The website discovery process.

Why discovery exists (and what happens without it)

Every website decision is a trade-off. Should the homepage lead with services or with proof? Should the booking system be custom or off-the-shelf? Do you need twelve pages or forty? Without discovery, those trade-offs get resolved by whoever has the strongest opinion in the room, on the day. With discovery, they're resolved by evidence.

The economics are brutal and well-established in software: a decision changed during research costs almost nothing; changed during design it costs hours; changed during build it costs days; changed after launch it costs a rebuild. Discovery is the cheapest place to be wrong. That's the entire argument for it.

There's a second, less obvious benefit. Discovery surfaces disagreement inside your own business while it's still free — the owner who wants to look premium and the sales manager who wants to compete on price, discovering mid-build that they've been describing two different companies. Better to have that conversation in week one.

Phase 1 — Business goals and KPIs

Everything downstream depends on this. The aim is to convert vague ambitions into a small set of measurable outcomes, ranked.

Getting to real goals

Start with the commercial reality: where does revenue come from now, where do you want it to come from, and what's blocking that? A plumbing firm doing 90% domestic work that wants commercial contracts doesn't need "a better website" — it needs credibility signals, case studies and content that ranks for commercial search terms. Same trade, entirely different site.

Useful questions: What's the single most valuable action a visitor can take? What's your average job value, and which jobs do you want more of? What do you currently turn down? What do customers ask before they commit? Which competitors do you lose to, and what do they have that you don't?

Turning goals into KPIs

A KPI needs a baseline, a target and a measurement method. "More enquiries" isn't a KPI. "Enquiry form submissions from 6/month to 20/month within six months, measured in analytics" is.

Note the last column. If a KPI can't be measured with tools you actually have, either set up the measurement during the project or pick a different KPI. Unmeasurable goals are how projects avoid accountability.

Phase 2 — Audience research, personas and journeys

Research that doesn't need a budget

Big agencies run interviews and surveys. A small business can get 80% of the value from four free sources: your own sales calls and emails (what do people actually ask?), your Google Business Profile questions and reviews, your competitors' reviews (what do their customers complain about?), and Search Console (what are people searching when they land on you?). An hour with those four beats a persona workshop.

Personas, used properly

Personas get mocked because they're often invented — "Marketing Mary, 34, likes yoga". A useful persona contains only decision-relevant information: what triggers their search, what they're worried about, what they need to see to trust you, what would make them leave, and how they prefer to make contact. Two or three is plenty; most small businesses genuinely serve two audiences and pretend they serve five.

User journeys

Map how each persona gets from trigger to conversion. For a local trade it's often: something breaks → Google search on a phone → scan three results → check credibility → phone whoever answers fastest. That journey tells you the phone number must be tappable and instant, the page must load in under two seconds on 4G, and credentials must be visible without scrolling. You've just made three design decisions from evidence rather than taste.

For a considered B2B purchase the journey is longer — research, shortlist, internal approval, procurement — and the site needs case studies, downloadable detail, and pages that survive being forwarded to a finance director who's never heard of you.

Phase 3 — Competitor and market research

The goal isn't to copy competitors; it's to know precisely what you're being compared against, because your customer will compare you.

For each of your top five competitors, record: what they lead with, their price positioning (if visible), their proof (reviews, case studies, accreditations), how fast their site is, how they handle mobile, and what they rank for. Run each through a website health checker — a competitor scoring 41 out of 100 is telling you the technical bar in your market is low, which is a genuine opportunity.

Then find the gaps. In our audit of 421 Essex business websites, 79% had at least one technical fault, 28.5% had no structured data at all and one in seven was broken outright. In most local markets, simply being fast, current and technically sound puts you ahead of the majority — before anyone writes a word of content.

Phase 4 — The three audits (content, SEO, technical)

If you have an existing site, audit it before you replace it. You're carrying value you can't see, and destroying it is the most common expensive mistake in web projects.

Content audit

List every page. For each, record: traffic, conversions, rankings, last updated, and a verdict — keep, rewrite, merge, or delete. Most small business sites have a handful of pages doing all the work and a long tail of pages doing nothing. Deleting dead weight often improves rankings, because it concentrates authority and gives search engines less thin content to wade through.

SEO audit

Export Search Console: which queries bring impressions, which bring clicks, which pages rank, and where you sit on average position. Any page ranking in positions 5–15 is a near-miss worth protecting and improving. This export is also your migration safety list — those URLs must survive, or redirect cleanly.

Technical audit

Speed, Core Web Vitals, mobile rendering, HTTPS, crawlability, indexing status, structured data, broken links, and whether anything is accidentally blocking search engines. Google's own Search Console and PageSpeed Insights cover most of it free; our checker gives you a plain-English summary in twenty seconds. If the site isn't appearing at all, our guide on why a website isn't showing on Google works through the nine causes in order.

Phase 5 — Information architecture

Information architecture is how content is organised and labelled: the sitemap, the navigation, the URL structure, and how pages relate to each other. It's the skeleton, and it's very expensive to change after launch because every URL you move needs a redirect and every ranking you've earned is attached to a specific address.

Principles that hold up: label things the way customers say them, not the way your industry does (people search "boiler repair", not "heating system remediation"); one page per intent, so each page has a clear job and doesn't compete with its siblings; depth costs clicks, so keep important pages within two clicks of the homepage; and URLs should be readable and stable, because they're seen by customers and search engines alike.

For most small businesses the shape is straightforward — the baseline is covered in what pages a website should have. Where it gets interesting is service and location structure: separate pages per service and per town, each genuinely differentiated, beat one page listing everything. That's not a trick; it's the only way a page can be specifically relevant to a specific search.

Phase 6 — Features, platform, hosting and integrations

Feature planning

Every feature has a build cost, a maintenance cost and a failure mode. Sort them ruthlessly into must have (the site fails without it), should have (real value, could follow in phase two), and could have (nice ideas). Most projects that overrun do so because a "could have" was quietly treated as a "must have" by someone.

Platform selection

Choose the platform from the requirements, not the other way round. The honest decision matrix:

Hosting requirements

Discovery should establish: expected traffic, whether server location matters (for UK businesses, UK or nearby hosting reduces latency), backup and uptime expectations, and who holds the account. That last point matters more than the spec — hosting held in a supplier's name is a hostage situation waiting to happen.

Integrations

List every system the site must talk to: accounting, CRM, email marketing, booking, stock, payment providers, review platforms. For each, confirm early that an API or supported integration actually exists — "we'll connect it to our CRM" is a sentence that has derailed a great many projects at the eleventh hour.

Phase 7 — What 2026 discovery adds: AI search, entities and E-E-A-T

This is where most discovery templates you'll find online are five years out of date. Three things now belong in every discovery, and they're cheap to plan and expensive to retrofit.

AI search readiness

A growing share of buying journeys now start with an AI assistant or an AI Overview rather than a list of links. Being citable by those systems depends on decisions made during discovery: does each page answer a specific question directly and early? Is your business information consistent and machine-readable? Is structured data planned for services, FAQs, reviews and location? Is the site crawlable by AI agents as well as search engines? Plan the answers now, because bolting them on later means rewriting pages.

Entity mapping

Search engines and language models understand the world as entities and relationships, not keywords. Your discovery should establish what your business is in those terms: name, location, services, service areas, credentials, people, and the relationships between them — then make those consistent everywhere (site, Google Business Profile, directories, structured data). Consistency is what turns scattered mentions into a recognised entity.

E-E-A-T planning

Experience, Expertise, Authoritativeness and Trust are the qualities Google's raters assess, and they need evidence on the page: named authors with real credentials, genuine case studies, accreditations, transparent pricing, real photography, reviews, clear contact details. Decide during discovery who the named expert is and what proof exists — because it changes the content plan, and often reveals that the most valuable asset the business has is the owner's own experience, currently invisible.

Content clustering

Plan content as clusters, not as a list of pages: a pillar page covering a topic broadly, supported by focused pages answering specific questions, all interlinked. That structure is how a small site builds topical authority against bigger competitors — and it's precisely the structure of the guides section you're reading now.

Local SEO planning

For any business serving a defined area, plan this explicitly: which towns you'll target and with what pages, how Google Business Profile will be maintained, how reviews will be generated systematically, and where your business details will be listed consistently. Our guide on what local SEO actually involves covers the mechanics.

Phase 8 — Conversion planning and future scalability

CRO planning before launch, not after

Conversion optimisation usually gets bolted on months later. Planned during discovery it's nearly free: decide now what the primary action is on each page type, where trust signals sit, how forms are structured (every extra field costs completions), what happens after someone converts, and what you'll measure. Design a site to convert and you rarely need to rescue it later.

Scalability

Ask what the business looks like in three years. A second location? A trade counter? Ten more staff with profile pages? Selling online? You don't build all of it now — but knowing it's coming changes the information architecture and platform choice, and stops you needing a rebuild in eighteen months. The cost of leaving room is usually nil; the cost of not leaving room is the whole project again.

Phase 9 — The deliverables and the launch roadmap

Discovery should produce artefacts, not just conversations. For a small business project, that's typically:

• A one-page summary of goals and KPIs with baselines and targets
• Two or three persona/journey notes
• A competitor summary with health scores
• Audit findings: content keep/rewrite/merge/delete, SEO must-preserve list, technical fault list
• A sitemap and URL plan, including redirect mapping if replacing a site
• A feature list sorted must/should/could
• Platform, hosting and integration decisions with reasons
• A content plan naming who writes what, by when
• The measurement plan — what's tracked and where
• A phased roadmap: launch scope, then phase two

The roadmap matters. Trying to launch everything at once is how projects slip for months; launching a strong core and adding phase two after real usage data is how they succeed. And "phase two" should be written down, not vaguely promised, or it never happens.

Running discovery yourself, without an agency budget

If you're not paying for a formal discovery phase, you can still do the valuable 80% in a couple of evenings. Work through this order: write your goals and KPIs with baselines; spend an hour reading your own enquiries and your competitors' reviews; run five competitor sites through a health checker; export your Search Console queries; list every page on your current site with a keep/rewrite/delete verdict; sketch your sitemap; list must/should/could features; decide who writes the content. Then put the answers into a website brief — our free brief builder is structured around exactly these outputs.

Do that, and you'll arrive at your first supplier conversation better prepared than most businesses ever are — which shows up directly in the quality of the quotes you get back, and in the price.

Frequently asked questions

What is the website discovery process?

The structured research phase before design begins. It establishes business goals and measurable KPIs, researches audience and competitors, audits existing content, SEO and technical health, then defines information architecture, features, platform, hosting and integrations — so every later decision has evidence behind it rather than opinion.

How long should website discovery take?

Proportionate to the project: an afternoon to three days for a simple local business site, one to three weeks for something substantial with integrations or e-commerce, longer for enterprise work. If a supplier proposes weeks of discovery for a five-page site, that's padding.

Is discovery worth paying for on a small website?

The questions are always worth answering; whether you pay someone to facilitate depends on budget. A small business can do most of it themselves in two evenings using their own enquiries, competitor reviews, Search Console data and a health checker — then capture the answers in a brief.

What's the difference between discovery and a website brief?

A brief is what you write before appointing anyone — it describes the problem so suppliers can quote. Discovery is the deeper research done at the start of the project, usually with your appointed designer, that turns those goals into evidence-based decisions about structure, features and content.

What are KPIs for a website?

Measurable indicators tied to business outcomes: enquiry form submissions, phone calls from the site, bookings, quote requests, average order value, rankings for named searches, or quote acceptance rate. Each needs a baseline, a target and a stated measurement method, or it isn't a KPI.

Do I need user personas for a small business website?

Two or three brief ones help, provided they contain only decision-relevant information: what triggers the search, what worries them, what they need to see to trust you, and how they prefer to make contact. Invented demographic detail adds nothing.

What is a content audit?

Listing every page on your existing site with its traffic, conversions, rankings and last update, then assigning a verdict: keep, rewrite, merge or delete. Removing thin, dead pages often improves rankings by concentrating authority.

What is information architecture?

How content is organised and labelled — sitemap, navigation, URL structure and the relationships between pages. It's expensive to change after launch because every moved URL needs a redirect and rankings attach to specific addresses, so it belongs in discovery.

How do I choose a CMS during discovery?

From requirements, not preference. Frequent self-editing and blogging favour WordPress built properly; a fast marketing site with minimal upkeep favours hand-coded; stock-heavy shops favour WooCommerce or a hosted platform; complex custom logic favours a custom build.

What is AI search readiness?

Planning a site so AI assistants and AI Overviews can find, understand and cite it: direct answers early on each page, consistent machine-readable business information, structured data for services, FAQs, reviews and location, and crawlability for AI agents. Cheap to plan, expensive to retrofit.

What is entity mapping in SEO?

Defining what your business is in the terms search engines and language models use — name, location, services, service areas, credentials, people and their relationships — then keeping that consistent across your site, Google Business Profile, directories and structured data.

What does E-E-A-T mean for a small business website?

Experience, Expertise, Authoritativeness and Trust need visible evidence: named authors with real credentials, genuine case studies, accreditations, transparent pricing, real photography, reviews and clear contact details. Discovery should identify who the named expert is and what proof already exists.

Should conversion optimisation be part of discovery?

Yes — planned upfront it's almost free. Decide the primary action for each page type, where trust signals sit, how forms are structured, what happens after conversion, and what you'll measure. Retrofitting conversion work later costs far more.

What deliverables should discovery produce?

Goals and KPIs with baselines, persona and journey notes, a competitor summary, content/SEO/technical audit findings, a sitemap and URL plan with redirects, a must/should/could feature list, platform and hosting decisions, a content plan with owners and dates, a measurement plan, and a phased roadmap.

What happens if you skip discovery?

The decisions still get made — just later, by whoever has the strongest opinion, and usually during build or after launch when changes cost days rather than minutes. Skipping discovery doesn't remove the cost; it relocates it to the most expensive part of the project.

Who should be involved in discovery?

Whoever holds commercial knowledge and whoever has final sign-off — often the same person in a small business. Include anyone who handles enquiries, because they know what customers actually ask. Name one decision-maker to prevent endless revision loops.

Related: How to write a website brief · Brief builder · What pages you need · Health checker · 2026 audit data