Design intelligence systems: a site that reads its own market

Most business sites are drawn once and then left alone. A design intelligence system is built the other way: the pages, the search structure and the enquiry path are decided from evidence about how the trade and the town actually search, and every paragraph on the page has a job you can name.

Scope mechanics

  • A written scope before any work starts: which pages exist, what each one must do, and what it costs, agreed in writing rather than implied over the phone.
  • Search evidence first: the exact phrases buyers in this trade and this town type, collected before a single layout is drawn, so the structure follows the demand instead of guessing at it.
  • A booking, quote or call path wired into the pages from day one, tested end to end, never bolted on after launch by a plugin nobody owns.
  • Local search structure from the first commit: titles, headings and service descriptions written for the neighbourhoods the business actually serves.

How the funnel works

  1. The search. A buyer types the trade and the town, or the problem in plain words. The page they land on answers that exact phrase in its heading, because the heading was chosen from the search, not from a template.
  2. The read. The visitor scans for the one thing that decides the call: price clarity, proof of the trade, or a reason to trust this business over the directory listing above it. The page puts that thing above the fold.
  3. The act. One clear next step per page: book, request a quote, or phone. Not three competing buttons, not a newsletter popup, not a form that asks for information the business never uses.
  4. The handover. The enquiry lands in accounts the owner controls, with the site, domain and analytics in the owner's name. There is no platform holding the business hostage, and nothing to rent back.

Engine parts

  • Search-led page architecture: the sitemap is drawn from what people search, so a trade with five ways of being found gets five pages that each answer one of them.
  • Heading and metadata discipline: every title and description written for a human reader and a search index at once, never copied from the last page with the noun swapped.
  • Mobile performance as a design material: pages that load on real cellular data in the town being served, because that is where the search happens.
  • Answer-ready content: direct answers to the questions a buyer asks before calling, written so a search assistant can quote them without distortion.

Artefacts you keep

  • The search map: the phrases, towns and trade words the site is built to be found for, written down before the design.
  • The page job list: one sentence per page naming what the page must make the visitor do, reviewed against the scope.
  • The speed record: load measured on a real phone connection, not a lab score, kept with the build paperwork.
  • The ownership inventory: site, domain, hosting and analytics accounts listed in the client's name at handover.

The field guide: Design intelligence

The problem

Most trade sites lose the search before the visitor ever sees the site. The homepage talks about the business: its history, its values, its gallery of finished work. Meanwhile the person doing the searching carries a specific job, a burst pipe in Kelowna, a cracked tooth in the evening, a roof leaking after a storm. That person types the problem, not the business name, and the first page that answers the problem earns the call. A site drawn from a template can look settled and still miss every one of those phrases, because nobody wrote the phrases down before drawing. The result is familiar: a handsome site, a quiet phone, and an owner blaming the internet when the real failure happened earlier, at the moment the pages were decided.

How it works

Design intelligence treats the search map as the architecture. Before any layout is drawn, the studio writes down the exact phrases buyers in this trade and this town type: the trade plus the town, the trade plus the problem, and the questions asked before a call. Each phrase becomes a test for the sitemap. If the map justifies a page, the page exists. If it does not justify a page, the page does not exist. Headings then use the buyers words rather than the studio adjectives, the metadata says what the page does, and the booking or call path sits where the eye lands first. Nothing on the page is decorative. A paragraph that does not help the visitor decide is cut, and a page without a written job is not built.

Worked example — Plumbing in Kelowna

A Kelowna plumber and two searches that must not share a page

Take a Kelowna plumber as the illustration, not as a client. The map separates two searches that a template would merge into one homepage: a flooded basement at midnight, and a slow drain on a weekday morning. The first buyer wants a phone number before scrolling, the service area in the heading, and a promise about response. The second buyer wants to know what the work involves, whether it is scheduled, and which neighbourhoods the van covers. So the system gives each search its own page. The night search puts the call path at the top and names the town in the title. The weekday page qualifies the job in prose, what the camera finds, what clears the line, what happens after the form is sent. Neither page exists because a theme offered it. Both exist because the map said a buyer was looking.

Checklist

  1. Write the search map before the first layout, with trade-plus-town and problem phrases in the buyers own words.
  2. Give every page one written job: the booking, quote or call it must earn, stated in a single sentence.
  3. Put the direct answer in the first screen, with the search phrase in the heading rather than buried in copy.
  4. Measure load on a real phone on cellular data in the town served, and keep that record with the build paperwork.
  5. Hand over the accounts in the owners name, site, domain, hosting and analytics, with the inventory written down.

Have the map drawn, then watch the pages fall into place.

Ask for your search map

Direct answers

What makes this a design intelligence system rather than a website?
The difference is what decides the pages. A website is drawn to look right; this is drawn from search evidence, scoped in writing, and held to a job list — each page exists to make a visitor book, request a quote, or phone. The pages are the visible layer of that decision-making.
Where does the search evidence come from?
From the phrases buyers already type: trade plus town, the trade plus the problem, and the questions asked before a call. The map is written before the design starts, and the pages are built to answer it. Nothing is invented — if the evidence is thin, the page says less, not more.
Do I own the design intelligence behind the site?
Yes. The search map, the page job list and the site itself are yours once the invoice is paid. If you take the system to another developer, the evidence and the paperwork go with it — there is no monthly rent and nothing held back.

What arrives for a business

Ten things, every build. The list is the ceiling, not a marketing sentence — if it is not on this list it is not in the price.

  • A written scope and one fixed price

    Agreed before any work starts. The number you sign is the number you pay.

  • The words, written for the customer

    Every page answers what a buyer asks, not what the owner wants to say.

  • The design, from a system already live

    Not a template you fill in. A layout built from the same system as every other build, adapted to the trade.

  • The build, fast on a real phone

    Tested on cellular data, not on the studio's wi-fi.

  • The enquiry path

    Form, notification, and a record of what came in — wired in, not bolted on.

  • Local search structure

    Real headings, structured data, a sitemap, and the town named in the words.

  • The launch

    Domain pointed, certificate checked from outside, redirects in place.

  • One revision round, in writing

    A list of changes, agreed, done, and closed — not an open-ended tab.

  • The handover

    Code, design, content and logins, in the business's name. Nothing held back.

  • Changes after launch, through the portal

    Queued and done when we are in there. No retainer required to ask.

The full list, tier by tier →

What it costs, line by line

These are the published prices, in Canadian dollars before tax, from the same table the plan page charges from. A build is quoted as one fixed number against a written scope — this is what the lines add up to, not an offer.

Published line prices for a Nimos build.
The site$2,500
The site plan$450
A funnel$450
Local search setup$350
Care$150
Site Watch$49
Price your own build →No card, no call, no email required to see a number.

What you can check before you call

There are no testimonials on this page, because there are none to show. What there is instead is four things you can open in another tab and judge for yourself.

LATELY, ON THIS SITE

  • 2026-11-06The field, named — Sixty-six tools and people get a page each, and every city gets a hub.
  • 2026-11-04The local engine — Twenty trades across a hundred and eighteen cities, from one data file.
  • 2026-10-20The build table — The six numbered steps become one table of real handoffs — HELLO to LIVE.

Nothing on this page is a claim we cannot point at. The four links above are the evidence, and every one of them is a click away.

Read next

Four short essays on the parts of a business’s site that decide whether it works. No jargon, no gated PDF.

Two ways to start, and neither is a sales call.

Price your system line by line on the plan page, or run the free leak report on the site you have now and read the findings yourself. Neither one is a sales call.

Design intelligence systems | Nimos