Case study · calculator archetype walkthrough · 2026-05-19

FermentCalc: applying the audit to a calculator archetype.

FermentCalc (fermentcalc.com) is a niche sourdough hydration calculator + reference site in the operator's fleet. One calculator, ~12 hydration-by-flour pages, ~18 hydration-by-recipe pages, plus glossary and reference content. Built by a solo operator (Paulo de Vries) over ~11 days in April 2026.

The calculator archetype is the structurally easiest path to a high Citation Readiness Score. This page walks through why — and the small set of fixes that move a calculator site from "already strong" to "top of the archetype".

Why calculator archetypes score so high

The Citation Readiness Score rewards specific structural properties. Calculator and tool archetypes hit several of them by default — not because the operator did anything special, but because the format inherently produces the right shape.

• Inherently fact-shaped first paragraphs. "Sourdough hydration is calculated as water weight ÷ flour weight × 100" is quote-extractable as-is. No editorial paragraph to rewrite. Dual Fit scores high without effort.
• Per-page unique data. The 65% hydration page has different recipe-output text than the 75% page — the calculator generates structurally distinct content per route. No template-thin-content risk.
• Definitional intent matches title. The title is "Sourdough hydration calculator" and users searching that exact phrase want what's on the page. SEO Foundation + Dual Fit both land high.
• Primary-source citations come naturally. Calculators cite a primary source (USDA, NCHFP, King Arthur Flour, The Bread Lab) for the formula or methodology. E-E-A-T signals are inherited from the citation structure.
• Mobile-perfect by necessity. Calculators are used on phones in kitchens. Designers prioritize mobile UX, which incidentally aligns with the SEO Foundation dimension's mobile-perfect signal.

The audit identifies fewer gaps — but they're still real

Calculator sites usually score above the publisher / SaaS / programmatic averages. But the audit still surfaces gaps. For FermentCalc shape:

• Entity Coherence — Person schema with sameAs. Even when Organization schema is solid, the Person schema for the operator/author is often missing. Solo operators tend to skip this step. Adding it + sameAs to LinkedIn + GitHub adds material score lift.
• GEO Readiness — DefinedTerm for niche concepts. "Hydration percentage", "brine percentage", "fermentation time at room temperature vs cold ferment" — wrapping each occurrence with DefinedTerm schema makes them LLM-extractable as terminology, not just prose.
• H2 patterns become repetitive at scale. Across the 47+ calc-result pages, H2 structures ("What is X%", "How to use Y", etc.) repeat. Diversifying the H2 structure across the long-tail lifts the SEO Foundation dimension.
• Wikipedia citation is the +95 archetype that's not in place yet. A sourdough-hydration calculator citing the right authoritative source AND being cited back by Wikipedia is the highest possible citation-graph signal. That's an operator-action (you can't self-cite Wikipedia) but it's the ceiling-raiser for any calculator site.

The lesson for calculator + tool sites

Calculator and tool archetypes are the easiest path to a high Citation Readiness Score. If you ship a calculator site:

1. 1. Make the result page's first paragraph fact-shaped (the calculator output IS the fact).
2. 2. Add Person schema for the operator with sameAs to ≥2 profiles.
3. 3. Add /llms.txt at root + explicit AI-bot allowlist in /robots.txt.
4. 4. Cite a primary source for the formula or methodology.
5. 5. Add DefinedTerm schema for the key concepts the calculator measures.

That's a 1-2 day setup for a strong Citation Readiness Score. The archetype carries most of the way.

What we honestly don't know yet

FermentCalc's real Citation Readiness Score — measured against the production calibration dataset — ships when Phase 1B's server-side crawler + polling engine lands. We'll publish the actual per-page score, the per-LLM citation rate across the 78 pages, and the citation-event timeline from real polling.

Subscribe via RSS or follow the changelog for the Phase 1B update where this case study gets its real data.