Marketing & Ads

What Should a Canadian Business Actually Publish? A Content System That Does Not Eat Your Week

Most Canadian businesses should publish one substantial piece per week, built into a topic cluster around a page that actually sells something. That is the content marketing system. Volume is not the lever. One...

Cody WiseCody WiseFounder, Wise Media16 min read
Wise Media card reading What a Business Should Publish, with the line One deep piece a week beats five thin ones.

Quick Answer

Most Canadian businesses should publish one substantial piece per week, built into a topic cluster around a page that actually sells something. That is the content marketing system. Volume is not the lever. One 2,000 word article that answers a real buyer question, links to the service page it supports, and gets refreshed twice a year will out-earn twenty thin posts published to hit a quota.

By Cody Wise, founder of Wise Media. Wise Media builds and runs content systems for Canadian agencies, operators and service businesses. This article describes the system we use on our own site and for clients.

Summary

  • Publish one deep piece a week, not five thin ones. Depth and internal linking move rankings. Raw post count does not.
  • Every piece belongs to a cluster. A cluster is one money page plus the posts that answer every question a buyer has before they enquire, all linked to that page and to each other.
  • Google says the test is who, how and why. Content made primarily to attract search traffic rather than to help the reader is the thing its guidance names as the problem.
  • There is no AI-specific markup to add. Google states there are no additional requirements to appear in AI Overviews or AI Mode beyond being indexed and snippet-eligible.
  • Budget 6 to 10 hours a week, or roughly $1,200 to $4,000 CAD a month outsourced, based on published Canadian agency and freelance rates. Verify current rates before you commit.
  • Measure nothing at 30 days except publishing consistency. Impressions at 90. Qualified enquiries at 180.
  • Content is the wrong channel when you need revenue inside 60 days, when your offer is not clear, or when nobody is searching for what you sell.

Table of contents

  1. What should a Canadian business actually publish?
  2. Why volume stopped working
  3. The cluster model, explained for an operator
  4. The one piece per week system
  5. The repurposing chain
  6. What to measure at 30, 90 and 180 days
  7. What a content system costs in Canada
  8. Nine mistakes that kill content systems
  9. When content is the wrong channel entirely
  10. FAQ
Cody Wise working on a laptop at an outdoor cafรฉ table in white linen
A content system has to survive a normal working week. If it only runs on the weeks you feel like writing, it is not a system.

What should a Canadian business actually publish?

Publish the answers to the questions a buyer asks in the two weeks before they hire someone. Not industry news. Not thought leadership. Not what your competitors are posting. The specific, slightly awkward questions people type into a search box when they are close to spending money and do not want to look uninformed on a sales call.

For most Canadian service businesses that list is between fifteen and forty questions, and it is almost entirely predictable. Pull it from four places:

  1. Your inbox and your call notes. Every question a prospect has asked you twice is a post. This is the highest-value source and almost nobody mines it.
  2. Google Search Console. Queries where you already get impressions and no clicks are pages you half-rank for. Those are the cheapest wins on the entire site.
  3. The “People also ask” box on the three queries closest to your offer.
  4. Communities where your buyers complain. Reddit, trade association forums, local operator groups. The phrasing people use there is the phrasing to put in your headings.

Then sort that list by how close each question sits to a purchase. A post titled “how much does X cost in Canada” belongs at the top. A post titled “the history of X” belongs nowhere. If you want this built and run for you rather than assembled in-house, that is exactly what our content production service does.

The four formats that carry almost all of the load

FormatWhat it answersBuyer distancePublish how often
Cost guideWhat will this cost me in CAD, and what moves the numberVery closeOne per service, refreshed twice a year
ComparisonShould I pick A or B, or do it myselfCloseOne per real decision your buyers face
Process explainerWhat actually happens if I hire you, step by stepCloseOne per service
Regulatory or complianceAm I allowed to do this in my provinceVaries, often very closeAs rules change, cited to the regulator

Notice what is missing. No “5 trends to watch in 2026.” No “why branding matters.” Those posts get written because they are easy, and they rank for nothing because nobody searches for them with money in hand.

Why volume stopped working

Because Google wrote down the test, and the test is not how many pages you have. Its documentation on creating helpful, reliable, people-first content asks three questions about any page: who made it, how it was made, and why it exists. Google calls the why “perhaps the most important question,” and says content should be created primarily to help people who visit it directly, not mainly to attract search engine traffic.

That same page lists the warning signs of search-engine-first content explicitly. Producing lots of content on many topics hoping some of it will rank. Summarising what others say without adding value. Changing the date on a page to make it look fresh. Those are not Wise Media opinions. They are in Google’s own guidance, in those words.

The scaled content rule applies regardless of who wrote it

Google’s spam policies define scaled content abuse as generating many pages for the primary purpose of manipulating search rankings, and the policy is explicit that it applies no matter how the content was created. A human writing forty thin posts a month is in the same bucket as a script doing it. The method is not the violation. The thinness is.

This matters for Canadian operators because the agency pitch of “we will publish you twenty blogs a month” is still being sold, and it is now an active liability rather than a neutral waste of money.

And AI search did not change the inputs

There is no separate optimisation layer for AI answers. Google’s guidance on AI features in Search states that to appear as a supporting link in AI Overviews or AI Mode a page must be indexed and eligible to appear with a snippet, and that there are no additional requirements. No new file to upload. No AI-specific schema. The fundamentals carry: let it be crawled, keep important content in text rather than images, use internal links, and make your structured data match what a reader actually sees.

Google also notes that AI Overviews traffic shows up in Search Console under the “Web” search type rather than in a separate report, which is worth knowing before you spend a morning looking for a filter that does not exist. If you want the mechanics of how AI answer engines treat a site, we covered the related question of whether Google penalises AI-generated content separately.

The cluster model, explained for an operator

A cluster is one page that sells, plus the posts that answer every question a buyer has before they fill in the form, all linked to that page and to each other. That is the entire model. Everything else written about topical authority is decoration on top of it.

Here is why it works in plain terms. A single post about “how much does a website cost” is one page competing against every other page on that query. A cost post that links to a services page, and is linked from a timeline post, a redesign post and a maintenance post that all link to each other and to the same services page, is a structure. Search engines read that structure as evidence that the site covers the subject rather than having touched it once.

The three links every post must carry

  1. One link to the money page, in the first third of the article, inside a sentence that would be there anyway. Not a banner. Not a footer button.
  2. Two links to sibling posts in the same cluster, placed where a reader would genuinely want them.
  3. One link to a post in a different cluster, which is what stops your site becoming a set of unconnected islands.

Four links. That is it. Miss them and you have a collection of separate pages. Add them and you have a site.

What a finished cluster looks like

LayerExample for a Canadian web design firmCount
Money page/services/website-packages/1
Cost and pricingWhat a website costs in Canada, what drives the number1 to 2
Process and timelineHow long a build takes, what you need to supply2 to 3
Decision and comparisonAgency vs builder vs freelancer, redesign vs rebuild2 to 3
Risk and complianceAccessibility, privacy policy, contract terms2 to 4
AftercareMaintenance, backups, what happens post-launch2

That is eight to fifteen posts. At one a week, a cluster takes a quarter. Two clusters a year is a realistic ambition for a business owner doing this alongside the actual work, and two well-built clusters beat eighty scattered posts every time.

The failure mode is not writing badly. It is writing well about eleven unrelated subjects, so that no single subject ever accumulates enough weight to rank.

The one piece per week system

One article per week, produced across four short working sessions rather than one long one. The reason to split it is that the single-sitting approach is what collapses the first week you have a bad client, and a cadence that collapses is not a system.

Cody Wise at a multi-monitor founder workstation in a home office
Four sessions of 60 to 90 minutes. The split is what makes the cadence survive a bad week.

The weekly loop

  1. Monday, 30 minutes. Pick and outline. Take the next question off the cluster list. Write the headings before you write a sentence of body copy. If you cannot write six headings that each answer something specific, the topic is not ready and you should take the next one instead.
  2. Tuesday, 90 minutes. Source it. Find the primary sources. Statutes, regulators, official documentation, published price lists. Paste the exact figures and the links into the outline. Anything you cannot source gets cut here, not softened later.
  3. Wednesday, 90 minutes. Draft. Write straight through from the outline without editing. Answer each heading in the first 40 to 60 words beneath it, then expand. This is also what makes a section quotable in a featured snippet or an AI answer.
  4. Thursday, 60 minutes. Edit, link and ship. Cut 15 percent. Add the four internal links. Add the metadata and schema. Publish.

That is four and a half hours of focused work, and it produces something publishable every week. Running it on a Friday afternoon in one block does not, because by hour three you are writing filler to reach the end.

Answer-first, every section

Put a direct 40 to 60 word answer immediately under every H2, then expand below it. This single habit does three things at once: it makes the page scannable for a reader in a hurry, it gives Google a clean block to lift into a featured snippet, and it gives an AI answer engine a self-contained passage it can summarise without mangling your meaning. It costs nothing and most sites still do not do it.

Refresh beats republish

Two refresh slots a quarter, used on the posts that already earn impressions. Update the figures, re-check that every cited source still says what you said it says, add anything that changed, and update the modified date honestly. Changing a date without changing the content is one of the behaviours Google names as a warning sign, so do the work or leave the date alone.

The repurposing chain

One article should produce roughly nine assets. Not by chopping it into fragments, but by taking the one idea in it that is genuinely contested and arguing that idea in each channel’s native form.

AssetSource from the articleTime
LinkedIn postThe single most contested claim, argued in 300 to 500 words15 min
X postThe sharpest contrast, under 280 characters, with the link5 min
Instagram captionThe reader-facing version of the problem, plus hashtags10 min
Email to the listThe answer-first intro, then the link10 min
Two to three short videosOne heading each, spoken to camera30 min
Carousel or slide deckThe comparison table, one row per slide20 min
Sales follow-up snippetThe objection the post handles, pasted into a proposal5 min

The sales follow-up snippet is the one most businesses skip and the one that pays fastest. If a prospect asks a question you have already answered in 2,000 researched words, sending the link closes the loop faster than typing a reply, and it quietly demonstrates that you have thought about this before.

Rules that keep repurposing from reading like repurposing

  • Never paste the article’s opening paragraph into a social post. Write the platform’s own opening.
  • One link per LinkedIn post. The preview card locks onto the first URL it parses, so a second link breaks the card.
  • Do not post all nine assets the same day. Spread them across the two weeks after publication, which is also when the post is being indexed.
  • If the idea was not contested enough to argue on LinkedIn, it was probably not worth 2,000 words either. That is useful feedback about the topic list.

What to measure at 30, 90 and 180 days

Measure almost nothing early. Content has a long and genuinely frustrating lag, and the main way owners kill a working system is by judging it at week six against a metric it cannot possibly move yet.

CheckpointThe only question that mattersWhat you should seeWhat to ignore
30 daysDid we actually publish four times4 of 4 published, cluster links in placeTraffic, rankings, conversions
90 daysAre we being shown for anything newImpressions rising on cluster queries in Search Console, new query strings appearingPosition averages, bounce rate
180 daysIs any of it producing enquiriesClicks on cluster pages, assisted conversions, people quoting your posts on sales callsSocial engagement, time on page

The three numbers worth a dashboard

  1. Impressions on cluster queries, in Search Console. This moves first and it is the earliest honest signal that the structure is working. Filter to the cluster’s page group, not the whole site.
  2. Clicks to the money page from blog posts. If the internal links are doing their job this number climbs even while total traffic is flat. If it does not climb, your links are decorative.
  3. Enquiries that mention a post. Add one optional field to your intake form asking what brought them in. Low-tech, and the single most useful content metric most businesses do not collect.

Note one nuance. Google states that clicks coming from AI Overviews pages tend to be higher quality, with users more likely to spend more time on the site, while total clicks can fall because the answer was given on the results page. A quarter where clicks drop and enquiries hold steady is not a failure, and treating it as one leads people to dismantle something that is working.

What a content system costs in Canada

Either about six to ten hours a week of your own time, or roughly $1,200 to $4,000 CAD a month outsourced for a genuine one-post-a-week system with research, editing and internal linking included. The ranges below reflect published Canadian freelance and agency rates at the time of writing. They are published rates rather than a surveyed dataset, no methodology is attached to them, and you should verify current pricing directly before budgeting against them.

ModelPublished range (CAD)What you getWhere it breaks
In-house, owner writesYour time, 6 to 10 hrs/weekBest subject knowledge available, zero cash costFirst busy quarter ends it
Freelance writer, per post$300 to $900 per 2,000 word postDrafting capacityYou still own topic choice, sourcing, linking and publishing
Freelance writer plus your editing$300 to $900 plus 2 hrs of your timeDrafting plus your judgementScales until your editing time runs out
Agency retainer, managed$1,200 to $4,000+ per monthStrategy, production, publishing, internal link graph, reportingOnly worth it if the strategy layer is real, not just drafting with a markup

The question to ask any agency before you sign

Ask them to show you the internal link graph they would build, cluster by cluster, with the money page named for each. An agency that answers with a volume number has sold you drafting. An agency that answers with a structure has sold you a system. The difference is not price, it is whether anything compounds. Our own website growth packages are built around that structure rather than a post count, for exactly this reason.

Nine mistakes that kill content systems

  1. Publishing to a quota instead of a cluster. Twenty unrelated posts is twenty unrelated pages.
  2. No internal links. The single most common and most expensive omission. The post ranks alone or not at all.
  3. Topics chosen by what is easy to write. Easy topics are easy because everyone already wrote them.
  4. No primary sources. Citing three other agency blogs is how a wrong number propagates for five years.
  5. Burying the answer. Four paragraphs of throat-clearing before the thing the reader searched for.
  6. Letting posts compete with each other. Two posts on the same query split their own rankings. Merge them and redirect the weaker URL.
  7. Measuring at 30 days and quitting. Nothing has happened yet. Nothing was going to.
  8. Changing dates instead of content. Named by Google as a warning sign, and readers notice too.
  9. Writing for peers instead of buyers. Your competitors are not going to hire you.

The one that is hardest to see

Writing in a register that signals the content was produced rather than thought about. Symmetrical paragraphs, identical section lengths, a hedge in every sentence and no specifics anywhere. Readers now pattern-match that instantly and so, increasingly, do the systems summarising your page. Vary your paragraph length. Name the province, the tool, the actual dollar figure. We wrote about this failure mode in more detail in how to use AI without your content sounding like AI.

When content is the wrong channel entirely

Content is a compounding asset with a six to twelve month lag. There are four situations where starting it is the wrong call, and an honest agency tells you so before taking the retainer.

SituationWhy content fails hereWhat to do instead
You need revenue inside 60 daysThe lag is longer than your runwayPaid search on high-intent terms, plus direct outreach
Nobody searches for what you sellThere is no demand to capture, only demand to createPaid social, partnerships, events, outbound
Your offer is not clear yetYou will publish forty posts arguing for something you change in MarchFix positioning and the money page first
You are a local service with no Google Business ProfileThe map pack is most of the available demand and you are not in itClaim and optimise the profile first

That last one is worth stating plainly. For a lot of Canadian local service businesses, trades and clinics especially, a properly built Google Business Profile produces more booked work in the first quarter than any blog will in the first year. Blogging before you have done that is spending on the second-best channel while the first one sits empty.

Content is also not a substitute for being cited elsewhere. Owned-media volume is the weakest lever available for getting mentioned in AI answers, because those systems overwhelmingly cite third-party sources. Reviews, directory profiles, partner content and press do work your own blog cannot do for you. Publish anyway, because it drives rankings and it closes sales, but do not expect your blog alone to make you the answer.

Frequently asked questions

How often should a small business post a blog?

Once a week, every week, inside a cluster. Consistency over twelve months matters far more than frequency in any given month. If once a week is not survivable, publish every two weeks and keep it, because a cadence you can hold is worth more than one you abandon in March.

How long should a blog post be?

Long enough to fully answer the question and no longer. Google has stated there is no preferred word count. In practice a commercial-intent post that genuinely covers a buying decision lands between 1,500 and 3,500 words, because that is what the subject requires, not because length is a target.

Does AI-written content hurt your rankings?

Not because it is AI-written. Google’s spam policies target scaled content abuse, which is generating many pages primarily to manipulate rankings, and the policy applies regardless of how the content was produced. Thin, unoriginal, mass-produced pages are the problem whether a person or a model wrote them.

Should I write for Google or for AI search?

The same content serves both. Google states there are no additional requirements to appear in AI Overviews or AI Mode beyond being indexed and snippet-eligible, with no AI-specific markup or files needed. Answer-first structure, clear headings, text rather than images, and structured data that matches the visible page cover both cases.

How many blog posts before I see results?

Expect first movement in impressions around month three, and enquiries attributable to content around month six. Eight to fifteen posts inside one cluster, properly interlinked, is the point at which most sites start to see a cluster behave differently from isolated pages.

Can I just repost my blog to LinkedIn?

You can, but it underperforms. Platforms rank native content and readers skim differently there. Take the one contested idea from the post and argue it in LinkedIn’s own form, then link to the article. One link per post, because the preview card locks onto whichever URL it parses first.

What should a Canadian business publish that a US competitor would not?

Anything where the rules differ by province or by country. Privacy and anti-spam compliance, accessibility legislation, trade licensing, tax treatment, pricing in CAD. Those posts are harder to write because they need primary sources, which is exactly why there is less competition on them and why they convert. Our post on building an email list in Canada without breaking CASL is a worked example of that format.

The short version

Pick one cluster. Write the eight to fifteen questions a buyer asks before they hire you. Publish one a week across four short sessions, link each post to the money page, two siblings and one post in another cluster, and refresh rather than republish. Ignore everything for 90 days except whether you actually published. That is the system, and it works because almost nobody sticks with it long enough to find out.

If you want it built and run rather than squeezed into your week, we do this for Canadian agencies, operators and service businesses as a managed system with the cluster map, the editorial calendar and the internal link graph included.

Tell us what you sell and who buys it in the Wise Media intake form and we will come back with the cluster map we would build first, whether or not you hire us to run it.

Cody Wise, founder of Wise Media

Written By

Cody Wise

Founder of Wise Media, the Calgary software development and product design studio. Cody builds brands, websites, apps and AI systems for founders across Canada and the United States, and writes here about what works in live client builds.

Reading About It
Or Getting It Built?

Thirty minutes with Cody. The bottleneck named, what we would build first, and what it costs.