UrbanTechFlow · Gatineau–Ottawa

What a Real Website Audit Checks (and What Most "Free Audits" Skip)

A heating and cooling contractor working out of a shop near Montée Paiement, in the Gatineau sector, gets an email at 6:40 on a Tuesday in October. Subject line: Free website audit, your site scored 42/100. The attached PDF has red bars, a headline that says "12 critical errors", and bullet points about missing alt text on 38 images and a missing H1 on four pages. At the bottom, a calendar link.

He reads it standing in a client's basement between two furnace calls, understands maybe a third of it, and files it. Nothing in that document told him why the phone rang eleven times last October and only six times this one. (This contractor is illustrative, a composite of the kind of business our team works with across the Outaouais, not a real company.)

Here is the uncomfortable part: the PDF wasn't wrong. It was answering a question nobody had asked.

What a "free audit" actually measures

An automated audit is a crawler. It fetches your pages, reads the HTML, and compares what it finds against a fixed checklist of machine-readable properties: image attributes, heading order, meta description length, broken links, a page-speed score from a lab test. It runs in about forty seconds and costs the sender nothing. That's not fraud. It's genuinely useful data, and if a scan says half your pages return errors, fix that.

But a crawler cannot know a single one of the things that decide whether you get the job:

• What trade you're actually in, and which of the dozen ways a customer might describe it in French or English
• Which businesses appear above you in the map pack when someone in Aylmer searches for your service at 7 p.m. on a Sunday
• Whether the phone number on the page is the one that rings
• Whether anyone calls back
• Whether your site is compliant with Quebec law

That last one is not a technicality in Gatineau. It's the one that carries fines.

What it costs, plainly

The study June 2026 consumer study (1,227 US consumers who had searched for a local business in the previous three months) found that 75% decide which business to use in under 30 minutes, and 72% consider three or fewer businesses before choosing. In a companion July 2026 release, 52% said they had looked at a business and decided not to contact it, and 65% of those rejections involved a problem with reviews. Only 6% said they simply click the top result.

Read that carefully, because it cuts both ways. Ranking first doesn't win the job. Being one of three, and being the easiest of the three to say yes to, wins the job. (Caveat worth stating: this is a US online panel. There is no equivalent Canadian or francophone study in the public record. Treat it as a directional prior, not an Outaouais measurement.)

As for what a lost shortlist spot costs you, don't take a number from an American software vendor. Take your own average job value, multiply it by the jobs you didn't hear about last season, and you'll have a figure that's actually true. In a region where the National Capital Region shed roughly 9,400 federal positions between 2024 and 2026 (Treasury Board of Canada open data, March 2026) and Outaouais unemployment sat at 6.6% against a Quebec average of 5.6%, that shortlist is shorter and harder-won than it was three years ago.

The shape of a real audit

A proper audit is a diagnostic, not a score. It runs in a specific order, because fixing one layer before the layer beneath it is wasted work. You can't improve conversion on a page nobody reaches, and you can't improve reach for a business Google has filed under the wrong trade.

1. Identity and eligibility, the map layer, first

What's checked: your fiche Google entreprise as a separate asset from your website. Primary category, business name, verified address, service area, hours, and the services list attached to the profile.
Why first: in the 2026 Local Search Ranking Factors, 47 independent local-search practitioners scored 187 factors, and the highest-scoring local-pack factors were primary category, proximity of the address to the searcher, keywords in the business title, having a physical address in the city of search, and being open at the time of the search. The study is blunt about what that is:

None of the experts have special access to the internal workings of Google's local search algorithm.

So call it what it is, the strongest available expert consensus, not a leaked weighting. It still points at something real: the single field most likely to be quietly wrong on a contractor's profile is the primary category. A furnace specialist filed as "Contractor" is competing in the wrong race.
What it reveals: whether you are even eligible to appear for the searches that matter, before anyone touches the website.

2. Reachability

What's checked: which of your pages search engines can actually retrieve, render, and index, and which exist only in your own navigation menu.
What it reveals: the gap between the site you paid for and the site that exists publicly. This is the one thing free scans do reasonably well, which is why they lead with it.

3. The contradiction sweep

What's checked: your name, address, phone, and hours as they appear across your site, your Google profile, your Facebook page, your directory listings, and your RBQ registration.
What it reveals: mismatches. An old phone number on a directory page from 2019 is still routing calls into a dead line. Holiday hours that were never updated tell a customer at 4 p.m. that you're closed. The study July 2026 data has "clear and complete information, hours, address, phone" as the top stated decision factor at 32%.

4. The decision surface

What's checked: what a stranger on a phone sees in the first fifteen seconds. Is the trade named? Is the service area named, not "the Ottawa region" but Hull, Aylmer, Buckingham, Val-des-Monts, Orléans, Rockland? Is there one obvious next action, and does it match how this customer actually wants to reach you? Where does a form submission land, and who reads it, and how fast?
What it reveals: whether your traffic problem is actually a traffic problem. Plenty of sites in the Outaouais get found and lose the job on the page.

5. Reviews as live inventory

What's checked: not just the star average. Count, recency, distribution across platforms, and whether anyone answers them.
Why it matters: The study Local Consumer Review Survey (February 2026, 1,002 US adults) found 68% will only use a business with four or more stars, and 74% look for reviews written in the last three months. Reviews are perishable stock. A 4.9 average built entirely in 2023 reads as a business that has gone quiet.

6. Speed, argued honestly

What's checked: loading behaviour on a real phone on a real mobile connection, not a lab score on office fibre.
What it reveals: a conversion problem, not a ranking problem. Google's own Vodafone case study (web.dev, 2021) is the strongest evidence available because it was a genuine A/B test, half of paid traffic to an optimized page, half to the baseline, and a 31% improvement in Largest Contentful Paint produced an 8% increase in total sales and a 15% lift in the lead-to-visit rate. Meanwhile, nothing in the study 187-factor survey puts page speed anywhere near the top local-pack factors. Anyone selling you speed work as a référencement local Gatineau ranking fix is overclaiming. Sell it as what it is: fewer people leaving before the page finishes.

7. Language and legal exposure

What's checked: on the Quebec side, whether a French version exists, whether it's a real version or machine output, and whether it is offered on terms no worse than the English one.
Why it matters: section 52 of the Charter of the French Language requires that brochures, catalogues and "any other documents of the same nature" available to the public be drawn up in French, and the OQLF lists websites and social media content among these. There is no employee-size threshold, a one-person contractor is covered exactly like a chain, and penalties for corporations run from $3,000 to $30,000, with each day of non-compliance treated as a separate offence. Commercially, the same check matters even if you ignore the law entirely: 25.9% of Gatineau residents know French only (2021 Census). A quarter of your market cannot read an English page. Confirm any specific legal question with counsel; this is a diagnostic flag, not legal advice.
What it reveals: the single most commonly missed item in every automated audit we've seen, because no crawler on earth is checking a Gatineau site against Quebec statute.

8. Measurement

What's checked: whether calls, form fills, and directions requests are being recorded at all, and whether the numbers in your reports reconcile with the jobs in your calendar.
What it reveals: whether anything you do next can be judged. Without this layer, every later decision is a guess with a confident voice.

Free scan versus diagnostic

Before and after

All figures below are illustrative, a composite built from patterns our team sees repeatedly in Gatineau and Ottawa trades, not a real client's results.

Before. Primary category set to a generic contractor label. Nine reviews, most recent fourteen months old. Site loads in about six seconds on LTE. The contact form emails an address the owner stopped checking in 2022. The French page is a machine translation of the English page, with the English version as the default landing. Roughly nine inbound calls a month in shoulder season, source unknown.

After (illustrative, ninety days). Primary category corrected to the trade actually searched for. Three service pages that name the trade and the sectors served. Review requests going out as a routine part of closing a job, producing four to six new reviews a month. Largest Contentful Paint down near two seconds on mobile. Form submissions routed to a phone that gets checked. French-first, with English available on equal terms. Result in this composite: inbound calls roughly doubled, and, more importantly, every one of them traceable to a source.

What it looks like when it's working

• You can name where your last ten jobs came from without guessing.
• Your newest Google review is weeks old, not seasons old.
• Someone searching your trade plus "Aylmer" or "Orléans" finds you in the map pack, and the hours they see are true.
• Nobody has to scroll or pinch to find the phone number on a phone.
• The French version is the real version, not an apology.
• When a month is slow, you can see which layer got slow instead of blaming the whole internet.

Why this doesn't get done

Not because owners are careless. Because that list is eight distinct disciplines, three of them changing every quarter, the study June 2026 tracking shows ads in the local pack going from about 1% visibility in early 2025 to 22% by December 2025, and Local Services Ads from 11% to 31% of tracked queries. The map pack a Gatineau contractor looked at eighteen months ago is not the one his customers see today.

You know your trade at a level no agency will ever match. This is simply a different trade, and it is a full-time one. That is the honest argument for hiring someone: not cleverness, dedication. A real audit is worth having even if you act on it alone, knowing which of the eight layers is broken is most of the value, and you now know what the categories are and what each one reveals.

If you'd like our team to run the diagnostic on your business, UrbanTechFlow works across Gatineau, Ottawa and the Outaouais, in French and in English. Reach us at 819-550-9153.