Research brief: What makes a marketing site do something (piece on brochure vs platform)

Status: Foundation research notes, not a finished article. Compiled May 22, 2026.

Thesis: Most marketing sites do nothing because the agencies building them can't build anything else. WordPress + Elementor + contact form is the capability ceiling for most agencies; above that line is custom software, and that's where compounding value starts — structured catalogs, real schema, freshness, citation eligibility, AI visibility.

Argument shape

1. The ceiling exists and is measurable. WordPress runs 43.2% of the web but only ~44% of WP sites pass mobile CWV (vs 65% for Shopify, 83.63% Duda). Elementor abstracts the DOM but not a data layer. Page builders cannot do query-driven catalogs, calculators, or KB-backed content without dropping into custom PHP.
2. Compounding vs decaying. Per Seer (Oct 2025), 65% of AI bot hits target content under a year old. A brochure site decays the day it ships. A database-backed site compounds — every SKU, every KB article, every dated update extends the publish-date freshness window.
3. Citation eligibility is the new ranking. AI Overviews now appear on 48% of tracked queries (BrightEdge, Feb 2026, up 58% YoY). Brands cited inside AI Overviews earn 35% more organic CTR and 91% more paid CTR (Seer, Sep 2025).
4. The Boucher & Jones case (Our observation). A regional lubricant distributor moved from a 12-page brochure to a 350+ SKU Postgres-backed catalog + KB-backed regulatory articles + government-data pricing visualizations. The post-move site has 350+ unique citable pages (vs 12), generates publish/update dates on real catalog changes, and provides extractable statistics (price, viscosity, OEM approvals) — every "addition" lever from the GEO paper, structurally.

Counter-arguments the piece must address

• "Most small businesses don't need a database-backed site." — Conceded. Narrow the claim to "at least one feature the site does."
• "WordPress is fine — it powers 43% of the web." — Fine for what page builders can do. The thesis is about the ceiling, not the floor.
• "Custom software is more expensive upfront." — True. The compounding-vs-decaying argument is the answer.
• "Schema doesn't cause citations." — Conceded (Ahrefs 2026). Schema is infrastructure for the database-backed content that does get cited.

Claims / sources that contradict or complicate the thesis (be honest)

• No primary research proves "small business sites with interactive features convert better than brochure sites." The Mediafly stat is from B2B sales decks, not marketing websites.
• WordPress CWV has improved YoY (28% → 44% mobile from 2023 to 2025-26). Platform is not standing still — gap is narrowing on raw CWV.
• The "20-30 plugins average" is industry rule-of-thumb (WPBeginner), not installed-base data. Use as industry-consensus.
• Page-builder market share is declining; Gutenberg FSE +145% YoY (Colorlib 2026). "Agencies stuck on Elementor" framing holds for SMB segment, broader market is moving.
• The GEO paper has been critiqued (Sandbox SEO) for the Subjective Impression metric construction.
• The internally-named "Manifold" viz library could not be verified as a public Vega-Lite-based tool used by news orgs/government. Use RULE: Never publish a tool or library name without a citation that can be verified by a reader..