Reference / database site growth audit
Reference sites (databases, encyclopedias, finite datasets) live on long-tail SEO depth + LLM citation. Users find a specific fact, leave, return weeks later for another fact. Different funnel shape than SaaS — different audit calibration.
What's different for reference sites
- Acquisition: programmatic SEO is the primary lever. Each record = one indexable page. LLM-citation (llms.txt, schema, quote-ready section headings) is the secondary lever.
- Activation: “user viewed ≥3 data pages” = activated. Target ≥35%. Search-engine entry-point matters more than homepage hero.
- Engagement: internal linking hub-and-spoke + related-record suggestions. Time-on-page secondary to PV/session.
- Retention: D7 target ≥15%. Direct-type-in rate ≥12% indicates durable brand. Bookmark patterns + brand search are honest signals.
- Advocacy: citations on Wikipedia + Reddit + LinkedIn drive ~30% of long-term traffic for reference sites. k-factor matters less than for SaaS.
- Monetization: AdSense + Mediavine (at 10k sessions/mo) are primary. Affiliate where contextually relevant. Pricing tiers usually don't fit the model.
- Performance: static export mandatory. Sub-1s LCP feasible. AI-crawler-friendly (no JS-gated content).
Common reference-site growth-friction patterns
1. JS-rendered content invisible to AI crawlers
GPTBot, ClaudeBot, PerplexityBot don't execute JavaScript. If your records render client-side, LLMs cannot cite them. Static export or SSR with content in HTML at first paint is mandatory for LLM-citation acquisition.
2. Thin per-record content
A database of 5,000 records where each page has 80 words and a table = Google's helpful-content-update penalty + AdSense rejection. Each record needs ≥250 unique words of genuine analysis, not template boilerplate.
3. Article schema mismatch
Reference sites that use Article schema with recommendation-style headlines (“Best [X]”, “Top 10 [Y]”) without backing data trigger Google's structured-data-abuse flag. Use Dataset, DefinedTerm, or factual Article schemas with content that actually matches.
4. No internal linking depth
Records linked only from a hub page = page-rank stranded. Cross-link records to related records (e.g., “X also relates to Y, Z”). Internal-linking depth lifts both SEO and Engagement simultaneously.
Run your reference-site audit
60-second 7-factor wizard, reference-site benchmarks baked in.