How we build Phase 5 combo pages — scaling programmatic SEO without doorway spam

Programmatic SEO fails when URL count becomes the goal. Search engines and enterprise buyers both punish thin clones — swapped adjectives, duplicated FAQ, fake local offices, and pages that exist only to rank. Arcloops Phase 5 adds combinatorial pages only when each URL has a distinct body, honest internal links, and a reason to exist beyond keyword variation.

This article is authority documentation for how we scale toward 310–400 indexable pages without doorway spam. It complements the execution plan in our SEO programme and the combo audit script we run before publish. Buyers can use it to judge our pages — and any vendor's.

If you are evaluating Arcloops content: expect explicit delivery honesty, links up to consulting and solutions, sideways links to siblings, and refusal to publish when overlap exceeds threshold.

Allowed combinations — and hard caps

Phase 5 allows use-case × industry, role × industry, and industry × market only when SERP or buyer need justifies a dedicated URL — not when a section on a parent page would suffice.

Each combo page requires 700+ unique words for use-case/role combos (800+ for industry/market), unique H1 and meta, ≥4 FAQ items, ≥3 internal links to live commercial pages, and explicit parent links via combo metadata.

Hard caps apply per pattern (+15 industry×market, +15 role×industry, +20 use-case×industry in the execution plan). Hard stop: if Jaccard overlap with any sibling exceeds 40%, we do not publish.

The uniqueness audit — how we enforce it

Before `published: true`, combo pages pass a word-count floor and pairwise overlap check against parent pages and nearest sibling combos. The audit script tokenises body copy and compares Jaccard similarity — human review catches false negatives.

Authors must document unique angle per URL: what integration surface, regulator, or stakeholder map differs from parents. Template swaps without new failure modes fail review.

Bangladesh Bank, UAE AI Strategy 2031, EU AI Act, and Vision programme claims appear only where accurate — geo pages carry geo claims; combo pages carry industry workflow claims.

What good combo pages look like

Banking AP: /use-cases/ai-invoice-processing-banking-financial-services — dual control, entity routing, withholding fields, month-end close discipline — not generic AP marketing.

RMG payroll: /use-cases/ai-payroll-anomaly-detection-rmg-garments — seasonality, line-level attendance adjacency, workforce scale — not generic HR analytics.

Fintech KYC UAE: /use-cases/ai-merchant-onboarding-kyc-fintech-uae — free-zone context, document classes, escalation to compliance — not generic onboarding.

SaaS helpdesk: /use-cases/ai-helpdesk-ticket-triage-saas-technology — deflection vs quality trade-offs, subprocessors, tier-2 escalation — not generic ITSM.

Each page links up to /ai-consulting, relevant /solutions paths, and parent /use-cases and /industries pages.

What we refuse to publish

Geo city farms — dozens of US or UK metros with identical copy — violate our market slug policy. One primary URL per country; metro callouts inside honest market pages when useful.

Combo grids where industry adjectives swap but integration paragraphs repeat — 40% overlap threshold exists to catch these.

Pages claiming local offices we do not have — stretch markets like /markets/sweden, /markets/france, /markets/nigeria, /markets/kenya, /markets/chile, /markets/colombia are remote/hybrid from Dhaka and Dubai.

Invented ROI, unapproved client names, and case metrics inside programmatic or insight copy — editorial and commercial honesty rules apply uniformly.

Internal linking law at scale

Market pages and top insights link down to industries, roles, use-cases, and guides. Combo pages link up to commercial pages and sideways to 1–2 sibling programmatic pages.

Insights Phase 5 (pillars 20–27) anchor stretch markets and combo MOFU — for example /resources/insights/eu-ai-act-nordic-operators-sweden links /markets/sweden; /resources/insights/use-case-industry-ai-combos-guide links exemplar combo URLs.

Every new page must answer: where does the buyer go next for BOFU action — readiness assessment, contact, or a specific consulting service — without orphan funnels.

Role × industry combos — higher bar

Role pages (/for/chief-financial-officer) and industry pages (/industries/banking-financial-services) already exist. Role × industry combos publish only when buyer committee and workflow map differ materially — example /for/chief-financial-officer-banking-financial-services when banking finance controls dominate the CFO agenda.

RMG role combos — /for/chief-operating-officer-rmg-garments, /for/hr-manager-rmg-garments — name plant reality, seasonality, and workforce dynamics generic COO pages skip.

If the role combo would repeat the industry page's stakeholder section, we strengthen the industry page instead of splitting a URL.

Stretch markets + combos — avoiding double doorways

Stretch markets get market pages first — unique regulatory and delivery copy. Industry × market dedicated URLs publish only when SERP needs them; otherwise industry pages gain a market section and link to /markets/*.

Phase 5 insights for Sweden, France, South Korea, Nigeria, Kenya, Chile, and Colombia exist to support market pages with editorial authority — not to duplicate market body copy.

Buyers comparing regions should use market pages for geo and combo pages for workflow × sector — see /resources/insights/use-case-industry-ai-combos-guide.

How buyers should use this documentation

Procurement teams can treat this article as a vendor content integrity test — ask any AI consultant how they prevent combo spam before RFP shortlist.

Internal SEO and growth teams can adapt the overlap audit pattern — word floors plus similarity checks beat honour-system editorial.

If a combo page you need does not exist, that may be intentional — ask for readiness on parent pages or contact rather than expecting infinite URL multiplication.

Arcloops entry for governed delivery: /ai-consulting/ai-readiness-assessment and /contact?intent=ai-readiness-assessment.

We revisit caps quarterly against GSC impressions and pipeline feedback — a combo that fails to earn clicks or assisted leads may be merged back into parent pages rather than maintained as orphan URLs.