calvincollins · xyz

The Forecast Desk · Research market

The Signal Layer

A viewer signs in on a laptop as an envelope passes through a prism and becomes a glowing token connected to a television and phone.
through 2029 5 priced outcomes Identity & Addressing
OpenIdentity & Addressingthrough 2029
38–48%

Most likely outcome

Messy Continuity — durable plural multi-signal, multi-graph stack

Identity in programmatic CTV — hashed emails, MAIDs, IP, ACR, clean rooms, and what survives to 2029

Read the full research →

01The Outcomes — 5 scenarios, priced by the research

Messy Continuity — durable plural multi-signal, multi-graph stack38–48%

No single currency, universal ID, or owner wins. Authenticated/first-party signals (HEM, login graphs, MVPD billing) harden into the spine; IP persists for DMA geo but is discounted as a household anchor; MAIDs fade on iOS, linger on Android; ACR scales for reach but stays siloed. UID2/EUID and RampID coexist below authenticated parity, capped by the FAST ceiling (~125M+ monthly unauthenticated). Consolidation (LiveRamp→Publicis, Oracle sunset) concentrates OWNERSHIP into holdco stacks that are more plural at the vendor layer yet NOT more interoperable; buyer routing-around spawns alternatives (ID5, ConnectID). Multi-currency measurement (Nielsen + challengers) persists; clean rooms become plumbing; regulation tightens incrementally. ABSORBS the 'decisive-but-re-papered' Salazar branch (narrow ruling + tokenization re-paper lands here, not as a reset).

How this number was derived
Reference-class M-of-N, re-weighted by cross-exam. Three CORRELATED continuity base rates: deprecations degrade-and-route-around ~4/5 (0.80); open-vs-platform concentration-without-enclosure ~4/4 (0.90); standards stay plural ~0.70. Because correlated (one adaptive-routing mechanism), take ~0.78–0.80 continuity NOT the naive product, then two forced subtractions: (i) driver-dominant garden-gravity + VPPA-in-wall pressure move mass to Walled-Garden Tilt; (ii) the SCOTUS 'decisive-but-re-papered' branch (~0.30 of the Salazar conditional) lands HERE, partly offsetting. Net 38–48%. GM-of-odds continuity cluster (base-rate .46, discontinuity .35, contrarian interop .29) = 0.36, consistent. Market check (Chrome ~2%, cookies-on YES) supports no-forced-break. Range overlaps Walled-Garden Tilt deliberately — the central surviving disagreement.

Drivers

  • Signal deprecations slip/degrade not break (cookie reversal Apr 2025; Sandbox retired Oct 2025; ATT routed-around)
  • FAST authentication ceiling cuts BOTH ways: caps UID2/HEM reach AND gardens' audited person-level reach on free inventory
  • Standards-vs-practice: OpenRTB EID-provenance/match-method/geo.type show minimal live bid-stream adoption mid-2026
  • Consolidation yields plural holdco stacks (IPG–Acxiom two-way precedent), not a commons and not a single winner
  • Multi-currency entrenched (Nielsen MRC but JIC-refusing; challengers JIC-certified)

Signals to watch

  • UID2/EUID authenticated-coverage share vs. FAST/AVOD share of impressions
  • Live bid-stream population of OpenRTB EID provenance / geo.type
  • Whether one TV currency wins JIC consolidation or 3–4 persist through 2027–28
  • Post-close LiveRamp non-Publicis retention; Omnicom Real ID / ID5 / ConnectID distribution by Q1 2028
Horizon · through 2029 (≈3.5-yr window)
Walled-Garden Tilt — gardens gain decisively (gradual grind, not a fast tip)28–38%

Inside continuity the balance tips harder to platforms. Amazon's authenticated graph (self-reported ~90%) + Amazon-Roku (~80M self-reported), Google's login graph, and Samsung/LG ACR estates capture a dominant majority of PREMIUM addressable CTV identity and measurement. AFAI/Fire-TV containment and interop friction starve open buyers of the best deterministic signals; UID2/RampID survive in long-tail and bridging roles. Independents keep folding in (Oracle, LiveRamp→Publicis). Retail-media clean rooms (AMC, Walmart) become the closed-loop measurement substrate. The contrarian's regulatory point cuts THIS way: VPPA pressure on OFF-platform sharing advantages IN-wall/clean-room activation, reinforcing the tilt. This is the GRADUAL grind; the fast network-cascade was defeated (no chokepoint; fragmented OS device IDs; FAST ceiling caps the enclosable pool).

How this number was derived
Base-rate + decomposition, deliberated NOT averaged. Outside view: first-party platforms reached a DOMINANT MAJORITY in ~2–3 of 4 prior digital eras ≈ 0.5–0.6 for a decisive tilt. CTV decomposition: P(gardens hold best authenticated signal | continuity) ≈ 0.85 × P(converts to DECISIVE share by 2029) ≈ 0.45 (raised: VPPA in-wall + IP collapse push budget into walls) × P(no countervailing interop/reg force) ≈ 0.75 = 0.287 ≈ 0.29. Driver-dominant Fermi lands 44–52%; base-rate/discontinuity ~26–29%; contrarian ~22–32%. REJECT the high end (double-counts continuity, rests on unaudited self-reported reach the reach-reckoning critique bites) and REJECT the fast cascade (no chokepoint, OS fragmentation, ceiling). Set 28–38%, overlapping Messy Continuity — co-modal, the gap is preserved disagreement. GM-of-odds garden cluster (.29/.48/.27/.26) = 0.32, upper-middle of range.

Drivers

  • First-party authenticated data is the strongest, most regulation-durable signal, activated in-wall at lower VPPA exposure (risk-inversion direction conceded — favors gardens)
  • Amazon-Roku ~80M is the largest footprint; AFAI containment raises the open-buy barrier
  • Consolidation base rate: independents fold into holdcos/platforms (Oracle, LiveRamp→Publicis)
  • Retail-media clean rooms concentrate CPG closed-loop measurement on platform-adjacent rails
  • IP-accuracy floor (~13% IP-to-household) degrades open probabilistic reach; gardens don't depend on it

Signals to watch

  • AUDITED (MRC/JIC) garden reach crossing ~85–90% — the reach-reckoning test
  • Share of CTV spend in login-tied DSPs (Amazon DSP, DV360) vs. open DSPs
  • A 2nd independent ID vendor absorbed by 2027–28 with no neutral alternative scaling
  • Whether VPPA litigation pushes publishers to pull eids/content fields from OPEN bid requests
Horizon · through 2029
Regulatory / Litigation Reset — a privacy step-change reshapes the stack10–16%

A discrete legal event resets the perimeter instead of incremental tightening: a BROAD/pro-plaintiff Salazar v. Paramount ruling (No. 25-459) making viewing-data sharing broadly actionable under VPPA with NO tokenization safe harbor, a federal opt-in law, or an enforcement wave (Disney $2.75M CCPA, FTC location orders, FTC Mobilewalla RTB ban) forcing GDPR-style consent flows. Identity contracts to consented first-party and clean-room activation; the broker/probabilistic layer is hit hardest. CORRECTED posture: the live case is Salazar v. PARAMOUNT — cert granted on the NARROW 6th-Cir. rule, NOT NBA v. Salazar (cert denied). A reset needs a BROAD reversal; the other tail (affirm narrow) RELIEVES the wave and feeds Continuity. The re-paper branch and the contrarian's 'inversion' mass were moved out (to Continuity and Garden), shrinking this bucket.

How this number was derived
Decomposition on the CORRECTED fork. The granted case reviews the NARROW 6th-Cir. rule, so reset needs a BROAD reversal: P(SCOTUS reverses broad, EXPANDING CTV exposure) ≈ 0.40–0.45 (coin-flip overturning a narrow rule, NOT the contrarian's 0.6 — refuted as backwards) × P(maximalist AND no safe harbor, i.e. resets not re-papers) ≈ 0.35–0.40 = 0.40×0.375=0.15 to 0.45×0.40=0.18. Federal-opt-in path ≈ 0.10–0.15. Combined ≈ 1−(1−0.16)(1−0.12) ≈ 0.26 IF the fork fires — then discounted hard for (a) the incremental-tightening base rate and (b) VPPA's first-order effect favoring in-wall activation (counted in Garden Tilt). Net 10–16%. GM-of-odds cluster 0.21 is inflated by the contrarian's 'inversion' number whose direction lost; stripping it leaves this near my range top.

Drivers

  • VPPA most-litigated CTV privacy law (250+ class actions 2024); Salazar v. Paramount a genuine binary fork ~early 2027
  • Active enforcement: Disney $2.75M CCPA; FTC location orders; FTC's first RTB-data ban (Mobilewalla)
  • Consent regimes elsewhere (GDPR) importable via ruling or statute
  • A maximalist 'consumer' read with NO safe harbor expands exposure for any off-platform viewing-data sharing

Signals to watch

  • Salazar SCOPE: affirm-narrow vs. broad-reversal, AND whether it blesses tokenized/aggregated sharing
  • A federal privacy bill, with or without an ad-tech safe harbor
  • Whether courts extend VPPA liability to DSP/SSP downstream recipients
  • State-AG/FTC settlements escalating into structural injunctions
Horizon · through 2029 (inflection at Salazar, ~early 2027)
Open-ID / Standards Consolidation — a universal ID and currency converge toward parity8–15%

The open coordination problem gets solved faster than base rate predicts. A dominant open universal ID (UID2/EUID, possibly RampID interop) reaches broad authenticated coverage AND a meaningful authenticated-FAST/clean-room bridge; OpenRTB EID-provenance/geo.type fields gain real bid-stream adoption; TV currency consolidates toward one or two JIC-blessed challengers displacing Nielsen as primary. The open layer reaches identity PARITY (not just inventory parity) with gardens. Cross-exam sharply bounded this: the contrarian's buyer-force argument yields plural VENDORS (Messy Continuity), not a converged interoperable STANDARD at transacted scale — 'more rails' ≠ verified emails on 125M FAST viewers — so it was reassigned away. The FAST authentication ceiling plus the OS-credential gap (OS owners don't pass email to channel publishers) is the hard cap keeping this a tail.

How this number was derived
Base-rate-led, capped by the binding constraint. Reference class = converging a fragmented ad ecosystem onto one neutral standard in ~3–4 yr (3p-ID replacement, header-bidding, prior universal-ID pushes): ~2–3 of ~12 reached genuine dominance in-window = 2/12=0.167 to 3/12=0.25 ≈ 15–25%. HARD CAP the base rate misses = FAST authentication ceiling + OS-credential gap → prior 12–20%. Cross-exam shaved the top: the buyer-force argument examined out to plural-vendors-not-converged-standard (reassigned to Continuity), and EID-provenance stays minimal in the live bid stream → 8–15%. GM-of-odds open-parity cluster (.135/.215/.095) = 0.14, in range. Distinguished from Messy Continuity by reaching PARITY/convergence, which the ceiling makes a tail.

Drivers

  • Near-universal US CTV publisher commitments to UID2 — latent coordination partly built IF authentication rises
  • Buy-side + JIC pressure for a non-Nielsen, identity-anchored, garden-independent census currency
  • Regulatory durability of consented IDs could push standardization onto them
  • Holdco scale post-LiveRamp/Publicis could fund interoperability IF commercially aligned (a weak 'if')

Signals to watch

  • UID2/EUID authenticated coverage crossing a majority of CTV impressions (binding: FAST ceiling)
  • A JIC currency winning primary-currency status, displacing Nielsen
  • Material rise in EID-provenance/geo.type bid-stream population from near-zero
  • Whether IAB Tech Lab assumes UID2 administration (decoupling from The Trade Desk)
Horizon · through 2029
Residual tail — exogenous shock, signal cliff, or non-modeled break4–9%

Acknowledged, not modeled as a trajectory: a forced Chrome/Google ad-tech divestiture in-window; an Apple/Google move removing a load-bearing CTV signal wholesale or defaulting IP obfuscation at scale; an Amazon graph reaching audited near-total lock-in; an IP-accuracy CLIFF (an FTC/state action reaching IP-DERIVED household inference, extending the location doctrine to IP, hitting the ~13% floor with no buffer and stranding IP-first graphs like Viant near-overnight); a binding cross-OS device-ID standard; a Blockgraph/MVPD consortium scaling into a third pole; or a UID2 bad-token/integrity failure at scale. Folds in the discontinuity member's Signal-Cliff list (Viant CORRECTED to MAID-agnostic, advantaged as MAIDs fall — its exposure is the IP forcing-event, not MAID drift) and the symmetric Open-Counter-Surge upside.

How this number was derived
Market-anchored residual, widened for the corpus-grounded cliff. Manifold 'divest Chrome by 2029' ≈ 2% YES (thin, play-money, 2026-06-19) anchors the highest-profile break path near zero; expired Polymarket <1% is STALE. Set 4–9% (above the ~2% single-path figure) for the broader unmodeled set: IP-cliff decomp P(IP forcing event by 2029)≈0.32 × P(cliff | event)≈0.4 (the 0.4 grounded in base rate #7 — the ~13% IP-to-household floor has NO buffer, so a forcing event tends toward a cliff not smooth degradation) ≈ 0.13 for IP-first graphs, but cross-exam established most of that mass ACCELERATES a repricing underway and flows into Reset/Garden, so only a slice lands here as a true exogenous shock; plus platform-policy and ID-integrity shocks and the symmetric upside. Tail discipline keeps it above the single-path figure. GM-of-odds residual cluster (.06/.085/.115) = 0.084, range top.

Drivers

  • Antitrust remedy uncertainty (Google/Chrome) — priced very low (~2% Manifold)
  • Platform-policy discontinuity (Apple/Google control load-bearing signals; OS-level IP obfuscation; Private Relay at scale)
  • IP-accuracy cliff: an FTC/state action reaching IP-derived household inference hits the ~13% floor with no buffer
  • Standards/security failure eroding trust in a dominant ID (UID2 bad-token incident at scale)

Signals to watch

  • A DOJ/court remedy ordering structural identity-asset divestiture with a 2026–29 timeline
  • Apple/Google removing a CTV identity signal, or IP-obfuscation default at scale
  • An FTC/AG order explicitly covering IP-derived household inference (Viant the named casualty)
  • A large-scale ID-integrity incident, or a binding cross-OS device-ID standard adopted
Horizon · through 2029 (tail)

Probabilities are the research's own scenario bands, priced as outcomes. The full argument — history, current state, drivers, and sources — lives in the corpus: read the Future Trajectory chapter.