Capstone: a programmatic SEO page system plan with a quality rubric

0:00 Capstone

This is your capstone, and it's where everything in the course becomes one plan. You'll design a programmatic page system that could survive Google's spam policies and a skeptical editor. The deliverable is a page system plan, a calibrated quality rubric, and three page mock-ups. In this lecture I'll walk through the scenarios, the eleven-section template, an abbreviated example, an AI review prompt, and how you'll be assessed.

0:30 Why it matters

Why does this capstone matter? Because a page system plan is a promise to users, to search engines, and to whoever pays for it, that every page will deserve to exist. Here's an analogy. Before a city approves a new housing development, planners check roads, water, schools and safety. They don't approve a thousand houses because the land is available. Your plan is the planning application. It proves demand, shows the infrastructure of data and templates, and explains how quality will be maintained after the first residents move in.

1:09 Scenarios

Pick a scenario or use your own business. Option A, a home services marketplace in the UAE and Saudi Arabia, with service in district pages in Arabic and English. Option B, a UK software company with integration and use case template pages. Option C, a Pakistani edtech with exam, subject, past papers and tutors in city pages. Each one tests different skills: localization, product data, or licensing and expert content.

1:39 Sections 1–3

Sections one to three are the foundation. Opportunity and intent: patterns, aggregated demand, intent checks and search features including AI Overviews. Policy fit: your unique value statement per page type, exactly what won't be published, and no shortcuts like site reputation or expired domain tricks. And data: sources, licenses, fields, freshness, owner, validation, the value-add layer and privacy handling.

2:05 Sections 4–7

Sections four to seven are the build. Template: block by block, the data behind each, conditional logic, noindex rules, performance and accessibility, including right-to-left where needed. Internal linking: hubs, spokes, related link rules, facets and pagination. Structured data: types, the generator, and validation tests. Indexing and rollout: sitemaps per type, pilot size, observation window, expansion criteria and monitoring.

2:30 Simple example: calibrating the gate (illustrative)

Let's do a simple worked example of calibrating your gate with mock-ups. For scenario A, home services in district pages. Mock-up one: a district with thirty providers scores high. Mock-up two: a district with twelve providers scores well. Mock-up three: a district with exactly your threshold of five providers. You score it, and the page feels thin: the price range is based on too few jobs, and the list barely fills a screen. So you raise the threshold to eight and add a rule that the price range only shows when there are at least twenty completed jobs. That's calibration: letting evidence set the gate.

3:16 Sections 8–11

Sections eight to eleven make it trustworthy. The AI workflow: tasks, grounding, prompts, validation, review levels, logging, model versions and privacy. The quality system: rubric, publishing bar, sampling, automated checks, accountability pages and AI disclosure. Measurement and pruning: scorecard, cohorts, decision framework, cadence and AI visibility tracking. And risks: your top five across policy, data, legal, technical and market, each with a mitigation.

3:43 Example: Scenario C (abbreviated)

Here's the abbreviated example for scenario C. Patterns: exam, subject, past papers, and subject tutor in city, with past-paper pages linking to tutor pages. Policy fit: host papers only where the exam board permits it, otherwise link to official sources, and add original teacher-written worked solutions as the unique value. The gate: worked solutions for most questions, illustratively seventy percent, and eight verified tutors for tutor pages. AI tags questions to syllabus topics and drafts hints that teachers edit, but no AI solution is published without teacher verification.

4:21 Rollout + AI review

The rollout for that example: pilot a hundred past-paper pages for two subjects, and expand after eight weeks if indexed share and engagement hit targets. Measure clicks per indexed page, engagement with solutions, tutor enquiries from past-paper pages, and AI feature impressions. Then use the lesson's AI review prompt. It asks a model to act as a senior SEO lead and a quality reviewer, find scaled content abuse risks, check data and gates for loopholes, test the AI workflow, and list legal questions, in a table. Tell it to be skeptical and not to praise.

5:02 Assessment

Your rubric is a deliverable too. Adapt it from the editorial QA lesson, with criteria, scoring guidance and examples, a publishing bar, and who scores. Then test it on your three mock-ups and include the scores. You'll be assessed on seven things: policy fit, data, template, architecture, rollout, AI and quality, and measurement. Excellent means clear unique value, proprietary and computed data, answer-first templates, controlled architecture, a staged rollout, grounded AI with a calibrated rubric, and cohort measurement with pruning.

5:37 Mistakes + try this now

Common capstone mistakes. A plan full of templates and URLs but no unique data. No explicit list of what won't be published. AI workflows without grounding or review levels. A rubric that's never tested on real mock-ups. And no plan for what happens after launch. Try this now: write the policy fit section first, in five sentences: the unique value per page type, what will never be published, the data thresholds, how AI is controlled, and who is accountable. If those five sentences are strong, the rest of the plan will follow.

6:17 Watch me do it: risks + mitigations (illustrative)

Watch me do it. Let's write the risks and mitigations section for scenario B, the UK software company building integrations and use-case template pages. Risk one, policy: template pages drift into generic AI text. Mitigation: every template page must include a real, downloadable workflow tested by our team, or it isn't published. Risk two, data: integration details go stale when partner APIs change. Mitigation: nightly checks against our integration catalog; pages auto-flag when a connector is deprecated, and support tickets link to the page owner. Risk three, legal: partner logos and trademarks. Mitigation: use them only under partner program terms; otherwise text names only. Risk four, technical: the documentation site and marketing site both publish similar setup instructions. Mitigation: one canonical location, with the other linking to it. Risk five, market: AI assistants answer does X integrate with Y directly. Mitigation: make pages worth the click with the downloadable workflow, setup video and limits table, and track AI impressions and citations monthly. Each risk has an owner and a review date. That table is often the first thing a senior stakeholder reads, so it earns trust for the rest of the plan.

7:40 Your next step

That's the course. You've learned where Google draws the line, how to find patterns and build unique data, how to design templates, links and structured data, how to manage indexing, how to use AI responsibly with editorial QA, how to stay visible in AI search, and how to measure and prune. Your next step: build your page system plan, score three mock-ups with your rubric, run the skeptical AI review, and fix what it finds.