How to Evaluate a Content Marketing Proposal (2026)

The decision in one paragraph

A content marketing proposal is good when you can read it and predict what will be on the website 12 months from now. Most proposals fail this test because they describe activity (blog posts per month, social posts per week) instead of outcomes and assets. The proposals worth funding describe a small number of compounding asset types (cornerstone pages, original research, authoritative comparisons), name the specific audience problem each asset solves, propose a publishing cadence the team can sustain, and report on outcome metrics (organic sessions to commercial pages, citation share in AI search, conversion-assisted revenue) — not vanity metrics. Canadian content marketing programs that produce 1–2 substantive assets per month usually outperform programs producing 8–12 short blog posts.

What 'content' should mean in 2026

The default 'content marketing proposal' from most agencies still describes a publishing schedule: 8 blog posts per month, 12 social posts per week, 1 newsletter every two weeks. This was a viable program in 2018 when surface-level content could rank and earn social engagement. It is not a viable program in 2026.

The shift, in practice: Google rewards substance and topical authority through the Helpful Content System and site-quality signals; AI assistants cite the asset that answers the underlying user question with the most depth and original perspective; LinkedIn and X engagement decays fast and produces minimal organic compound. A proposal that doesn't reflect this shift is selling 2018 content marketing in a 2026 wrapper.

• Substance over volume

  Two cornerstone assets per month, each 1,800–3,500 words with original perspective, will outperform 8–12 short posts in nearly every Canadian B2B and consumer market we've measured.
• Asset durability over campaign activity

  Content that compounds over 12–36 months matters more than content that drives a one-week traffic spike. Ask the proposal what the assets will still earn in month 18.
• Distribution-aware production

  Content that's built to be cited (clean structure, FAQs, original data, quotable framing) earns AI citations and earned PR. Content built to be 'shared' usually doesn't get either.

The 6 asset types worth paying for

• Cornerstone pillar pages

  1,800–4,000 words on a topic central to your business. Designed to rank for the head term, earn citations, and serve as the hub for related cluster pages. 1–2 per quarter is a reasonable cadence.
• Original research and survey reports

  Surveys, audit datasets, benchmark studies. Earn coverage and citations on their own merit. 1 per quarter for most Canadian SMBs; 1 per month for larger programs.
• Authoritative comparison pages

  'X vs Y', 'Best X for Y', 'Alternatives to X'. High commercial intent, defensible if written from real product experience. 6–10 per year typically.
• Operator-led case studies and methodology pieces

  What you actually did, with numbers and the parts that didn't work. The most defensible content type because it's hard to fake. 1–2 per quarter.
• Vertical-specific playbooks

  Long-form guides for a specific industry or use case. Earn organic search and act as sales-cycle assets simultaneously. 4–8 per year for established programs.
• AEO-optimized FAQ clusters

  Cleanly-structured Q&A around the questions your buyers ask AI assistants. Earn AI citations, support featured snippets, and reduce sales-call burden.

The 10 questions that surface real strategy

• Walk me through 3 cornerstone assets you'd build for us in the first 6 months

  Real strategists will name specific topics with real intent rationale. Sales pitches describe 'editorial calendars'.
• Who specifically will write the content?

  Names, credentials, vertical experience. 'A team of writers' is rarely an answer that produces depth.
• How will you handle subject-matter accuracy in our vertical?

  Real programs describe an SME-review workflow. Vague answers produce content that reads like it was written by a non-expert (because it was).
• What's the editorial review process before publish?

  Real programs describe stages (outline review, draft review, SME review, final QA). 'We have an editor' isn't enough.
• How are you using AI tools in production?

  Honest answers describe specific use cases (research synthesis, outline generation, fact-checking) and explicit limits (no AI-generated final copy without human rewrite). Vague 'we use AI carefully' is a tell.
• What's the cadence and why?

  1–2 cornerstone assets/month is defensible. 12 short posts/month rarely is.
• How do you measure success?

  Outcome metrics (organic sessions to commercial pages, AI-citation count, conversion-assisted revenue) — not page views or social shares.
• Show me a content brief you'd send to a writer

  Real briefs are detailed and unfun to read. The quality of internal documentation predicts the quality of the output.
• What happens to the assets when our engagement ends?

  All assets, source files, briefs, and editorial calendars stay with you. If the agency claims partial ownership, walk away.
• Can you show me 2 published assets you've produced for clients in our stage?

  Live URLs, not anonymized PDFs. If 'NDAs prevent it', the work is either thin or the client wouldn't be proud of it.

What the deliverable should actually look like

A real content asset, in 2026, is not a 600-word blog post. It's a structured, depth-substantial page with the supporting elements that earn citations and conversions:

• 1,800–3,500 words of substantive content

  Length isn't the goal — depth is. But the questions worth ranking for in 2026 generally require this length to be answered well.
• Original framing or data

  An angle, a framework, a dataset, a real example. Not a synthesis of what other articles already say.
• Structured headings and TOC

  Skimmable, anchor-linkable, AI-extractable. Clear H2/H3 structure with descriptive headings.
• FAQ section with FAQPage schema

  5–10 real questions with real answers. Earns AI citations, featured snippets, and direct user value simultaneously.
• Author identity and date

  Real byline with a real bio and a publish/updated date. Schema.org Person and Article schema.
• Internal links to commercial pages

  The asset should support the buyer journey, not exist in isolation. Every cornerstone asset connects to 3–8 commercial-intent pages.
• Specific distribution plan

  How will the asset be promoted? What earned-PR pitches will it support? What sales-team enablement will it provide?

Outcome metrics vs. vanity metrics

Honest C$ ranges by program shape

Red flags in content proposals

• Volume-led pricing ('X posts per month for C$Y')

  Tells you the agency competes on production volume, not asset quality.
• AI-generated content with light human edit

  If they describe 'AI-assisted production' without naming a substantive human-rewrite step, the deliverable will read like AI.
• Vague subject-matter expertise process

  Without a real SME-review workflow, the content will be wrong on the details that matter to your buyers.
• Outsourced offshore writing without editor oversight

  Common cost structure that produces content that doesn't sound like you, doesn't reflect your market, and doesn't convert.
• Agency retains rights to assets, briefs, or templates

  All work product should belong to you. Anything else creates a future ransom situation.
• No connection between content cadence and business calendar

  If the proposal doesn't reference your launches, sales cycle, or seasonal demand, it's not strategic content — it's a publishing schedule with your logo on it.

Frequently Asked Questions

The questions Canadian operators ask before signing a contract.