Case walkthroughs: four page systems from idea to iteration

0:00 Case walkthroughs

Theory is useful, but seeing a whole page system from idea to iteration is where it clicks. In this lecture we'll walk through four illustrative composite cases, a UK software integrations directory, a UAE rental property system, a Pakistani tutor marketplace, and a money transfer comparison site for the UK and UAE to Pakistan corridor. Then we'll pull out the lessons they share, so you can apply them in your capstone.

0:31 Why it matters

Why do case walkthroughs matter? Because principles only click when you see them applied end to end, including the parts that went wrong. Here's an analogy. Pilots train in simulators with realistic scenarios, not just by reading manuals. Walkthroughs are your simulator. You'll see the same sequence of decisions play out in four very different markets and business models, and you'll start to recognize the patterns that predict success before you build anything.

1:03 Case 1: UK integrations

Case one, a UK software integrations directory. The pattern: our product plus a tool integration. The data: the real integration catalog, supported triggers and fields, plan availability, setup steps, and aggregated usage. The template answers first, yes, it syncs contacts and deals both ways on the pro plan, then gives setup steps, a fields table, limits and a short video. The gate: only integrations that actually exist. Requests get a vote page that stays out of the index. AI drafts setup steps from internal docs, and product specialists verify them.

1:42 Case 2: UAE rentals

Case two, UAE rentals. Property type for rent in community pages, in English and Arabic. The data: live listings, community facts like service charges, nearby schools and metro distance, computed rent trends, and verified agents. The template leads with count, rent range and trend, then filters, a map, and a local editor's community guide. Communities need ten active listings, otherwise the page is noindexed with an alert sign-up. They rolled out emirate by emirate, with per-community sitemaps, and later merged micro-communities.

2:17 Case 3: Pakistan tutors

Case three, a Pakistani tutor marketplace. Subject tutor in city, and exam board subject tutor, like O-Level physics, with Roman Urdu variants validated in Search Console. Data: verified tutors, aggregated rates from bookings, languages, online or in-person, and availability. Senior teachers wrote how-to-choose guides per subject. The gate is eight verified tutors. Tiny areas merged into city pages, and the team created online tutor pages for cities with little local supply.

2:48 Mini-case: Jeddah wedding venues (illustrative)

Let's add a simple fifth mini-case to practice the sequence yourself. A Jeddah wedding venue directory. Pattern: wedding halls in district, plus women-only wedding halls in district, validated as separate intents in search results. Data: verified venues with capacity, price range, separate sections, catering options and availability. Template: answer first with count, capacity range and price range. Gate: at least five verified venues per district. Linking: district hubs and nearest districts. AI: Arabic and English summaries of venue features from verified data, fully reviewed. Measurement: enquiries per indexed page. Notice how quickly you can now reason through a new case.

3:31 Case 4: Transfer comparison

Case four, money transfer comparison, UK and UAE to Pakistan. Patterns like send money from a country to Pakistan, currency to rupee, and provider versus provider. Data: live mid-market rates, provider fees and rates captured under agreements or public APIs, delivery times and limits. The template has a live comparison table, a total cost calculator, methodology and update time, regulatory notes, and rate alerts. Compliance is built in: financial promotion rules, clear affiliate disclosure, no misleading best claims.

4:05 AI search shift

Case four also shows the AI search shift. Simple rate queries were increasingly answered directly by AI, so the team leaned into what AI can't easily replace: comparison tables, calculators and alerts. Those pages earned citations and clicks, and adding Urdu and Arabic versions, plus the alert feature, drove repeat visits.

4:27 Shared lessons

Now the shared lessons. Data thresholds prevented most problems in every case. Pages that double as product surfaces performed best: integrations, listings, tutor booking, comparison tools. Merging beat keeping thin pages. Localization into Arabic, Urdu and Roman Urdu was a real moat. AI worked best on grounded, reviewed tasks. And compliance was part of the template, not an afterthought.

4:53 Your walkthrough skeleton

The lesson gives you a YAML skeleton to write your own walkthrough: the pattern, how you checked intent, your proprietary, licensed and computed data, template blocks, the gate, linking, indexing and rollout, AI tasks with grounding and review level, measurement, compliance, and the triggers for iteration. Watch for four red flags: a pattern with no proprietary data, no gate or a gate that's too low, AI tasks without grounding, and no plan for stale data.

5:26 Mistakes + try this now

Common mistakes when learning from cases. Copying a case's surface, like its URL structure, without its data advantage. Assuming a pattern that worked in one market transfers directly to another language. Ignoring the compliance details, like financial promotion rules. And forgetting that each case iterated after launch. Try this now: take the case closest to your business and list three things you'd do differently in your market, and why. That's where your competitive edge will come from.

5:59 Quick self-check

Quick self-check. In the money transfer case, the team wanted a headline saying cheapest way to send money to Pakistan on every corridor page. What's the risk? Pause. It's a comparative claim that depends on amount, method and timing, and in the UK financial promotions must be fair, clear and not misleading. A safer approach is to show the calculation for the user's amount and say which provider is cheapest for that amount right now, with the update time and your methodology.

6:35 Watch me do it: YAML skeleton live (illustrative)

Watch me do it. Let's fill in the YAML skeleton live for an illustrative Manchester gym-comparison site. Page system: gyms in area. Pattern: gyms in area, twenty-four hour gyms in area, women-only gyms in area, the last validated as a separate intent because its results differ. Intent check: done in both desktop and mobile results. Data: proprietary, class timetables from partner gyms and member reviews; licensed, gym locations and facilities from a data provider; computed, monthly price ranges and peak-time busyness from check-in data shared by partners, aggregated. Template blocks: answer first with count and price range, list with filters, peak times chart, a guide on contract traps written by a personal trainer, methodology, related areas, and a free pass call to action. Gate: at least six gyms with verified pricing. Linking: city hub, area hubs, nearest areas. Indexing: per-pattern sitemaps, pilot twenty areas. AI: summarize reviews by theme, grounded, sample review. Measurement: free-pass claims per indexed page. Compliance: pricing must include joining fees to avoid misleading comparisons. Iteration trigger: any area below fifty percent indexed after eight weeks gets reviewed.

7:54 Recap and next step

Recap. Four very different page systems followed the same sequence and learned the same lessons: gate on data, build product surfaces, merge thin pages, localize, ground your AI, and design for compliance. Your next step: fill in the walkthrough skeleton for your own page system, then ask a colleague to predict what three sample pages will contain. If they can, your plan is clear enough to build.