Forecast scorecard · Edition 1 · October 2026
Forecasts we publish, then grade in public
Each forecast is logged with the indicators that would confirm or invalidate it and a resolve-by date. When it resolves we grade it here — hit, partial, or miss — including the misses. That is how you can tell whether our intelligence is worth acting on.
- Forecasts logged
- 12
- Open / hit / miss
- 12 / 0 / 0
- Graded so far
- First forecasts resolve January 2027
- Next resolve-by date
- Jan 31, 2027
| # | Horizon | Forecast | Confirms / invalidates | Resolve by | Status |
|---|---|---|---|---|---|
| F1 | 0–3m | US online holiday sales (Nov 1-Dec 31, 2026) grow >=6% YoY | Adobe/Salesforce holiday trackers; Census Q4 e-commerce report · Invalidated by: UMich sentiment <44 plus parcel network disruption; growth prints <6% confirming · checked Oct 20, 2026 — Deloitte finds 28% of holiday budgets spent by end of October (6-year high) and 42% planning October sales events, signaling robust early-season demand supportive of >=6% YoY growth. | Jan 31, 2027 | open |
| F2 | 0–3m | Effective Q4 2026 residential parcel cost inflation for DTC sellers lands >=8% despite 5.9% headline GRIs | Carrier surcharge tables; shipper invoice analyses · Invalidated by: Surcharge rollbacks or peak-fee waivers confirming · checked Oct 20, 2026 — Diesel at record $6.529/gallon with surcharges not fully offsetting costs; Walmart's carrier expansion adds options but doesn't reduce underlying cost inflation pressure. | Jan 31, 2027 | open |
| F3 | 0–3m | AI-referred retail traffic at least doubles YoY during Cyber Week 2026 and its conversion premium stays >=25% | Adobe Analytics holiday telemetry · Invalidated by: Premium compresses below 25% as volume scales; traffic growth <100% confirming · checked Oct 20, 2026 — AI-attributed orders grew 105% from BFCM 2025 to July 2026 per Tamebay profitability piece; directionally supports doubling threshold, though Cyber Week-specific conversion premium data absent. | Jan 31, 2027 | open |
| F4 | 0–3m | TikTok Shop US full-year 2026 GMV >= $20B | Momentum Works / eMarketer full-year updates · Invalidated by: Post-divestiture operational disruption; estimates land <$20B confirming · checked Oct 20, 2026 — TikTok Shop BFCM campaign runs Nov 11–30 with Best Buy, Ulta, Sephora joining; 2025 BFCM drove $500M+ with 84% YoY growth, supporting trajectory toward $20B annual GMV. | Jan 31, 2027 | open |
| F5 | 3–6m | Amazon 2027 US fee announcement contains no headline referral increase but raises net seller cost via surcharge/restructure layering | Seller Central fee notice (expected Dec 2026-Jan 2027) · Invalidated by: A flat, surcharge-free 2027 schedule or headline referral increase | Apr 30, 2027end of horizon | open |
| F6 | 3–6m | Walmart extends fee-cut or WFS-incentive playbook to additional categories to sustain >40% marketplace growth | Seller Center policy updates; Q3/Q4 FY27 earnings · Invalidated by: Walmart raises marketplace fees instead confirming · checked Oct 20, 2026 — Walmart added 8 approved carriers Oct 13 to reduce friction for seller-fulfilled orders, consistent with Walmart sustaining marketplace growth incentives rather than raising fees. | Apr 30, 2027end of horizon | open |
| F7 | 3–6m | At least one additional top-10 US marketplace (excluding Amazon) signs a formal AI-agent access integration | Platform announcements (UCP, ACP successor, or equivalent) · Invalidated by: Industry-wide agent blocking following Amazon precedent confirming · checked Oct 20, 2026 — Meta Muse and OpenAI Dots now live with Walmart, Best Buy, Sephora, Wayfair as connector partners — multiple top-10 US marketplaces have signed formal AI-agent integrations. | Apr 30, 2027end of horizon | open |
| F8 | 6–12m | US average effective tariff rate moves >=2 points in either direction from ~10.9% within 12 months | Yale Budget Lab effective-rate series; Federal Register actions · Invalidated by: Twelve months of rate stability within +/-2 points | Oct 14, 2027 | open |
| F9 | 6–12m | Used/refurbished smartphone unit growth outpaces new-unit growth in full-year 2026 data | IDC / Counterpoint annual used-smartphone reports · Invalidated by: New-unit resurgence on an upgrade supercycle | Oct 31, 2027end of horizon | open |
| F10 | 12–24m | De minimis is not restored before the statutory repeal (July 2027) takes effect | Federal Register / CBP rulemaking record · Invalidated by: Congressional reversal restoring the exemption | Jul 1, 2027 | open |
| F11 | 12–24m | Consolidated multichannel operation (one console + one fulfillment network serving >=3 channels) becomes the default stack for new multichannel sellers, with Amazon a top-2 provider | Seller surveys; MCF adoption disclosures; console market share data · Invalidated by: Antitrust action or broad seller refusal of cross-channel data sharing | Oct 31, 2028end of horizon | open |
| F12 | 12–24m | Agentic commerce exceeds 2% of US e-commerce sessions by October 2027 | Adobe / Contentsquare session-share measurements · Invalidated by: Platform-wide agent blocking or stalled consumer adoption confirming · checked Oct 20, 2026 — Meta Muse and OpenAI Dots are live in US/Canada with retail connectors; Google Universal Cart adds another agentic-style aggregation layer — adoption infrastructure is accelerating ahead of the 2027 target. | Oct 31, 2027 | open |
From the analyst’s desk
Daily agent run · Oct 20, 2026 · 12 articles reviewed
Today's coverage converges on three structural pressures for Q4 2026 operators. First, the distribution landscape is fragmenting fast: Google Universal Cart, Meta Muse, OpenAI Dots, and TikTok Shop are simultaneously pulling purchase intent off the open web, making product-data completeness and AI-agent connector status existential rather than optional. Second, logistics cost inflation is compounding — Trans-Pacific on-time delivery at 26.8% and diesel at a record $6.529/gallon both point toward higher landed costs and tighter margins precisely when volumes peak. Third, early-season budget pull-forward (28% of holiday spend by end of October, a 6-year high) means operators who aren't in-market now are already losing wallet share. Walmart's carrier-list expansion is a modest relief valve on last-mile flexibility. The through-line: sellers need margin protection, channel diversification, and clean product data all at once, entering the highest-stakes quarter in recent memory.
Google Universal Cart adoption curve
Aggregating Walmart, Target, Shopify sellers into one Google checkout could rapidly shift where conversion happens, with major implications for attribution and platform fee structures.
AI agent connector partner lists
Meta Muse and OpenAI Dots are live now; which merchants gain or miss connector status will determine agent-driven GMV distribution in the critical holiday window.
Trans-Pacific reliability / year-end capacity crunch
On-time delivery at 26.8% with tightening capacity into Q4 raises stockout and markdown risk for import-dependent sellers at the worst possible time.
Diesel / freight rate trajectory
Record $6.529/gallon diesel with unrecovered surcharges signals further freight cost pass-through to sellers, directly pressuring forecast id 2 on effective parcel cost inflation.
TikTok Shop BFCM Nov 11–30 campaign
First appearance of Best Buy, Ulta, and Sephora on TikTok Shop BFCM, combined with $500M+ 2025 base, is the clearest near-term data point for the $20B GMV forecast.
Written by our analyst agent, which runs daily over the newly classified coverage with its own persistent memory, checks every open forecast’s indicators against the evidence, and corrects its earlier conclusions when the data turns. Its indicator reads appear on the ledger above.
Methodology
- Forecasts are evidence-based judgments, not model outputs. Each one is grounded in the cited evidence in the Intelligence Dossier.
- Resolution is fixed in advance. Each forecast names the indicators that confirm or invalidate it. Where the forecast does not name a date, the resolve-by date is the end of its stated horizon from the October 2026 edition.
- Scoring. Each resolved forecast receives one of four grades — hit, partial, miss, or unresolvable — and the scorecard shows the running count of each. Only a fully correct call is a hit: a partially correct call is marked partial, and a forecast whose measuring source stops publishing is marked unresolvable rather than quietly dropped. The first forecasts resolve in January 2027.
- Indicators are checked by an agent, daily. An analyst agent reviews each open forecast’s confirming and invalidating indicators against the day’s classified coverage and posts its read on the ledger. When a forecast resolves, it writes a post-mortem to its memory, and future runs — and future forecasts — are made with those post-mortems in context.
- Nothing is deleted. Misses stay on the ledger. New editions add forecasts; they do not rewrite old ones.
The first grades arrive when the holiday forecasts (F1–F4) resolve in January 2027.