Enter passphrase to open the recap
Strategy reset. Before scaling industry × suburb across the country, we prove the model on four industry hubs first. Tight clusters, real schema, real PAA, real case studies, real internal links.
We pivoted off the broad 15-page sprint and committed to a four-industry proof-of-concept before scaling the matrix.
Gayan walked the competitor analysis output (OptWeb, Shout, Clearwater) and we converted it into an HTML dashboard live using the front-end design skill. Read-out: Shout owns Melbourne head terms via age + backlink authority; PMGS has a real backlink-quality problem (volume looks healthy, ranking doesn't match → likely PBN penalty residue from the 2020 era).
Audited the live PMGS site. Some industry pages indexed (healthcare, e-commerce), some not (automotive, childcare). Old programmatic suburb pages still echoing in GSC (Bendigo AdWords, Ballarat AdWords) because Google's index still holds them. New site narrowed scope to ~20 suburbs around the Epping office — right call, but pages aren't getting traction yet because content is too generic and not differentiated enough page-to-page.
Ran a content-similarity audit across the four candidate industry pages. Healthcare is the most differentiated; the others share repeated copy in "How we build your strategy" and FAQ blocks. FAQ content was Claude-generated without real PAA grounding — that's the fix we're making.
Decision: park the suburb expansion. Build one cluster per industry first — hub page + case study + 3 supporting blogs + LinkedIn + video — across four industries (healthcare, automotive, childcare, e-commerce). Measure GSC delta over 2–3 weeks. If it moves, replicate across the rest. If it doesn't, we triage before burning effort on 100+ industry pages.
Live medical-centre paid campaign generating leads — case study to be created. Healthcare hub already on site, most differentiated content of the four. Anchor schema: Medical Practice + YMYL signals.
Existing case study (used-car salesperson, "card yard" client). Two local KWs into top 9, branding clicks +17. Page exists but not yet indexed — push for index after rebuild.
Multi-location early-learning client as the case study. Lives under "industries / retail-hospitality-education-technology" — sub-bucket needs surfacing as its own clean industry hub. Multi-location schema candidate.
Currently bundled with retail — split out into its own hub. Different intent, different schema (Product / Online Store), different funnel. Needs new dedicated industry page built.
Service + LocalBusiness + industry-specific schema. Healthcare = MedicalBusiness/MedicalClinic. Automotive = AutoDealer. Childcare = ChildCare/Preschool. E-commerce = OnlineStore. Plus FAQ + AggregateRating where review data exists. Validate every page in Schema.org validator.
Industry-specific differentiator paragraph (no generic agency copy). Healthcare hits YMYL: compliance, patient privacy, AHPRA awareness. Automotive: dealership stock + finance pages context. Childcare: multi-location intent. Each hub ≥ 50% unique vs the others.
Pull questions from real PAA + AI Overview fan-outs (not invented by Claude). Weave LSI keywords from LSI Graph. FAQs internal-link out to a supporting blog that answers the question in depth.
domain level (PBN era cleanup)by May 09Volume of referring domains is healthy; ranking is not. Likely residue from the 2020-era PBN campaigns. Action: domain-level disavow on the iffy ones via Ahrefs export. Cut the infected finger to save the hand.
Old suburb pages (Bendigo AdWords, Ballarat AdWords) still appearing in queries. They're not on the new site, but Google's index hasn't fully cleared. Don't recreate them. Let them decay; focus the new effort on the four-industry test.
Different intent, different funnel, different schema. Splitting e-commerce into its own hub. Retail can stay in the bundled hub for now until we know whether to invest there.
Hub pages currently show all case studies regardless of industry. Filter so each hub only shows the case studies relevant to that industry. Tighter signal to user + Google.
Existing system instruction is solid, but prompts are missing context. Before each page, feed Claude: real PAA from SERP, fan-out queries, LSI keywords, the case study, the service definition. Walkthrough video coming Friday.
The 20-suburb matrix isn't dead — it's paused. Once a hub cluster moves the needle, we replicate the cluster pattern for /seo-for-{industry}/{suburb}/. Quality first, then scale.
Move the four hubs, then scale. If the cluster pattern works, every other industry gets the same treatment.
Per hub URL, baseline today and re-check at +14d and +21d after the rebuild ships:
→ Impressions in GSC (search appearances)
→ Clicks (intent capture)
→ Average position across the page's keyword cluster (lower = better)
→ Indexation status of the hub + the 3 supporting blogs
→ Branded vs non-branded query mix shift
If 3 of 4 hubs show meaningful improvement, the cluster pattern is validated and we apply it to the rest. If not, we triage before spending on backlinks (Alexander intro stays parked).
Fathom recording, transcript, competitor HTML dashboard, hub-rebuild checklist (when ready).
Walk the four rebuilt hubs, audit the schema, review the first case-study video, and decide whether the cluster engine is ready to replicate.