioning statement. One paragraph: what you make or do, for whom, and why a buyer would pick you over the alternatives they're comparing. The test: your sales team recognizes it, and a competitor couldn't use it unchanged.
- Buyer notes. The roles involved in a purchase and what each one checks before getting in touch. The test: it names real checks, like a certification, a tolerance or a lead time.
- Page list (sitemap). Every page, the job it does and the question it answers. The test: every page has a job; any page dropped was checked against its search traffic and is marked for a redirect.
- Content plan. Who writes each page, from what source, by when. The test: no page says "client to provide" without a name and a date.
- URL inventory and redirect plan. Every current address, each with a keep, rewrite, merge or remove decision, with a redirect target for anything that moves or goes. The test: it covers every URL, not just the top twenty; the final redirect map is due before build.
- Functional and technical scope. Platform or CMS, every form and where its submissions go (inbox, CRM), integrations such as product data or drawing uploads, and who holds hosting, domain and analytics access. The test: nothing in the build is first discovered during the build.
- Baseline. How quote requests reach you today and how they'll be counted after launch, plus current search traffic, rankings and top landing pages, recorded before anything changes. The test: it includes requests that go straight to an estimator's inbox.
- Open decisions. What's still unresolved, and who decides it. The test: the list is short, and each item is named and dated.
Mood boards, persona posters and competitor collages can come along as attachments, but none is a decision. Discovery should produce the full URL inventory and a decision for every address. Slugs shift during design, so the one-to-one redirect map is finished before build, not the week before launch. The website migration checklist covers that work.
Discovery should also end with the build's scope and price confirmed, or re-quoted against these decisions. A quote written before discovery was partly a guess about what's in that list.
How can you tell the discovery was real before you pay for design?
Read the output with four questions in mind.
- Is there a sentence in it you didn't say? Discovery that only reorganizes your own answers is transcription. A useful one tells you something about your buyers or your site you hadn't put into words.
- Is there a sentence you'd argue with? If every line is agreeable, nothing was chosen. A real positioning statement leaves some work out of the first line, and someone in your company will mind.
- Would the person who fields quote calls recognize the buyers in it? Give them the buyer notes. If they say customers never ask about any of that, the document isn't finished.
- Could design start on Monday? If the page list, URL decisions, content owners, technical scope and open decisions aren't settled, they'll get answered during design instead, more slowly and at a higher price.
Settle two things in writing before you sign: what the discovery output will be, ideally with a redacted example from a past project, and whether you keep it if you don't continue into design.
The real decision discovery hands you is whether to carry on with the same agency into design. It's the first piece of their work you can judge on its own. If it passes these tests, the rest of the B2B website redesign has something to be checked against. If it doesn't, you've found out before the larger invoices, and if you settled ownership up front, the document goes with you.
