{"id":2349,"slug":"ramp-genuine-unknowns-base-rate-attribution-ai-discontinuity","title":"Genuine unknowns in ramp economics — unrecorded denominator of failed sites, structural attribution problem, 2026 AI-search discontinuity reducing predictive power of historical models","kind":"reference","scope":"business","status":"internal","audiences":["kevin","candid-team"],"topics":["seo-roi-payback-honest","ai-era-zero-click-impact"],"reference_body":"**Claim:** What is genuinely unknown — and why precise per-business payback cannot be forecast:\n\n- **The base rate is unpublished.** Published ROI case studies are **winners only**. The denominator — sites that were abandoned, underperformed, or never paid back — is essentially unrecorded. The \"compounding asset\" story is told by survivors. Any typical-ROI claim built on published cases overstates expected value by an unknown but material amount.\n\n- **Attribution is structurally hard.** Organic value arrives over multiple visits and touchpoints with no clean last-click; an evergreen page \"has no fixed return period\" ([[ramp-owned-vs-rented-attribution]]). This alone makes a precise payback number unidentifiable.\n\n- **The 2026 AI-search discontinuity.** AI search is actively changing organic click-through behaviour; the historical relationship between rank → click → revenue is in flux, so even good historical models have reduced predictive power going forward ([[ramp-ai-era-verification-older-studies-upper-bound]]).\n\n- **Forecastability of Stage 3.** Whether/when compounding materialises for a specific site is one of the unknowns; the six failure modes ([[ramp-when-organic-does-not-compound]]) cannot be predicted at launch.\n\n**Source:** Compass_artifact research document, June 2026.\n\n**Confidence:** High on enumeration. The honest framing is exactly that these ARE unknown.\n\n**Caveat:** When a vendor claims to have decoded any of these via a \"proprietary methodology,\" the appropriate response is skepticism, not curiosity. See [[rule-quarantine-vendor-roi-multiples]].","rationale_body":null,"metadata":null,"links":{"outgoing":[{"slug":"rule-quote-payback-as-range-never-point-estimate","title":"Rule: quote payback as a 6-24-month range with non-trivial failure probability — never as a point estimate","kind":"rule","scope":"business","link_type":"relates-to"},{"slug":"rule-quarantine-vendor-roi-multiples","title":"Rule: quarantine vendor ROI multiples (748% / 22:1 / similar) — never quote as typical results","kind":"rule","scope":"business","link_type":"relates-to"}],"incoming":[{"slug":"research-brief-new-site-ramp-economics-june-2026","title":"Research brief: the time dimension of a new website — ramp economics, the J-curve, owned vs rented, and the AI-era verification (June 2026)","kind":"reference","scope":"business","link_type":"relates-to"},{"slug":"ramp-owned-vs-rented-attribution","title":"Owned vs Rented (attribution) — organic: HIGH difficulty (multi-touch, multi-visit, no clean last-click, \"no fixed return period\"); paid: LOW (campaign-level CAC readily available)","kind":"reference","scope":"business","link_type":"relates-to"},{"slug":"ramp-contested-payback-precision-false","title":"Contested claim — \"Payback can be precisely forecast for a specific business.\" Verdict: FALSE. Too many compounding unknowns: vertical, competition, content, authority, conversion, LTV, churn, attribution, AI disruption","kind":"reference","scope":"business","link_type":"relates-to"}]},"created_at":"2026-06-25T16:11:00.113Z","updated_at":"2026-06-25T17:07:10.774Z"}