The State of Small-Business Websites in Canada 2026

Does the business have a website at all?

Before speed comes existence. Of 146,074 Canadian business listings examined, four in ten give no web address. Among those that do, a third of the “websites” are either somebody else’s page or nothing at all.

A listing is the first place a customer looks a business up, and the website field on it is the door. For 60,360 of the businesses in this sample that field is empty — 41.3%. The figure is a ceiling rather than a count, because a business can own a site its listing does not mention, but it is a high ceiling, and it matches what any tradesperson’s phone will tell you about how their customers actually find them.

The more surprising finding is what sits behind the addresses that are filled in. Of the 85,714 businesses that list a website, 15,668 (18.3%) point at a page they do not own: a directory listing, a marketplace profile, a booking platform, or a social page. In Canada this is almost entirely one thing. 98% of those addresses are Facebook pages. For one Canadian small business in five, “our website” means a Facebook page, with everything that implies about ownership, discoverability and what happens when the platform changes its rules.

A further 11,139 addresses (13.0%) are dead. The domain has lapsed, the host has gone, or the address never resolved in the first place. These are businesses that once paid for a website, put its address on their listing, and are now sending every customer who clicks it to an error page. Most will not know.

One Canadian business in eight that lists a website is sending customers to a page that no longer exists.

The mobile speed picture

Google scores a page from 0 to 100 and paints it red, orange or green. The median Canadian small-business homepage scores 61, which is orange. Fewer than one in twelve reach green. Nearly one in four is red.

The performance score is a composite. Lighthouse, the engine behind PageSpeed Insights, loads the page on a simulated mid-range Android phone over a throttled mobile connection, measures when things appear and when the page responds to a tap, and blends those into one number. It is the number a business owner sees first when they test their own site, and it is the number this chapter reports.

The distribution is narrow. Half of all sites score between 51 and 71, and the two most populated bins — 50–59 and 60–69 — hold just under half the sample between them. The tails are not symmetric. The green band above 90 holds 8.1% of sites, and only 2.7% score a perfect 100. The red band below 50 holds 23.0%, nearly three times as many, with a long tail of 3.6% scoring below 30 — which on a phone means a page that is effectively unusable until the visitor has given up.

Two combined measures are worth holding onto. A site that both scores 90 or above and shows its main content within Google’s 2.5-second threshold is doing everything right; 5.8% of Canadian small-business sites do, roughly one in seventeen. At the other end, 22.9% both score under 50 and take more than four seconds to show their main content. That is the group for whom the website is, on a phone, an obstacle to the business rather than a door into it.

For every Canadian small business with a fast mobile site, three have one that Google rates as poor.

The eight-second homepage

The score is a summary. Largest Contentful Paint is the thing a visitor feels: how long until the biggest thing on the screen — usually a headline or a hero photo — has appeared. On the simulated phone Google uses, the median Canadian small-business homepage takes 8.2 seconds. A quarter take longer than 13.7.

Google’s own guidance is that a good Largest Contentful Paint is 2.5 seconds or less, that 2.5 to 4 seconds needs improvement, and that anything over 4 seconds is poor. Against those bands, Canadian small-business sites are not spread across the spectrum. They are stacked at the slow end. 6.4% are good. 11.0% need improvement. 82.6% are poor, and two-thirds of all sites — 66.1% — take more than six seconds, which is the point at which a visitor on a bus has already switched to the next result.

The shape of the curve is the finding. It rises slowly through the first four seconds, where Google wants every site to be, then climbs steadily for another sixteen. There is no cliff and no cluster of good sites: the population simply drifts from slow to slower, with a tenth of it past twenty seconds.

Two-thirds of Canadian small-business homepages take longer than six seconds to show their main content on a phone. Google’s threshold is two and a half.

What is actually wrong with them

A slow page is a symptom. Lighthouse also records why, as a list of failed checks, and across 57,897 homepages the list is monotonous: the same handful of problems appear on nearly half of all sites. None of them requires a redesign to fix.

The audit records up to ten failed checks per homepage, the ones Lighthouse weighs as most significant. 99% of sites carry at least five. The top of the frequency table is dominated by loading mechanics rather than content: resources that load one after another instead of in parallel, main content that paints late, pages that look ready before they respond, and servers that tell the browser to download everything again on every visit.

Three things stand out. First, images: 40.3% of sites deliver images larger than the screen they are shown on or in formats a decade old, and 23.2% do not tell the browser how big an image will be, so the page jumps around while it loads. Second, caching: 42.3% make a returning customer download the whole page again, which is a one-line server setting. Third, code the page does not use: 42.0% ship JavaScript that never runs and 37.2% ship styling for elements that are not there — almost always the residue of a theme or a page builder that includes everything in case it is needed.

What is not on the list is telling too. Heavy pages — the “enormous network payload” check — appear on 19.0% of sites. The problem is rarely that a site has too much on it. The problem is the order and manner in which an ordinary amount is delivered, and that is exactly the class of problem whoever built the site can fix in an afternoon.

The five most common problems on Canadian small-business sites are all fixable by whoever built the site, without touching a word or a picture.

The basics of being found

This is the chapter that runs against expectation. Lighthouse also grades a page on the basics of search-engine readiness, and Canadian small businesses mostly pass. The median SEO score is 92; three-quarters score 90 or above, and 37% score a perfect 100. The failures that remain are the embarrassing ones.

These checks are not about ranking. They test whether a page is crawlable and describes itself: a title, a meta description, readable link text, alt text on images, a valid robots file, a mobile viewport, HTTPS. These are things a page builder does by default in 2026, and the scores reflect it — only 8.6% of sites score below 80.

Two of the failures deserve a sentence each. No HTTPS on 6.5% of sites means that in 2026, roughly 3,800 Canadian small businesses in this sample alone run a site every browser marks “Not secure” in the address bar — the first thing a customer sees, before they see the business. And no mobile viewport on 5.1% means the page is not built for a phone at all: it renders as a desktop page shrunk to fit, with text too small to read. For those sites the speed numbers in the earlier chapters are almost beside the point.

Taken together, the search basics are the good news in this report, and the report says so plainly, because it is what makes the speed numbers credible. The Canadian small business has, by and large, done what it was told to do about being found. What it has not been told is that being found is worth little if the page then takes eight seconds to arrive.

In 2026, one Canadian small-business site in fifteen still has no HTTPS, and one in twenty is not built for a phone at all.

Province by province

The provincial story is that there is no provincial story. Every province’s median score sits between 59 and 62, and the share of poor sites runs from a fifth to a quarter. Website quality does not follow the map; it follows who built the site.

Quebec (20.8% poor), New Brunswick (19.5%) and Newfoundland and Labrador (19.9%) have the smallest shares of poor sites. Alberta (24.6%) and Prince Edward Island (26.3%) have the largest, though the Island’s 209 sites make its figure the least certain in the table. Prince Edward Island is also, oddly, the only province where a tenth of sites meet the main-content threshold; the other nine sit between 4.3% and 6.9%. The Prairie provinces have the fewest fast sites: Saskatchewan 4.3%, Newfoundland 4.9%, Manitoba 5.1%.

Quebec’s marginal lead deserves a sentence, because it will be quoted. Quebec’s sites are, on the median, one point better than Ontario’s and three better than Alberta’s, and have the second-smallest poor share in the country. The likely reason is not that Quebec’s developers are better, but that a larger share of Quebec’s small-business sites were built recently and in French, by a smaller set of local agencies and platforms, which narrows the variance. Whatever the cause, the lead is small: three-quarters of Quebec’s sites still fail Google’s speed test.

Quebec’s small-business websites are marginally the fastest in the country, and still three-quarters of them fail Google’s speed test.

City by city

Nineteen cities have at least 500 audited sites. Oshawa, Hamilton and Victoria have the smallest share of poor sites; Vancouver and Toronto the largest. The two cities that house the country’s biggest web-design industries have its slowest small-business websites.

That pattern is real and it is probably not about the agencies. Toronto and Vancouver have the country’s most expensive commercial rents and its most competitive local markets, and a small business under that pressure is the one most likely to have bought a template site five years ago and never touched it again. The mid-sized cities — Oshawa, Hamilton, Victoria, London, Halifax — sit at the other end with poor shares around a fifth and the country’s highest shares of good sites: Oshawa’s 11.8% green is the best figure in the table.

The city table counts Montréal proper at 1,622 sites, because many businesses on the island list a borough or a demerged municipality as their city. Grouping those gives a fairer read. The Island of Montréal, 4,244 sites, has a median of 61, a poor share of 22.4% and a good share of 8.9%. The off-island suburbs, 2,950 sites from Laval to Vaudreuil-Dorion, do slightly better on every measure: median 62, 20.3% poor, 10.5% good. Québec City and Lévis, 1,096 sites, have the lowest poor share of any large Canadian market at 19.7%. Ottawa–Gatineau, 2,696 sites across the river, sits at 23.3% poor and has the region’s smallest share of fast main-content times, 5.2%.

Toronto and Vancouver, home to the country’s largest web-design industries, have the highest share of slow small-business sites among Canada’s big cities.

Industry by industry

Industry is where the sample finally spreads out. Electricians, accountants and plumbers have the fastest sites in Canada; real estate, insurance, mortgage and fitness the slowest. The industries that sell with photographs pay for it in load time.

Forty-seven industries have at least 300 audited sites. Their median scores run from 65 down to 54, and their poor shares from 12.8% to 39.5% — a spread three times wider than anything the provinces or cities produced. The fast end is populated by the trades and the professions: Electrical (median 65, 12.8% poor), Accounting (64, 15.0%), Plumbing, Renovation and Painting (all 64). These are businesses whose sites are built to do one thing — produce a phone call — and tend to be light: a headline, a service list, a number.

The slow end is populated by industries that sell with images and with third-party widgets. Real Estate has the worst figures in the report: a median of 54, 39.5% poor, and a median main-content time of 10.6 seconds, the product of listing feeds, map embeds and full-width photography. Insurance (57, 29.7% poor) and Mortgage (58, 28.5%) carry quote calculators and lead forms. Fitness (58, 30.7%), Med Spa (57), Bridal (58) and Restaurant (59) sell atmosphere, and atmosphere is heavy: restaurants and photographers both post median main-content times near ten seconds.

The middle of the table is where the large health professions sit, and they deserve attention because of how many of them there are. Dental (2,720 sites), Veterinary (1,546), Physiotherapy (1,438) and Optometry (1,416) all post a median of 60, a poor share between 24% and 28%, and the smallest shares of fast sites in the report — between 4.6% and 7.5% in the green band. A dental practice’s website is nearly always built by one of a handful of specialist vendors on a shared template, and the numbers say those templates are uniformly middling.

A Canadian plumber’s website is, on average, faster than a Canadian real-estate agent’s. The industries that sell with photographs pay for it in load time.

Can a customer reach them?

Speed is only worth measuring on a page that leads somewhere. The audit also recorded what each homepage offers a visitor who wants to make contact: a phone number nearly always, an email address two times in three, and a link to a social profile just over half the time.

The phone number is the one thing the Canadian small business gets almost universally right: 91.8% of homepages carry one. An email address is on 63.1%, and where one is published, 53.3% are a generic inbox — info@, contact@, hello@, bonjour@ — rather than a person. That is not wrong, but it is the address most likely to be checked weekly rather than hourly, and the one a customer is least likely to write to twice.

The social figures describe where the Canadian small business actually lives online. Facebook is linked from 48.9% of sites and Instagram from 40.5%; LinkedIn, at 16.2%, is a distant third and mostly confined to the professions. YouTube (12.5%) and TikTok (4.6%) barely register. And 42.4% of homepages link to no profile at all — which, set beside chapter 1’s finding that a fifth of businesses use a Facebook page as their website, describes two populations that barely overlap: businesses that are on social media instead of the web, and businesses that are on the web and nowhere else.

Four Canadian small-business websites in ten link to no social profile, and one in three publishes no email address.

Canada and the United States

The same audit ran on 76,748 American small-business homepages in the same weeks. Canadian sites are slightly better on score and slightly worse on main-content time; the striking difference is in what a listed “website” turns out to be.

On the numbers that describe a real site, the two countries are close. The Canadian median score is 61 to the American 59. The United States has more sites in both tails — more poor (27.4% against 23.0%) and more good (10.0% against 8.1%) — and more sites meeting the main-content threshold (8.2% against 6.4%). Canada clusters in the middle. Median SEO scores are identical at 92.

The listings differ far more than the sites. A third of American listed websites (34.4%) point at a directory or platform page, against a fifth in Canada, while Canadian listed websites are more than twice as likely to be dead (13.0% against 5.5%). Part of this is source mix: the American sample drew more heavily on search-result sweeps, which by construction find live sites, and the Canadian sample on directory listings, which preserve addresses long after the sites behind them have gone. The report notes the gap and does not read a cultural difference into it.

The largest American states look like the Canadian provinces: a narrow band. Texas (12,119 sites), California (10,176), Florida (8,235) and Pennsylvania (4,515) all post a median of 60 and poor shares between 25.0% and 26.8%. New York (59, 28.1%), Illinois (58, 29.3%) and Colorado (58, 30.0%) sit a little below. Nothing in either country’s geography moves the needle the way industry does.

The United States has more fast small-business sites and more slow ones; Canada clusters in the middle.

What a business owner should do with this

Most of what this report measures can be fixed by whoever built the site, in a working day, without changing a word or a picture on it. This is the short version of what to ask for.

First, run the test yourself

Google’s PageSpeed Insights is free and needs nothing but a web address. Test the homepage on mobile. If the performance score is above 90 and the Largest Contentful Paint is under 2.5 seconds, the site is in the top 6% of this report and this chapter does not apply. If the score is under 50, the site is in the bottom quarter and every item below almost certainly does.

Then ask for these five things, in this order

1. Images. Every image resized to the largest size it is ever shown at, saved in a modern format (WebP or AVIF), with its width and height declared in the page. This addresses the two image checks that 40% and 23% of sites fail, and is usually the single largest gain.
2. Caching. The server told to let browsers keep files for a year. It is one setting, it costs nothing, and 42% of sites have not done it.
3. Fonts. Web fonts set to show text immediately in a fallback face while the real one loads, and the number of fonts cut to two. A quarter of sites leave text invisible while a font arrives.
4. Unused code. The theme or page builder’s optional modules switched off, and any plugin the site no longer uses removed. Two-fifths of sites ship code that never runs.
5. Redirects. The address on the listing, the address on the business card and the address the site answers on made the same, with one redirect at most. A fifth of sites bounce a visitor two or three times before the page begins.

And the three questions to ask whoever built it

• Is the site on HTTPS, and does the http:// address redirect to it?
• Does it have a mobile viewport, or is it a desktop page shrunk to fit?
• What does the homepage score on PageSpeed Insights today, and what will it score when you are done?

If the answer to the third question is a shrug, the site is in the wrong hands.

All 47 industries

Industries with at least 300 audited Canadian sites, sorted by median mobile performance score.

Methodology and data

Population

146,074 Canadian business records assembled by FutureSource between July and August 2026 from public listings: 116,916 from Yellow Pages Canada, 26,796 from Google search results by industry and city, 2,346 from Google Maps, and a small remainder. It is a large convenience sample of small and mid-sized service businesses across roughly eighty industries, not a census, and it over-represents the industries FutureSource works in: trades, clinics, professional firms and local retail. Cite it as “Canadian small businesses drawn from public listings,” never as “all Canadian businesses.”

Website status

85,714 records (58.7%) listed a website. Each address was classified before auditing: own site (56,764), directory or platform page (15,668 — any address on a host the business does not control, using a fixed list of directory, marketplace, social and booking hosts; 98.0% of the Canadian ones were facebook.com), dead (11,139 — no response, DNS failure, or a parked domain), and 2,143 not classified or timed out. Booking-platform storefronts (Vagaro, Booksy, Fresha, Mindbody) are deliberately not treated as directory pages, because the business builds that page itself.

Audit

57,897 own-site homepages measured with Google PageSpeed Insights (Lighthouse) between 8 and 30 August 2026, mobile profile, one run each, from Canadian infrastructure. Recorded per site: performance score, SEO score, Largest Contentful Paint, and up to ten failed audits. Desktop was not run. The 76,748 United States sites in chapter 10 were audited the same way in the same weeks.

Thresholds

Google’s own. Performance and SEO scores: 0–49 poor, 50–89 needs improvement, 90–100 good. LCP: 2.5 s or under good, 2.5–4 s needs improvement, over 4 s poor.

Cuts

Province and city come from the listing address after normalisation to ISO province codes; both spellings of Montréal are merged. Industry is FutureSource’s canonical bucket, assigned from the listing category. Tables show only groups with at least 200 (province), 500 (city) or 300 (industry) audited sites.

Reachability

Phone, email and social links were read from each business’s own homepage by a crawler in August 2026. A profile that exists but is not linked from the homepage counts as absent; a generic inbox is any address whose local part is info, contact, hello, admin, office, sales, support, reception or bonjour.

Privacy

All figures are aggregates. No business is named, no address is printed, and no cell with fewer than 200 sites is published. Individual audit results are available only to the business they concern.

Limits

• Lab, not field. LCP and the performance score are measured on a simulated mid-range phone over a throttled connection, the standard Lighthouse mobile profile. Real-user timings are typically faster. Comparisons between groups are fair; the absolute seconds are a benchmark.
• One homepage, one run. Lighthouse scores vary by a few points between runs. Medians over thousands of sites are stable; an individual site’s figure is not, which is one reason none are published.
• “No website” means no website in the listing. The 41.3% figure is a ceiling on absence, not a count of it.
• The failed-audit list is capped at ten per site, so the frequency table describes the most significant failures, not every failure.
• Source mix differs between countries. The American sample drew more on search results and less on directories than the Canadian one, which likely explains most of the directory and dead-site gap in chapter 10.
• Not measured: desktop performance, search rankings, advertising, reviews, inner pages, and accessibility beyond what the SEO checks cover. Each would be a separate study.

The Greater Montréal grouping in chapter 7 uses the following city names as they appear in listings. Island: Montréal, Montreal, Lachine, Pointe-aux-Trembles, LaSalle, Pierrefonds, Saint-Léonard, Rivière-des-Prairies, Verdun, Saint-Laurent, Montréal-Nord, Dollard-des-Ormeaux, Pointe-Claire, Westmount, Dorval, Kirkland, Mont-Royal, Anjou, Outremont, Côte-Saint-Luc, Hampstead, Beaconsfield, Sainte-Anne-de-Bellevue, Baie-D’Urfé, Senneville, Montréal-Est, Montréal-Ouest. Off-island: Laval, Longueuil, Brossard, Terrebonne, Repentigny, Châteauguay, Blainville, Vaudreuil-Dorion, Saint-Jérôme, Mirabel, Boisbriand, Saint-Eustache, Saint-Hubert, Sainte-Julie, Boucherville, Saint-Lambert, Candiac, La Prairie, Mascouche, Sainte-Thérèse, Rosemère, Deux-Montagnes, Saint-Constant, Delson, Varennes, Chambly, Beloeil, Saint-Bruno-de-Montarville, Pincourt, L’Île-Perrot, Notre-Dame-de-l’Île-Perrot, Vaudreuil, Hudson, Saint-Lazare.

How to cite

FutureSource (2026). The State of Small-Business Websites in Canada, 2026. First edition, September 2026. Montréal: FutureSource.

Figures and charts may be reproduced with attribution and a link to this page.