Candidate markets — the waiting list

Liquidation · proposed group: Tier 2 · Adjacent: Heavy equipment auctions; Dealer used-car auctions; 'Surplus laptops' — as a unified supply source it is new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 3 · 3 · 3 · 3 · 3 · 1 · 3 · 3 → structural 92.6, gates Y/Y/Y/Y, actionable 92.6
• Reject reasons: Attention · Capacity · Operational fit
• Access: GovDeals free registration, 12.5% premium, pay in 5 days, remove in 7–10 days; GSA no premium, pay 2 days, remove 10 days, default = 20%/$200; GovPlanet 15% fee, DEMIL items need End Use Certificate. Also Public Surplus and Municibid (~50k+ listings across fleet, IT, lab, athletic, misc.); treat the municipal/institutional seller set as one reject menu.
• Scale: GovDeals 15,000–25,000 active lots/day, 900+ categories, 4.5M+ auctions completed; GovPlanet ~3.5k gov items + 7.4k trucks/trailers.
• Sales tape: Win → pay → pickup (deadline) → title/transport → resale over weeks → storage/abandonment fees
• Consumption sink: Durable (score 1).
• Primary risk / boundary: As-is, photo-only; hard removal deadlines and pickup logistics decide margin; titling.
• First experiment: Use BidProwl-class archives + a reseller's bid log (won/lost) with pickup cost; score lots within 300 miles; measure abstention on title-issue lots.
• Confidence: high
• Sources: https://featured.govdeals.com/by-the-numbers · https://bidprowl.com/guides/govdeals-faq · https://govauctions.app/guides/gsa-auctions-faq · https://www.govplanet.com/buyer-faq

Perishables · proposed group: Industrial · New — no floral entry in the 135

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 3 · 2 · 3 · 3 · 2 · 3 · 3 · 3 → structural 92.6, gates Y/Y/Y/Y, actionable 88.0
• Reject reasons: Attention · Time · Capacity
• Access: Register as RFH customer; screened; expected turnover ≥€5k/week (~€250k/yr); collateral or eWallet prepay; Floriday access; remote clock buying (KOA).
• Scale: 2025 product turnover €5.4B; clock >€2B on ~5bn stems; 18,930 distinct products auctioned; >100,000 transactions/day; 35 clocks; ~6,000 growers; 4 hubs.
• Sales tape: Clock purchase → cold-chain dispatch → customer sale (florist/exporter) → returns/credit → destruction of unsold stems
• Consumption sink: Flowers die in days; unsold volume is destroyed (2020: ~20%).
• Primary risk / boundary: Turnover minimum + Dutch entity for a new operator; unsold/'doorgedraaid' rates not published; vase-life and downstream sell-through hidden at purchase.
• First experiment: Log 4 weeks of clock lots for 200 products incl. passes; shadow-score against a wholesaler's realised sales; measure predicted vs realised margin by product/grade.
• Confidence: high
• Sources: https://www.royalfloraholland.com/en/buying-2/marketplace/become-a-customer · https://www.floraldaily.com/article/9801629/nl-almost-eur5-billion-in-flowers-and-plants-traded-via-the-clock-in-2025/ · https://www.royalfloraholland.com/en/news-2026/week-13/royal-floraholland-succesvol-2025-productomzet-groeit-naar-5-4-miljard-euro

Commerce · proposed group: Core · Adjacent: Amazon wholesale; E1 'Overlooked supplier catalogue' — the small-retailer, marketplace-supply version

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 3 · 3 · 3 · 3 · 2 · 3 · 2 · 3 → structural 85.2, gates Y/Y/Y/Y, actionable 85.2
• Reject reasons: Attention · Prediction · Scale
• Access: Retailer signs up with business info; net-60 for eligible retailers; free returns on first order from any new brand; brand pays 15% + $10 first-order fee (Faire). Ankorstore: 60-day terms, ~€100 minimums.
• Scale: Faire: 100,000+ brands; Ankorstore FY2025: 30,000+ brands, 300,000+ retailers, €1.2B GMV, €160 AOV, ~45% repeat.
• Sales tape: Order → receipt → store sell-through over weeks–months → markdown/return (first order only)
• Consumption sink: Consumer goods consumed.
• Primary risk / boundary: Sell-through only visible in own POS; taste-driven demand; catalogue snapshots needed for declined set; net-60 hides cash cost.
• First experiment: Independent retailer: 18 months of Faire orders + weekly catalogue snapshots + POS sell-through; shadow-score reorders and first-time brands; measure abstention on slow categories.
• Confidence: medium-high
• Sources: https://www.faire.com/how-faire-works · https://garnetmarketplace.com/marketplaces/ankorstore.html

Equipment · proposed group: Tier 2 · Adjacent: Office clearance lots; Industrial surplus equipment — distinct vertical

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 3 · 3 · 3 · 3 · 1 · 2 · 3 → structural 85.2, gates Y/Y/Y/Y, actionable 85.2
• Reject reasons: Capacity · Operational fit · Attention
• Access: Open bidder registration (ID + card); buyer's premium 10–18% (15% verified at Restaurant Auction Company); fixed pickup days (e.g., 3 days × 6 h); buyer removes/rigs.
• Scale: BidSpotter snapshot Sep 2026: 1,853 restaurant/bar lots live; US restaurant closures 886–1,650/month (2025); 2026 chain closures ~1,000 locations.
• Sales tape: Win → removal (disconnect, rigging) → test/clean → resale (eBay/local) over 2–8 weeks → scrap
• Consumption sink: Durable; wears (score 1).
• Primary risk / boundary: 'Working when removed'; refrigerant/gas faults; removal deadlines; geographic clustering (WA/GA/FL/CA).
• First experiment: Used-equipment dealer's 12 months of bids (won/lost) + removal cost + resale; score lots; measure abstention on far-away/heavy lots.
• Confidence: medium
• Sources: https://www.bidspotter.com/en-us/for-sale/food-and-beverage-equipment/restaurant-and-bar-equipment · https://restaurantauctioncompany.com/faqs/ · https://datassential.com/resource/restaurant-closures-2025-trends-data/

Commerce · proposed group: Core · Adjacent: Amazon wholesale / online arbitrage; Walmart and marketplace wholesale (domestic) — cross-border is new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 3 · 2 · 3 · 3 · 2 · 3 · 2 · 3 → structural 88.9, gates Y/Y/Y/Y, actionable 84.5
• Reject reasons: Complexity · Attention · Time
• Access: MercadoLibre Global Selling: no local entity, one account → MX/BR/CL/CO/AR, Full fulfilment from Texas, biweekly payouts; noon: trade licence + VAT; Allegro export 0% commission on new export offers. Ozon/Wildberries: sanctions exposure — exclude.
• Scale: Whole catalogues per country; ML Full storage $0.0005–$0.024/unit/day; category counts unpublished.
• Sales tape: Source purchase → inbound to ML Full / cross-border parcel → sale → returns/customs delays → biweekly payout
• Consumption sink: Consumer goods consumed.
• Primary risk / boundary: Duties/VAT and returns are the hidden costs; FX; marketplace rule volatility; Russian platforms off-limits for a US entity.
• First experiment: Reuse the Amazon OA pipeline: snapshot ML/Allegro catalogues weekly for 3 months, ship 200 SKUs, replay landed margin; measure abstention vs duty surprises.
• Confidence: medium
• Sources: https://global-selling.mercadolibre.com/landing/pricing · https://helpcenter.noon.partners/en/category/onboarding-and-registration/documents-required-to-sell-on-noon · https://help.allegro.com/en/marketplaces

Electronics · proposed group: Tier 2 · Specifies: Refurbished electronics — the auction-lot, best-instrumented version

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 3 · 2 · 3 · 3 · 3 · 2 · 3 · 3 → structural 88.9, gates Y/Y/Y/Y, actionable 84.5
• Reject reasons: Prediction · Complexity · Time
• Access: B-Stock mobile storefronts: US-only, state reseller certificate, $500 deposit, screening up to 2 months, R2 certification for non-functional lots; pay within 2 business days; buyer ships. Back Market: seller onboarding/quality charter.
• Scale: ~1,000 mobile auctions closing per week on B-Stock; 5M+ phones/yr; Mobile Carrier ~500 lots/wk (grades AA+–RR); hundreds of model × storage × carrier SKUs.
• Sales tape: Win → receive → test/grade → activation-lock & yield → refurb → resale over 2–6 weeks → returns
• Consumption sink: Depreciating; used up over its life (score 2).
• Primary risk / boundary: Functional yield, locks, battery health only after receipt; weekly price decay; reseller-cert/R2 gate.
• First experiment: Refurbisher's 12 months of B-Stock bids (won and lost) + yield/resale ledger; score lots; measure predicted vs realised recovery by grade.
• Confidence: high (B-Stock)
• Sources: https://bstock.com/blog/buyers-guide-mobile-storefronts-on-bstock/ · https://bstock.com/mobilecarrier/faq/ · https://sustainableelectronics.org/knowledge-base/ctia-grading-qa/

Perishables · proposed group: Industrial · Adjacent: Agricultural surplus and off-grade lots; Short-dated food closeouts (both cover off-spec — this is primary spot wholesale)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 3 · 2 · 3 · 3 · 3 · 3 · 3 · 3 → structural 96.3, gates Y/Y/Y/P, actionable 82.3
• Reject reasons: Attention · Time · Operational fit
• Access: PACA licence for >2,000 lb/day ($995/yr); Hunts Point entry fee; buy full cases/pallets; negotiated spot pricing per merchant, no central order book, no API.
• Scale: Hunts Point: ~47 merchants, $2–2.3B/yr, ~23,000 restaurant customers, ~9% of US produce; ~13 USDA-reported terminal markets; daily lot counts not published.
• Sales tape: Purchase → cold storage → restaurant/retail sale (hours–days) → rejections/credits → shrink write-off
• Consumption sink: Eaten or spoiled within days; retail loss 4–43% by fruit (USDA ERS).
• Primary risk / boundary: Declined offers must be captured by the agent (no electronic book); shrink attribution per lot; PACA prompt-pay trust; hours-scale cadence stresses fit latency.
• First experiment: Instrument one wholesaler's daily buys + passed offers for 30 days against USDA terminal prices; replay shrink-adjusted margin; test abstention on volatile commodities.
• Confidence: medium
• Sources: https://www.ams.usda.gov/rules-regulations/paca/licensing · https://mymarketnews.ams.usda.gov/mymarketnews-api · https://www.nycfoodpolicy.org/hunts-point-distribution-center-brief-overview-spotlight-produce-market/

Commerce · proposed group: Tier 2 · New — refusal-first market (accept/decline per item offered)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 3 · 3 · 3 · 3 · 3 · 1 · 2 · 3 → structural 81.5, gates Y/Y/Y/Y, actionable 81.5
• Reject reasons: Attention · Prediction · Capacity
• Access: Open; TRR commission tiers (~40–85% to consignor by price band), Fashionphile quote in 3–5 days, buyout 70% (85% >$3k); local shops set own rules.
• Scale: TRR FY2025: consignment revenue $535.9M of $693M; 1.056M active buyers; take rate 36.5% (Q4'25).
• Sales tape: Intake → authentication/regrade → listing → sale → returns → consignor payout (day 15–21) → unsold return after ~90 days
• Consumption sink: Wears (score 1).
• Primary risk / boundary: Shelf capacity is the binding constraint; regrade after receipt; low consumption.
• First experiment: Consignment shop: 24 months of intake decisions incl. declines + sales; shadow-score intake; measure inventory-turn gain.
• Confidence: medium-high
• Sources: https://www.globenewswire.com/news-release/2026/02/26/3246067/0/en/the-realreal-announces-fourth-quarter-and-full-year-2025-results.html · https://www.underpriced.app/blog/the-realreal-consignment-guide-2026 · https://www.thriftandtell.com/guides/fashionphileselling

Electronics · proposed group: Tier 2 · Specifies: Refurbished electronics / 'Surplus laptops and IT gear' (landing page)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 3 · 3 · 3 · 3 · 3 · 1 · 2 · 3 → structural 85.2, gates Y/Y/Y/Y, actionable 80.9
• Reject reasons: Complexity · Attention · Capacity
• Access: GovDeals/AllSurplus/Liquidation.com free registration; enterprise ITAD contracts need R2v3/e-Stewards; buyers inherit data-bearing-media liability.
• Scale: ITAD market $28.3B (2025) → $31.9B (2026); Liquidity Services 500+ categories; servers hold value 3–5 yrs from release.
• Sales tape: Win → sanitisation (NIST 800-88) → component test → resale (weeks) → yield write-offs
• Consumption sink: Durable; obsolescence (score 1).
• Primary risk / boundary: Component yields, BMC/BIOS locks, missing transceivers; data-sanitisation obligations.
• First experiment: ITAD reseller's 12 months of lot bids (won/lost) + yield + resale; score lots by class/age.
• Confidence: medium
• Sources: https://www.grandviewresearch.com/industry-analysis/it-asset-disposition-market · https://liquidtechnology.net/how-much-is-used-it-equipment-worth/ · https://www.sktes.com/news/what-is-r2v3

Industrial remnants · proposed group: Industrial · New — distinct from recovered-material bales (this is unused stock)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 3 · 3 · 3 · 2 · 3 · 2 · 3 → structural 88.9, gates Y/Y/Y/P, actionable 80.0
• Reject reasons: Scale · Complexity · Attention
• Access: Go2Paper: free registration, sellers pay 5% ($100 min), buyers free, reverse auctions/RFPs; PaperIndex: lead-gen directory (64,046 members claimed), no transactions handled.
• Scale: No listing/volume counts published; lot structure (basis weight × brightness × width × core) unverified on the help pages.
• Sales tape: Buy → jumbo-roll freight → slit/convert → sale to printer → off-spec claims
• Consumption sink: Printed on — consumed (score 3).
• Primary risk / boundary: Spec matching is the business; freight on jumbo rolls; counterparty and quality risk off-platform; slow RFQ cycles on PaperIndex.
• First experiment: Paper merchant's 12 months of stocklot offers (taken/passed) with conversion cost and sale price.
• Confidence: medium (fees) / low (lots, prices)
• Sources: https://www.go2paper.com/help/help.asp · https://www.paperindex.com/faq

Commerce — venture · proposed group: Core · Adjacent: Amazon wholesale / OA — launch selection (no supplier approval) is a different decision

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 3 · 3 · 3 · 3 · 3 · 3 · 1 · 3 → structural 92.6, gates Y/Y/Y/P, actionable 79.2
• Reject reasons: Prediction · Complexity · Capacity
• Access: Amazon Professional plan + identity verification; Brand Registry needs a registered/pending trademark; Alibaba/1688 suppliers; inspection agencies $300–1,000.
• Scale: Unbounded niche menu (Helium 10 / Jungle Scout); Jungle Scout 2025 survey: 40% cite rising costs, 30% fewer sourcing from China; de minimis ended May/Aug 2025.
• Sales tape: PO → inspection → freight → FBA → sales/PPC → reviews → returns → reorder or liquidation
• Consumption sink: Venture-class: the launch lives or dies; goods consumed.
• Primary risk / boundary: Data floor: $4–12k per observed launch → ≥100 observed launches ≈ $0.5–1M of history for one operator; 3–6 month cycles; tariff drift; platform suspension risk.
• First experiment: Pool anonymised launch ledgers from several aggregators/agencies (enterprise); test if P34 abstains on launches that later died; membership version = only for operators with 100+ past launches.
• Confidence: medium (costs/tariffs high; failure rates low)
• Sources: https://sell.amazon.com/blog/brand-registry-requirements · https://www.unicargo.com/de-minimis-ended-2026-tariffs-guide-amazon-fba/ · https://www.junglescout.com/resources/reports/amazon-seller-report-2025/

Perishables · proposed group: Industrial · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 2 · 2 · 2 · 2 · 3 · 3 · 3 → structural 81.5, gates Y/Y/Y/Y, actionable 77.4
• Reject reasons: Time · Operational fit · Capacity
• Access: Bank guarantee lodged with each physical auction → buyer number; paid Pefa Auction Clock subscription (free trial); buyer arranges transport. US: Portland Fish Exchange registered-buyer display auction with remote e-bidding.
• Scale: Dutch first sale 2025: 26,682 t / €145M across 12 auctions (7 on Pefa); dozens of species × size × grade per port; lots/day not published.
• Sales tape: Clock buy → same-day cold-chain → processor/retail sale → yield loss/shrink → claims
• Consumption sink: Eaten within days; unsold fish is discarded.
• Primary risk / boundary: Per-auction bank guarantees; lot counts unverified; yield/shrink only after processing; same-day cadence.
• First experiment: Subscribe to one Pefa clock; record every lot and pass for 6 weeks; shadow-score against a fishmonger's realised resale; measure abstention value on gluts.
• Confidence: medium
• Sources: https://www.pefa.com/how-to-buy-fish-with-the-pefa-auction-network/ · https://fishery-aquaculture-market-observatory.ec.europa.eu/en/country-and-species-profiles/country-profiles/netherlands · https://www.pfex.org/auction/

Electronics · proposed group: Industrial · Adjacent: Spare-parts stocking; Industrial surplus equipment — MPN brokerage is distinct

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 3 · 3 · 3 · 3 · 3 · 3 · 2 · 2 → structural 85.2, gates Y/Y/Y/P, actionable 76.7
• Reject reasons: Attention · Prediction · Complexity
• Access: Open; ERAI membership for counterfeit screening; Sourcengine (1B+ parts, 3,500+ suppliers, Part/Offer/Order APIs), Octopart/Nexar API, NetComponents/Broker Forum subscriptions.
• Scale: 1B+ MPNs; 621,909 components EOL'd in 2025 (52% without PCN); 2026 shortage: DRAM contract +90–95% QoQ Q1 2026, FPGA lead times 40 wks, MCU up to 55 wks.
• Sales tape: Buy → AS6081 test (days) → sale to OEM/EMS → returns/nonconformance → write-off of unsold reels
• Consumption sink: Soldered into products; obsolescence ends the market for the part.
• Primary risk / boundary: Counterfeit/nonconforming risk (ERAI 748 reports 2025; 24% pass electrical-only tests); shortage cycles are multi-quarter regimes; spreads unpublished.
• First experiment: Broker's 24 months of RFQs, quotes, buys and sales; join test results; score buy/pass per line under shortage/normal regimes with lifecycle features.
• Confidence: medium-high
• Sources: https://www.sourcengine.com/api-integrations · https://nexar.com/api · https://www.dasenic.com/blog/knowledgeHub/obsolete-components-in-2026-eol-list-shortage-statistics-what-to-do-about-it · https://www.oxebridge.com/emma/erai-report-counterfeit-electronics-on-the-decline-despite-increase-in-sales/

Equipment · proposed group: Tier 2 · Adjacent: Industrial surplus equipment

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 3 · 3 · 3 · 3 · 1 · 2 · 3 → structural 85.2, gates Y/Y/Y/P, actionable 76.7
• Reject reasons: Complexity · Attention · Capacity
• Access: Free accounts (LabX, EquipNet, BioSurplus); Heritage Global biotech/pharma auctions on Bidspotter; buyer's premium 5–20%; decon certificates. University lots also on Public Surplus.
• Scale: Listing counts not published; HGP ~20+ biotech auctions Jan–Jul 2026; university surplus via Public Surplus/GovDeals.
• Sales tape: Win → rigging/decon → test → resale (weeks) → OEM licence transfer issues
• Consumption sink: Durable; wears (score 1).
• Primary risk / boundary: Scale unverified; software licences and calibration unknown until powered; export controls on some analytical gear.
• First experiment: Lab-equipment dealer's auction bids (won/lost) + test outcomes + resale for 12 months; score by instrument class.
• Confidence: low-medium
• Sources: https://www.hgpauction.com/biotech-pharmaceutical-auctions/ · https://www.labx.com/ · https://www.excedr.com/blog/laboratory-equipment-auctions

Liquidation · proposed group: Tier 2 · Adjacent: Liquidation and B-stock pallets; Returned-goods resale — insurer supply is distinct

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 2 · 3 · 3 · 3 · 2 · 2 · 3 → structural 85.2, gates Y/Y/Y/P, actionable 76.7
• Reject reasons: Complexity · Attention · Capacity
• Access: Salvex open signup; Fr8Auctions wholesale buyers only (business licence, in-person orientation, approval not guaranteed); Liquidity Services marketplaces free registration.
• Scale: Salvex ~5k live lots (heavily industrial/aviation); Liquidity Services FY2025 GMV $1.57B, 6.0M registered buyers; Fr8 weekly.
• Sales tape: Win → freight → inspection/sort → resale by channel → destruction of de-labelled goods
• Consumption sink: Consumed downstream (score 2).
• Primary risk / boundary: Damage extent hidden; control-of-damaged-goods clauses (brands destroy instead of salvage); Fr8 gate.
• First experiment: Salvage buyer's 12 months of bids (won/lost, bid counts) + recovery ledger; score lots by damage type.
• Confidence: medium
• Sources: https://www.salvex.com/ · https://fr8auctions.com/registration/ · https://liquidityservices.com/press-releases/liquidity-services-announces-fourth-quarter-fiscal-year-2025-0 · https://www.munichre.com/en/insights/mobility-and-transport/high-price-brand-protection.html

Industrial remnants · proposed group: Industrial · Adjacent: Deadstock and branded apparel lots; Industrial by-products — remnant rolls into small jobs is new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 3 · 3 · 3 · 2 · 3 · 2 · 3 → structural 85.2, gates Y/Y/Y/P, actionable 76.7
• Reject reasons: Scale · Complexity · Operational fit
• Access: Textiles: jobber channel (NYC/LA districts, DeadstockLA, Fabscrap wholesale) buying end-of-roll, cancellations and overruns; 1–10 yd minimums vs 50–500 yd at the mill. Film remnants: converter-to-converter and scrap/regrind brokers — no dedicated marketplace verified.
• Scale: No published volumes; jobber deadstock $5–15/yd at 50–70% below mill price (Style3D, marketing source).
• Sales tape: Buy → LTL → cut/convert or resell → yardage shortfall, defects, dye-lot disputes
• Consumption sink: Cut into garments / converted into packaging.
• Primary risk / boundary: Thin comps; freight can exceed remnant value; actual yardage, flaws and (for film) treatment decay hidden until receipt; film-remnant side unverified.
• First experiment: Jobber's 12 months of lots offered/taken and sell-through by spec; test spec-match features vs realised margin.
• Confidence: low
• Sources: https://www.style3d.ai/blog/what-are-fabric-jobbers-and-how-do-they-benefit-fashion-designers/

Perishables · proposed group: Tier 2 · Adjacent: Rare plants and seeds (collectibles) — this is trade wholesale

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 2 · 3 · 3 · 3 · 3 · 2 · 3 → structural 88.9, gates Y/Y/Y/P, actionable 76.0
• Reject reasons: Attention · Complexity · Operational fit
• Access: Resale certificate; state nursery-dealer licence in some states (CA/FL); LandscapeHub free for buyers (300+ suppliers, one cart); PlantANT availability aggregator.
• Scale: USDA 2019 census: $13.8B, 20,655 operations; nursery stock $4.55B; PlantANT claims 613,182 listings / 36,871 varieties; CA+FL+OR = 66% of sales.
• Sales tape: Purchase → freight → hold in yard → job installation or retail sale → mortality/shrink write-off
• Consumption sink: Planted (leaves the market) or dies in the yard; shrink models 2.5–15%.
• Primary risk / boundary: Seasonality and weather; mortality attribution; phytosanitary quarantines; no central price archive.
• First experiment: Aggregate 20 growers' availability lists weekly for a season; join a landscaper's job pipeline and yard shrink; score buy-vs-pass per line.
• Confidence: medium
• Sources: https://www.nass.usda.gov/Newsroom/archive/2020/12-08-2020.php · https://blog.landscapehub.com/how-to-source-nursery-products-wholesale-without-the-hassle · https://www.plantant.com/

Energy feedstock · proposed group: Industrial · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 2 · 2 · 2 · 2 · 3 · 2 · 3 → structural 77.8, gates Y/Y/Y/Y, actionable 73.9
• Reject reasons: Operational fit · Scale · Time
• Access: BALTPOOL (Lithuania): registered participants trade weekly auctions with distance-constrained matching; KYC/registration per BALTPOOL docs; participants from LT/LV/EE/PL/SE/FI. US: OTC, relationship-driven.
• Scale: 2020–21 season: 4M MWh traded, 453 registered participants (117 international) — dated; 2025 volumes unverified. Avg supply prices then LT €10.89/MWh, LV €12.15, EE €12.90.
• Sales tape: Award → delivery → moisture/ash measurement → settlement on delivered energy → penalties
• Consumption sink: Burned (score 3).
• Primary risk / boundary: Moisture and haul dominate value; EU sustainability certification (RED III/SBP); auction results public incl. unmatched bids — good declined-option record.
• First experiment: Replay BALTPOOL auction history for a supplier: bids won/lost vs delivered-energy settlement; test abstention on long-haul lots.
• Confidence: medium (2021-vintage scale)
• Sources: https://www.svebio.se/en/press/pressmeddelanden/increased-volumes-traded-international-biomass-exchange/ · https://www.baltpool.eu/en/biomass-exchange/how-to-trade/

Logistics · proposed group: Capacity · Adjacent: Truckload boards; Backhaul matching (this is per-vehicle, multi-car routing)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 3 · 2 · 3 · 3 · 2 · 1 · 3 · 3 → structural 85.2, gates Y/Y/Y/P, actionable 72.8
• Reject reasons: Operational fit · Time · Capacity
• Access: Carrier: DOT + MC authority, liability ≥$750k (brokers want $1M), cargo $100–250k; subscription ~$115–135/mo; broker authority + BMC-84 bond.
• Scale: Central Dispatch: 10,000,000+ vehicles posted/yr (~27k/day), 20,000+ carriers; cancellations +24.8% H1-26, 4.64% of orders cancelled for inadequate rate.
• Sales tape: Accept → pickup → delivery (BOL inspection) → damage claims → payment (net terms/quick-pay)
• Consumption sink: Truck-hours perish; the vehicle itself is not consumed (score 1).
• Primary risk / boundary: Rolling menu with interference (paper: rapidly expiring menus); declined loads must be logged by the dispatcher agent; damage-claim rates not published.
• First experiment: Carrier with 3–9 car trailers: 6 months of board snapshots + accepted/declined loads + realised margin per trip; score load sets under routing constraints.
• Confidence: medium
• Sources: https://www.centraldispatch.com/ · https://www.ccjdigital.com/economic-trends/freight-demand/article/15834395/how-much-does-it-cost-to-ship-a-car-in-2026-rates-trends · https://superdispatch.com/blog/carrier-insurance/

Logistics capacity · proposed group: Capacity · New — distinct from Ocean freight capacity and Container repositioning

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 2 · 3 · 2 · 2 · 3 · 3 · 3 → structural 85.2, gates Y/Y/Y/P, actionable 72.8
• Reject reasons: Operational fit · Time · Capacity
• Access: Shippers/brokers quote via barge lines or OpenTug (multi-vendor quotes on predetermined routes); owning the fleet is Jones Act-gated; shipper-side quoting is open.
• Scale: Rates quoted as % of the 1976 benchmark tariff × $/ton (St. Louis $3.99, Twin Cities $6.19, Cairo–Memphis $3.14); USDA GTR publishes weekly grain-barge tons and rates.
• Sales tape: Booking → loading → transit (locks, river stage) → delivery → demurrage/cleaning → payment
• Consumption sink: Capacity dies if the barge sails underutilised (score 3).
• Primary risk / boundary: Draft restrictions, lock outages and low-water spikes (Oct 2022) are regime shocks; fleeting/demurrage hidden; declined quotes must be self-logged.
• First experiment: Grain shipper/broker's 12 months of quotes accepted/declined vs USDA weekly rates and realised transit cost.
• Confidence: medium-high (benchmarks)
• Sources: https://opentug.com/blog/comprehensive-guide-to-barge-rates · https://agtransport.usda.gov/stories/s/Barge-Dashboard/965a-yzgy/ · https://www.ams.usda.gov/sites/default/files/media/GTR08062026.pdf

Commerce — venture · proposed group: Core · New (Marketplace sponsored listings is the ad side; this is product-test selection)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 3 · 3 · 3 · 3 · 3 · 3 · 3 · 2 → structural 92.6, gates Y/Y/P/P, actionable 71.3
• Reject reasons: Time · Prediction · Attention
• Access: US individual: ID + SSN/ITIN + US bank; business: EIN + docs; approval 24–48h; Seller Center API; 24–48h order-processing SLA.
• Scale: US: ~475k–800k shops (sources conflict), ~216k with sales; US GMV 2025 ~$15B (Canopy); global $64–66B; ~2,000 US shops >$1M.
• Sales tape: Launch → ad/affiliate spend → orders → delivery → returns/refunds → payout; kill or scale within days
• Consumption sink: Venture-class product bets that die in days; goods consumed. Closest ordinary-commerce analogue to an ICO.
• Primary risk / boundary: Platform-rule volatility (referral fee 6%→8% Aug 2026; gated categories); attribution of affiliate/ad demand; constraints change under you (gate 'Constraints' = Partial).
• First experiment: Operator with 300+ past product tests: replay launch/skip decisions with ad spend and return tape; test 7-day kill/scale policy vs P34 abstention; keep fees as dated features.
• Confidence: medium
• Sources: https://canopymanagement.com/tiktok-shop-eligibility-what-you-need-to-get-started/ · https://delzonic.com/blogs/tiktok-shop-fee-increase-2026/ · https://redstagfulfillment.com/how-many-tiktok-shop-sellers/

Industrial / agri · proposed group: Industrial · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 1 · 2 · 3 · 2 · 3 · 2 · 1 · 3 → structural 74.1, gates Y/Y/Y/Y, actionable 70.4
• Reject reasons: Capacity · Operational fit · Complexity
• Access: WA DNR monthly sealed-bid auctions with bid bonds; Oregon ODF monthly bid windows; USFS open-sales dashboard; some states require licensed timber buyers with surety (e.g., Indiana).
• Scale: USFS FY2025 sold 2.94 BBF (goal 4 BBF by FY28); WA DNR monthly auctions with results published since 1969; sample GNA sales 2,442–11,547 MBF, 2 of 4 with no bids.
• Sales tape: Award → logging → scaling at mill → delivered-log payments over months → defect/grade recovery variance
• Consumption sink: Milled into lumber/pulp (converted, score 2).
• Primary risk / boundary: Feedback arrives over months–years (paper boundary: too slow vs regime); $1M+ tickets on public sales; operator fit (crews, mills) dominates; cruise vs scale error is the market.
• First experiment: Backtest 10 years of DNR/ODF results incl. no-bid sales against delivered-log price series; test whether P34 abstains on single-bidder tracts that later lost money.
• Confidence: medium-high
• Sources: https://www.dnr.wa.gov/programs-and-services/product-sales-and-leasing/timber-sales/timber-auction-results · https://www.oregon.gov/odf/working/pages/timbersales.aspx · https://www.fs.usda.gov/about-agency/accomplishments · https://timbermart-south.com/

Services · proposed group: Contracts · Adjacent: Freelance-platform work arbitrage; Lead purchasing; B2B outbound prospecting

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 0 · 2 · 2 · 3 · 3 · 3 · 2 · 2 · 3 → structural 74.1, gates Y/Y/Y/Y, actionable 70.4
• Reject reasons: Attention · Prediction · Capacity
• Access: BountyJobs admits preferred vendors/niche agencies (screening ≤72h; employer sets fee; earnings pending until 60 days of employment); Top Echelon TE Network $350 + $150/mo + 6% brokerage; state employment-agency licences in some states.
• Scale: Open job orders per network unpublished; fees 15–25% of first-year base; time-to-fill ~36 days; Paraform avg placed comp $260k.
• Sales tape: Submit → interview → placement → 60–90 day guarantee → fall-off / non-payment → fee received
• Consumption sink: A filled req is consumed (score 2).
• Primary risk / boundary: Fall-off and non-payment 60–90+ days later; capacity (recruiter hours) is the constraint; no logistics.
• First experiment: Agency's 3 years of ATS job orders (worked/passed) + placements/fall-offs; shadow-score req selection.
• Confidence: medium
• Sources: https://bountyjobs.com/recruiter-faq · https://topechelon.com/product/te-network/ · https://www.paraform.com/blog/contingency-recruiting-guide

Perishables / agri · proposed group: Industrial · Adjacent: Small livestock and equipment auctions

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 3 · 2 · 1 · 2 · 3 · 2 · 3 → structural 77.8, gates Y/Y/Y/P, actionable 70.0
• Reject reasons: Operational fit · Scale · Time
• Access: Open; auctions sold by the ton (e.g., Rock Valley IA: $125 handling, $40 no-sale fee); HayMap / Internet Hay Exchange classifieds; no API.
• Scale: USDA AMS state reports (CA 9/25/2026: 25,550 t + 44,544 bales confirmed); HayWire index 83 quotes/19 markets weekly; national $166/t (wk of 21 Sep 2026).
• Sales tape: Buy → truck → store → resale or feed-out over weeks → moisture/mold rejection
• Consumption sink: Eaten; spoils if stored badly.
• Primary risk / boundary: Few asset types (breadth low); freight 30–50% of delivered price dominates; quality tests 'honor system'; regional spreads are the edge.
• First experiment: Take 12 months of AMS + auction data; join a hay dealer's buys, freight and sales; test whether regional spread minus freight predicts realised margin and when to pass.
• Confidence: high (data) / medium (marketplace scale)
• Sources: https://www.ams.usda.gov/mnreports/ams_2904.pdf · https://haywireag.com/prices.html · https://www.rockvalleyhay.com/web/faqs/

Packaging / logistics · proposed group: Tier 4 · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 3 · 3 · 2 · 1 · 2 · 2 · 2 · 3 → structural 77.8, gates Y/Y/Y/P, actionable 70.0
• Reject reasons: Scale · Operational fit · Capacity
• Access: No licence; buy cores from generators, sell graded 48x40 by the truckload (~560 pallets); PalletTrader / Repackify marketplaces; UN/DOT reconditioning rules for hazmat-service IBCs/drums.
• Scale: 48forty: 240+ facilities, 850+ partners; Q1-2026 report: 396 verified supplier listings in 50 states; core buy avg $2.75 (range $1.60–12.84), Grade A sell avg $9.04, regional spread 85%.
• Sales tape: Buy cores → sort/repair → truckload sale → dock re-grade / rejection
• Consumption sink: Pallets break and are consumed in service (score 2).
• Primary risk / boundary: Few SKUs (breadth low); freight decides everything; grade drift at the dock; hazmat residue on totes.
• First experiment: One recycler's 12 months of core buys, sort yields, repair labour and sales by lane; score buy/pass per source and lane; test regional spread capture.
• Confidence: medium-high
• Sources: https://www.repackify.com/blog/what-is-a-gma-pallet-48x40-standard-grades-pricing · https://usedpalletrecycling.com/blog/2026-used-pallet-pricing-report · https://www.48forty.com/company-48forty

Collectibles · proposed group: Tier 1 · Adjacent: Retired LEGO sets; Sealed TCG; Sneakers (Tier 1 family)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 3 · 3 · 3 · 3 · 2 · 1 · 3 · 2 → structural 77.8, gates Y/Y/Y/P, actionable 70.0
• Reject reasons: Prediction · Time · Attention
• Access: Open retail/secondary; StockX (authenticated), eBay, Mercari, Whatnot; retailer purchase limits.
• Scale: Funko ~800 licensed properties; Hot Wheels 30 Super Treasure Hunts/yr (resale $19–350); Pop Mart #1 collectibles brand on StockX since Oct 2024, cooling in 2026 (latest drop −14.3%).
• Sales tape: Buy → hold → sale on StockX/eBay (days–months) → fees 10–15% → fad reversal write-down
• Consumption sink: Durable collectible; blind-box mechanics only (score 1).
• Primary risk / boundary: Fad-driven regimes (Labubu 2026 cooling); counterfeits (11,000+ seized Seattle 2025); low consumption → capital competition.
• First experiment: Reseller's ledger of drops bought/passed + StockX/eBay realised; test P34 on the 2025→26 Labubu regime shift.
• Confidence: medium
• Sources: https://stockx.com/about/stockxs-midyear-trend-report-has-asics-and-salomon-among-fastest-growing-sneaker-brands-sees-new-records-for-labubu-and-triple-digit-sales-growth-for-trading-cards/ · https://www.autoevolution.com/news/hot-wheels-super-treasure-hunt-prices-2025-vs-2024-253713.html · https://www.iam-media.com/article/pop-mart-intensifies-anti-counterfeiting-efforts-labubu-products-fakes-the-rise-and-prices-decline

Collectibles / fashion · proposed group: Tier 2 · Adjacent: Entry-tier watches and microbrands (authentication-risk family)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 3 · 3 · 3 · 3 · 3 · 1 · 2 · 2 → structural 77.8, gates Y/Y/Y/P, actionable 70.0
• Reject reasons: Prediction · Attention
• Access: Sell-side open; eBay Authenticity Guarantee mandatory ≥$500; TRR/Fashionphile/Rebag buyouts and consignment; no public APIs.
• Scale: TRR FY2025 GMV $2.13B, 3.59M orders, AOV $594; Rebag Clair 2025: Hermès 138%, Goyard 132% value retention; item-level Kelly Mini II 282%.
• Sales tape: Buy → authentication → listing → sale (weeks) → 15-day returns → take rate realised at sale
• Consumption sink: Wears; Hermès is a store of value (score 1).
• Primary risk / boundary: Authentication/regrade after receipt; take rates 36.5% (TRR); fashion regime shifts.
• First experiment: Reseller's quote history (Fashionphile quotes accepted/declined) + realised sales for 12 months.
• Confidence: medium-high
• Sources: https://www.prnewswire.com/news-releases/rebag-releases-its-sixth-annual-clair-report-a-comprehensive-luxury-appraisal-index-for-resale-302637452.html · https://pages.ebay.com/authenticity-guarantee-handbags-seller · https://www.globenewswire.com/news-release/2026/02/26/3246067/0/en/the-realreal-announces-fourth-quarter-and-full-year-2025-results.html

Perishables / consumables · proposed group: Industrial · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 3 · 3 · 2 · 2 · 3 · 2 · 3 → structural 81.5, gates Y/Y/Y/P, actionable 69.7
• Reject reasons: Attention · Complexity · Scale
• Access: Business account with importer (Cafe Imports, Royal Coffee): no minimum, full bags (59–70 kg) or 22–50 lb boxes; prepay first orders then credit; CoE auctions need ACE membership + sample sets.
• Scale: Hundreds of live spot lots per importer (spot/afloat/future); CoE ~10 auctions/yr × 30 lots; Nicaragua 2025: ~2,800 bids from 100+ companies.
• Sales tape: Buy spot → warehouse release → roast → wholesale/retail sale over weeks → fading write-down
• Consumption sink: Roasted and drunk; cup quality fades with age.
• Primary risk / boundary: Declined lots only recorded if offer sheets are snapshotted; value is operator-mediated (roasting + sales channel); seasonality of arrivals.
• First experiment: Snapshot two importers' spot lists daily for 90 days; join a roaster's purchase + sell-through ledger; test whether cup score/price residuals predict margin.
• Confidence: medium-high
• Sources: https://www.cafeimports.com/north-america/blog/purchase-planning/coffee-ordering-faq/ · https://royalcoffee.com/how-to-buy/ · https://cupofexcellence.org/about-coe-auctions/

Recycling · proposed group: Industrial · Adjacent: Scrap metal and e-waste recovery (different materials)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 3 · 2 · 1 · 3 · 3 · 2 · 3 → structural 81.5, gates Y/Y/Y/P, actionable 69.7
• Reject reasons: Operational fit · Scale · Complexity
• Access: No licence to broker; relationships with MRFs/generators and mills; BaleBid marketplace (open, £2/wk); truckload (18–22 t) or container lots; Basel PIC for mixed plastics export.
• Scale: BaleBid snapshot: 590 listings (1,153 t paper, 1,328 t cardboard, 7,959 t plastic); ReMA PS-2026 defines dozens of paper grades; broader flow bilateral.
• Sales tape: Buy → truck/ship → mill receipt & grading → claims within 2 business days → downgrade/rejection → payment
• Consumption sink: Pulped / re-melted (destroyed as a tradable bale).
• Primary risk / boundary: Contamination grade only at mill receipt; commodity-like index (few grades) limits long-tail edge; export rules; brokers earn $/t spreads.
• First experiment: A broker's 18 months of loads incl. quotes not taken; join mill claims; test whether P34 abstains on high-claim generators/lanes.
• Confidence: medium-high
• Sources: http://resource-recycling.com/recycling/2026/09/09/bale-pricing-mostly-flat-for-september/ · https://www.recycledmaterials.org/wp-content/uploads/Guidelines-for-Paper-Stock-Revisions-Jan-2026-1.pdf · https://www.balebid.com/

Logistics / liquidation · proposed group: Tier 2 · Adjacent: Liquidation and B-stock pallets; Storage unit auctions (different supply)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 1 · 3 · 3 · 3 · 3 · 2 · 2 · 3 → structural 85.2, gates Y/P/Y/P, actionable 69.0
• Reject reasons: Complexity · Attention · Capacity
• Access: CBP GO / Treasury auctions via contractors (CWS Marketing, Apple Auctioneering): free public registration, deposit >$5k, no premium, pickup 15–30 days; Fr8Auctions weekly sealed-bid pallets for approved wholesale buyers.
• Scale: GO goods saleable 6 months after import (19 CFR 127.4); auction frequency/lot counts unpublished; Fr8 weekly.
• Sales tape: Win → pay in 2 days → pickup ≤15 days → sort → resale by channel → unsellable disposal
• Consumption sink: Mixed consumer/industrial goods, largely consumed downstream (score 2).
• Primary risk / boundary: Contents/condition largely hidden; export-only lots; sealed bids = your own passes only; sporadic supply.
• First experiment: Log every GO/Fr8 lot and bid decision for 6 months; join resale ledger; test if manifest features + inspection notes predict recovery.
• Confidence: medium
• Sources: https://cwsmarketing.com/faqs/faqs-us-customs-go-merchandise/ · https://www.ecfr.gov/current/title-19/chapter-I/part-127 · https://fr8auctions.com/turning-lost-freight-into-lucrative-opportunities/

Construction reuse · proposed group: Tier 2 · Adjacent: Construction material surplus (new excess) — reclaimed/deconstruction supply is new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 3 · 3 · 3 · 3 · 2 · 2 · 3 → structural 88.9, gates Y/P/Y/P, actionable 68.4
• Reject reasons: Operational fit · Complexity · Capacity
• Access: Habitat ReStores (affiliate-run, donation-fed, walk-in), deconstruction contractors, salvage yards, reuse platforms (PlanetReuse, Rheaply — unverified); deconstruction ordinances create supply in some cities.
• Scale: ReStore count commonly cited ~900 US (unverified); no sales figures; inventory unlisted.
• Sales tape: Buy → deconstruct/haul → de-nail, re-mill → sale to a job or yard → yield loss, lead/asbestos disposal
• Consumption sink: Installed into buildings (score 2).
• Primary risk / boundary: Condition variance and haul cost dominate; structural reuse needs re-grading/engineer sign-off; lead paint and asbestos; walk-in inventory means the menu must be built by hand.
• First experiment: Salvage yard's 12 months of lots offered/taken with haul and re-milling cost vs realised sales.
• Confidence: low-medium
• Sources: https://www.habitat.org/restores · https://thegreenmissioninc.com/the-lumber-market-in-2026/

Auto consumables · proposed group: Tier 2 · Adjacent: Auto parts, OEM take-offs and cores

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 2 · 2 · 3 · 2 · 3 · 2 · 3 → structural 77.8, gates Y/Y/Y/P, actionable 66.5
• Reject reasons: Scale · Attention · Operational fit
• Access: Dealer account with distributors (ATD, US AutoForce); scrap/used-tire hauler registration and manifests by state; no public API.
• Scale: ATD: 90,000+ SKUs, 110+ DCs, 80,000+ customers; ~250M scrap tires/yr in US (79% to end markets); closeout flow unverified.
• Sales tape: Buy → LTL → fit/resell over weeks → warranty/returns → scrap disposal fee
• Consumption sink: Worn out — literal consumable.
• Primary risk / boundary: Closeout flow and tire-only margin unverified; bulky low value density; DOT-age risk on aged closeouts.
• First experiment: A regional tire shop's 12 months of distributor closeout offers + used-lot buys vs sell-through; score by size/age; measure pass value on slow sizes.
• Confidence: medium-low
• Sources: https://www.atd.com/about-us/article/we-are-american-tire-distributors/ · https://www.ustires.org/advocacy/tire-recycling-markets · https://www.moderntiredealer.com/home/news/55263033/mtds-facts-issue-examines-us-tire-market

Energy hardware · proposed group: Industrial · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 2 · 2 · 2 · 3 · 2 · 2 · 3 → structural 77.8, gates Y/Y/Y/P, actionable 66.5
• Reject reasons: Complexity · Operational fit · Prediction
• Access: EnergyBin members-only B2B; SanTan Solar / A1 SolarStore open with pallet minimums (20–25 panels); secondsol (EU) open.
• Scale: Global secondary ≈1 GW/yr; US exports >50 MW in 2025; EnergyBin 14,462 modules posted early 2026; 98% of 'resale' modules in 2025 were excess NEW stock.
• Sales tape: Buy → flash/EL test → pallet freight or export → sale to installer/exporter → breakage/hazardous-waste determination
• Consumption sink: Installed for 20+ years (leaves the market); failed panels become regulated waste (score 2).
• Primary risk / boundary: Sellers list without flash data; margins 'often too thin'; EPA hazardous-waste determination if unfit; UFLPA traceability.
• First experiment: Reseller's 12 months of lots with test results and sales; test whether P34 abstains on untested/low-W lots.
• Confidence: medium-high
• Sources: https://resources.energybin.com/solar-resources/global-solar-panel-resale-poised-to-take-off-despite-challenges · https://www.epa.gov/hw/solar-panel-frequent-questions · https://www.secondsol.com/en/index.htm

Venture · proposed group: Tier 3 · Adjacent: Micro-website and newsletter acquisitions (Tier 3, sub-$5k) — main-street tier

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 2 · 2 · 3 · 3 · 3 · 3 · 1 · 3 → structural 77.8, gates Y/Y/Y/P, actionable 66.5
• Reject reasons: Complexity · Prediction · Capacity
• Access: Free browsing + broker NDAs (BizBuySell ~50,000 tracked); Acquire.com 2,200+ live listings, ~13% of submissions approved, paid buyer tiers; SBA 7(a) needs US-person ownership, equity injection, guarantees.
• Scale: BizBuySell 2025: 9,586 closed deals, $7.95B; median price $350k, 0.69× revenue, 2.61× cash flow, 170 days to close; Acquire.com SaaS 3.9× profit, ~81 days on market.
• Sales tape: LOI → diligence → close → post-close cash flow over years → seller-note default / earn-out
• Consumption sink: Venture-class: businesses die (score 3).
• Primary risk / boundary: Paper boundary: feedback arrives years later; one operator cannot reach 100+ observed deals → pooled/enterprise (search-fund or roll-up) data only.
• First experiment: Roll-up or search-fund platform with 100+ acquisitions and a decline log; otherwise blueprint only.
• Confidence: high (BizBuySell/Acquire figures)
• Sources: https://www.bizbuysell.com/blog/2025-year-in-review/ · https://blog.acquire.com/acquire-com-biannual-acquisition-multiples-report-jan-2026/ · https://ctacquisitions.com/guides/small-business-ma-statistics-2026/

Renewable feedstock / routes · proposed group: Tier 4 · New — not in the 135

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 2 · 2 · 2 · 2 · 3 · 2 · 3 → structural 77.8, gates Y/Y/Y/P, actionable 66.5
• Reject reasons: Operational fit · Scale · Capacity
• Access: Place containers at restaurants, haul to renderer/aggregator (Darling/DAR PRO, Mahoney/Neste, regionals); accounts won with per-gallon rebates and free equipment; state grease-hauler registration (e.g., CA CDFA inedible kitchen grease), DOT; renderer acceptance specs.
• Scale: Restaurant-side rebates $2.00–4.50/gal (avg ~$3, 2025); national volume not verified (industry-cited 3B+ lb/yr, unverified).
• Sales tape: Pickup → volume & FFA/moisture test at renderer → deduction → payment; theft losses
• Consumption sink: Burned as biodiesel / renewable diesel or rendered into feed.
• Primary risk / boundary: Sticky incumbents and permitting are the binding constraints; price keyed to RFS/45Z/LCFS credit values; theft (Dec 2025 multistate ring); per-stop yield hidden until pickup.
• First experiment: Collector's 12 months of stops (signed and passed), pickup volumes, deductions and route costs; score stop selection under route density.
• Confidence: medium-low
• Sources: https://greaseconnections.com/used-cooking-oil-pricing-guide/ · https://thejacobsen.com/about/methodology/uco-spec/ · https://www.argusmedia.com/en/methodology/key-commodity-prices/argus-us-uco-price-assessments

Recycling / apparel · proposed group: Industrial · Adjacent: Deadstock and branded apparel lots (new goods) — this is used/export

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 3 · 3 · 2 · 3 · 2 · 1 · 3 → structural 81.5, gates Y/P/Y/P, actionable 66.0
• Reject reasons: Complexity · Operational fit · Attention
• Access: No licence; graders/exporters (Bank & Vogue, Buffalo Export, US Dynamics) quote container lots (40'HC ≈ 40,000 lb credential); Queen of Raw / FabScrap for fabric.
• Scale: Buffalo Export ~50M lb/yr; Bank & Vogue in 34 countries; grades supplier-defined (Cream/A/B/C); listing counts unverified.
• Sales tape: Buy → ocean freight → sort → graded resale by destination → unsold/rag write-off
• Consumption sink: Worn out over time (score 2).
• Primary risk / boundary: Yield of saleable grade unknown until sorted; bale-level attribution hard (gate 'Outcome' = Partial); weeks–months cycle; destination import bans.
• First experiment: Sorter's 12 months of bale purchases with sort yields and graded sales; test if source/grade features predict yield-adjusted margin.
• Confidence: low-medium
• Sources: https://www.bankvogue.com/blog/credential-clothing-bale-guide-for-buying/ · https://www.buffaloexportllc.com/credentialclothing.html · https://resale.queenofraw.com/

Demand — offline media · proposed group: Demand · Adjacent: Small-space advertising; Display and programmatic inventory — offline remnant is new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 3 · 2 · 3 · 3 · 2 · 3 · 3 · 2 → structural 81.5, gates Y/P/Y/P, actionable 66.0
• Reject reasons: Time · Attention
• Access: Through remnant agencies, direct station/network desks or the DR/infomercial market; no licence.
• Scale: No volume or discount figures verified; agencies claim large discounts to rate card (unverified).
• Sales tape: Buy → airing (no guarantee; preemptable) → response/attribution → make-goods
• Consumption sink: Unsold inventory is destroyed when the slot passes (score 3).
• Primary risk / boundary: Hidden variable is delivery of the asset itself (airing not guaranteed); attribution weak; keep the novel slice offline — DOOH overlaps the catalogue's demand group.
• First experiment: DR advertiser's 12 months of remnant buys/passes with response data; test abstention on preempt-prone dayparts.
• Confidence: low-medium
• Sources: https://www.simulmedia.com/blog/expert-guide-to-remnant-inventory-and-tv-advertising-strategies

Logistics · proposed group: Capacity · Adjacent: Container repositioning (operator side) — this is trading/SOC leasing

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 2 · 2 · 2 · 2 · 1 · 2 · 3 → structural 70.4, gates Y/Y/Y/P, actionable 63.4
• Reject reasons: Operational fit · Attention · Capacity
• Access: Container xChange: trading buyer $399/mo, wholesaler $899/mo, leasing $399–649/mo (12-mo commitment), vetted partners; retail via broker app.
• Scale: xChange: 1,700+ vetted partners, 2,500+ locations, 29,000+ containers across 150+ NA depots; 2026 US delivered: 20' CW $2,800–4,000, 40'HC WWT $3,500–5,500.
• Sales tape: Buy/lease → survey → trucking → resale or drop-off → damage/repair charges
• Consumption sink: Durable; some boxes exit to storage conversion (score 1).
• Primary risk / boundary: Subscription gate; regional imbalances shift quickly; survey vs grade disputes; small spread on trading.
• First experiment: Use xChange history for 12 months of listings + one-way credits; join a trader's buys/leases; test if imbalance features predict per-box margin.
• Confidence: medium-high
• Sources: https://www.container-xchange.com/pricing · https://www.container-xchange.com/blog/one-way-container · https://containerone.net/blogs/news/how-much-does-a-shipping-container-cost-complete-2026-price-breakdown

Energy hardware — crypto-adjacent · proposed group: Industrial · New (hardware is ordinary commerce, but value keyed to BTC hashprice)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 3 · 2 · 1 · 2 · 2 · 3 · 2 → structural 70.4, gates Y/Y/Y/P, actionable 63.4
• Reject reasons: Time · Prediction · Scale
• Access: Open; Kaboomracks (tested/graded, 7-day not-DOA), Luxor Hardware, Bitmain/MicroBT secondary, eBay; freight; hosting $/kWh.
• Scale: Small model universe (~10 current-gen SHA-256); prices $10–30/TH (Hashrate Index 2026); hashprice $31.89/PH/day (Aug 2026) near breakeven.
• Sales tape: Buy → burn-in → resale or hosting → hashboard failures → obsolescence write-down
• Consumption sink: Economically consumed by difficulty growth (score 2).
• Primary risk / boundary: Paper boundary: target dominated by one irreducible shock (BTC price/hashprice); tariffs 21.6–57.6% on imports; breadth low.
• First experiment: Only as a hardware-dealer ledger test with hashprice as a feature; expect low incremental value over a hashprice rule.
• Confidence: medium
• Sources: https://hashrateindex.com/blog/top-10-bitcoin-mining-asic-machines-for-2026/ · https://kaboomracks.com/faq/ · https://www.theblock.co/post/365793/trump-tariffs-bitcoin-miners

Liquidation · proposed group: Tier 2 · Adjacent: Returned-goods resale; Appliance and electronics repair-and-flip

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 1 · 2 · 2 · 3 · 1 · 2 · 3 → structural 70.4, gates Y/Y/Y/P, actionable 63.4
• Reject reasons: Capacity · Operational fit · Scale
• Access: Distributor programs (NeuAppliance for Best Buy/Costco/LG/GE/Samsung/Whirlpool): business buyers, pre-made manifest truckloads only; several programs 'enrollment cap met'.
• Scale: Truckload = 40–80 units, $20–35k; Best Buy S&D loads 95% no repair / 4% minor / 1% parts; returns loads 70/10/20.
• Sales tape: Receive → test/repair → showroom/online sale over weeks → warranty claims
• Consumption sink: Durable (score 1).
• Primary risk / boundary: Rationed supply (enrolment caps) → menu is thin; hidden functional faults; warehouse and delivery needed.
• First experiment: Discount appliance store's 12 months of loads (taken/passed) + repair and sell-through ledger.
• Confidence: medium
• Sources: https://neuappliancewholesale.com/pages/best-buy-appliance-liquidation-program · https://neuappliancewholesale.com/pages/wholesale-appliance-liquidation-programs

Collectibles · proposed group: Tier 1 · Adjacent: Trading cards — sport and TCG (same family, missing lines)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 3 · 3 · 3 · 3 · 1 · 0 · 2 · 1 → structural 63.0, gates Y/Y/Y/Y, actionable 63.0
• Reject reasons: Attention · Prediction
• Access: Free registration; weekly internet auctions per category; buyer's premium ~20% (Heritage), ~10% (GreatCollections, unverified).
• Scale: Heritage archive: millions of lots (free with registration); weekly auctions across coins/comics/currency; GreatCollections $250M+ sold.
• Sales tape: Win → hold/regrade → resale at auction (weeks) → round-trip fees 20–30%
• Consumption sink: Pure store of value — no consumer (score 0).
• Primary risk / boundary: Boundary example: cleanest telemetry but no consumption and low ops → open capital competes away the edge; round-trip fees dominate.
• First experiment: Cheap benchmark: replay 5 years of Heritage archives; likely shows P34 ≈ comps rule; use to calibrate, not to operate.
• Confidence: medium
• Sources: https://www.ha.com/information/about-auction-archives.s · https://www.pcgs.com/publicapi · https://www.cgccomics.com/population-report/

Logistics capacity · proposed group: Capacity · Adjacent: Container repositioning; Equipment rental — chassis/trailer pools are new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 3 · 1 · 2 · 2 · 2 · 3 · 3 · 3 → structural 81.5, gates Y/P/Y/P, actionable 62.7
• Reject reasons: Capacity · Time · Operational fit
• Access: Chassis pools (DCLI, TRAC, Flexi-Van) require a UIIA-registered motor carrier with insurance (new UIIA effective 5 May 2026); unregistered users pay DCLI $55/day; trailer/reefer spot marketplaces unverified.
• Scale: DCLI daily rates from 1 Jul 2026: $28.55 (Southeast) to $47.50 (Pacific SW) incl. damage waiver; LA/LB short-term $20/day → $10 after 90 days.
• Sales tape: On-hire → use → gate fees, repositioning penalties ($250), damage disputes, tolls → off-hire; detention/demurrage
• Consumption sink: Capacity-days perish (score 3).
• Primary risk / boundary: Pool access gated; detention/demurrage accounting messy (FMC rule under OSRA 2022); attribution of per-day cost to a load is partial.
• First experiment: Drayage carrier's 6 months of daily chassis/trailer decisions with per-diem, penalties and load revenue.
• Confidence: high (chassis rates) / low (trailer marketplaces)
• Sources: https://dcli.com/daily-market-rates/ · https://uiia.intermodal.org/documents/uiia/newuiia-Home.pdf

Travel capacity · proposed group: Capacity · Adjacent: Air cargo capacity

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 1 · 2 · 2 · 2 · 3 · 3 · 2 → structural 70.4, gates Y/Y/Y/P, actionable 60.2
• Reject reasons: Time · Operational fit
• Access: Avinode broker tiers $740–2,119/mo (API on Ultimate); consumer apps (JetASAP, XO) list legs; DOT Part 295 disclosure duties; no licence.
• Scale: Avinode 3,400+ aircraft, 13,500+ professionals; ~300k requests/month (2018); ~40% of European bizav flights are empty legs; legs/day unpublished.
• Sales tape: Book → flight flies or cancels (operator may cancel) → client payment → refunds
• Consumption sink: Seat-hours perish (score 3).
• Primary risk / boundary: Leg may not fly; matching market rather than inventory; subscription gate.
• First experiment: Broker's 12 months of Avinode requests, quotes and booked legs; score match decisions.
• Confidence: medium
• Sources: https://avinode.com/pricing/ · https://nbaa.org/aircraft-operations/part-135/air-charter-brokers/14-cfr-part-295-air-charter-broker-questions-answers/ · https://theflyingengineer.com/private-jet-empty-legs/

Real estate · proposed group: Tier 3 · Adjacent: Tax lien certificates (flagged) — outright parcel purchase is ordinary property

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 3 · 3 · 3 · 3 · 3 · 0 · 1 · 2 → structural 70.4, gates Y/Y/Y/P, actionable 60.2
• Reject reasons: Attention · Prediction · Complexity
• Access: No licence to buy/sell own parcels; county delinquent lists; seller financing on vacant land largely outside Dodd-Frank/Reg Z (no dwelling); Reg N and state subdivision statutes apply.
• Scale: LandWatch 526,047 US listings (Sep 2026), ~192k listed in last 60 days (≈3,000/day); off-market parcel counts unverified.
• Sales tape: Offer → title/escrow → resale or seller-financed note over months–years → buyer default
• Consumption sink: Land is not consumed (score 0).
• Primary risk / boundary: Slow realisation; title/access/wetland defects; seller-financed notes stretch feedback to years; no consumer.
• First experiment: Land flipper's 3 years of mail-campaign offers (accepted/declined) + resale; test if P34 beats their offer formula.
• Confidence: medium
• Sources: https://retipster.com/dodd-frank/ · https://www.landwatch.com/land · https://www.landwatch.com/land/listed-60-days

Feed / agri · proposed group: Industrial · Adjacent: Industrial by-products and side streams; Hay, forage and feed lots

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 1 · 2 · 1 · 2 · 3 · 2 · 3 → structural 70.4, gates Y/Y/Y/P, actionable 60.2
• Reject reasons: Operational fit · Scale
• Access: Plant-gate/FOB-plant sales negotiated with merchandisers; no exchange; needs feed-buyer relationships and trucking; state commercial-feed licensing and FDA/FSMA animal-food rules (unverified).
• Scale: Plant count/volumes not verified this pass (RFA data).
• Sales tape: Buy → truck → feedlot delivery → quality claims → payment
• Consumption sink: Fed to animals (score 3).
• Primary risk / boundary: Few product types (breadth low); quality variance (moisture, protein, fat after oil extraction) not in the published price; corn/ethanol correlation; relationship-heavy access.
• First experiment: Feed merchandiser's 12 months of plant-gate offers taken/passed vs USDA weekly and realised delivered margin.
• Confidence: medium (data) / low (practice)
• Sources: https://mymarketnews.ams.usda.gov/viewReport/3618

Industrial / agri · proposed group: Industrial · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 2 · 1 · 2 · 1 · 3 · 2 · 2 → structural 63.0, gates Y/Y/Y/Y, actionable 59.9
• Reject reasons: Scale · Operational fit
• Access: AWEX: register via declaration forms; open-cry auctions Sydney/Melbourne/Fremantle ~46 weeks/yr; brokers publish catalogues with AWTA presale tests. Hides: packer tenders via relationships.
• Scale: Weekly offerings 34k–41k bales (8–14% passed in); ~85% of Australian wool sold at auction; bale ≈180 kg greasy.
• Sales tape: Buy → dump/ship → mill sale → claims
• Consumption sink: Spun/tanned — consumed (score 3).
• Primary risk / boundary: Boundary example: objectively measured pre-sale → little hidden mispricing; passed-in lots are an excellent declined-option record but the edge is thin.
• First experiment: Cheap benchmark: AWEX passed-in vs re-offered price history; expect small P34 edge.
• Confidence: medium-high (wool) / low (hides)
• Sources: https://www.awex.com.au/auction/ · https://www.sheepcentral.com/wool-prices-tumble-under-weight-of-bigger-offering/ · https://www.ams.usda.gov/mnreports/nw_ls444.txt

Finance / commerce — regulated · proposed group: Core · Adjacent: Micro-lending — individual (Active experiment) — collateralised variant

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 1 · 3 · 1 · 3 · 3 · 3 · 1 · 2 · 3 → structural 74.1, gates Y/Y/Y/Y, actionable 51.9
• Reject reasons: Prediction · Attention · Complexity
• Access: State and/or local pawnbroker licence (bond $5k–100k+), police reporting (LeadsOnline), 30–90 day holds; fee caps 2–25%/month by state.
• Scale: ~7,771 US stores (NPA); FirstCash 1,207 US stores, avg loan $312 (2025), redemption ~85%, retail margin ~42–43%, inventory turns 2.8×/yr.
• Sales tape: Loan → fees → redemption or forfeiture → inventory at principal → resale over months → markdown
• Consumption sink: Goods redeemed or resold; not consumed (score 1).
• Primary risk / boundary: Regulated consumer credit → gated launch like micro-lending; declined offers exist but only in POS; forfeiture resale months later.
• First experiment: Enterprise lane with a chain using Bravo: 2 years of offers (incl. declines) + redemption + resale; shadow-score offer amounts.
• Confidence: medium-high
• Sources: https://www.stocktitan.net/sec-filings/FCFS/10-k-first-cash-holdings-inc-files-annual-report-b004a84f37ec.html · https://pawn-software.com/pawn-licensing-us.htm

Consumables — regulated · proposed group: Core · New (regulated commerce)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 3 · 1 · 3 · 3 · 2 · 3 · 2 · 3 → structural 81.5, gates Y/Y/Y/P, actionable 51.3
• Reject reasons: Attention · Prediction · Complexity
• Access: Valid state cannabis licence required (retailer/distributor/processor); LeafLink verifies every licence; platform free for retailers; cannabis-specific banking.
• Scale: LeafLink: 34 states, 12,000+ companies, >$5B annual GMV, 'millions of orders'; ~398,000 unique SKUs (141,488 flower) in the 2025 pricing guide.
• Sales tape: Order → licensed transport (manifest) → retail sell-through over weeks → price compression → expiry/destruction
• Consumption sink: Consumed; flower degrades in months; expired product destroyed under state rules.
• Primary risk / boundary: Federal illegality (280E, banking, no interstate) — HyperC would gate or exclude; licence required to hold inventory; retail sell-through is the real outcome.
• First experiment: Enterprise-only: a multi-store retailer's 2 years of LeafLink orders + catalogue snapshots + POS sell-through; shadow-score reorders; do not run on membership.
• Confidence: high (scale) / medium (economics)
• Sources: https://www.leaflink.com/press/leaflink-marks-a-decade-of-powering-cannabis-commerce-growth-across-34-markets-and-billions-in-sales/ · https://www.leaflink.com/press/leaflinks-2025-wholesale-cannabis-pricing-guide-now-available/ · https://www.cannabisbenchmarks.com/reports/

Aviation · proposed group: Industrial · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 1 · 3 · 3 · 3 · 2 · 2 · 3 → structural 81.5, gates Y/Y/Y/P, actionable 51.3
• Reject reasons: Complexity · Capacity · Attention
• Access: ILS / PartsBase subscriptions (PartsBase 83,000 members, 7,600+ companies); ASA-100 accreditation for credibility; parts need FAA 8130-3 / EASA Form 1 and back-to-birth trace.
• Scale: PartsBase 15B+ line items, 3.3M repair capabilities; USM market $7.8B (2025) → $12.1B (2035); ~1,100 aircraft retirements/yr.
• Sales tape: Buy → inspection/repair (shop visit, BER risk) → sale → warranty returns
• Consumption sink: Expendables consumed; rotables repaired (score 2).
• Primary risk / boundary: Accreditation and trace paperwork gate; shop-visit findings only after purchase; ITAR/EAR; unapproved-parts liability.
• First experiment: Enterprise lane with an ASA-100 distributor: 3 years of RFQs, buys, teardown results and sales; test abstention on BER-prone P/Ns.
• Confidence: medium
• Sources: https://aviation.partsbase.com/partsbase-landing-page-homepage/ · https://www.snsinsider.com/reports/used-serviceable-material-market-8691 · https://www.aviationsuppliers.org/asa-100

Consumables — regulated · proposed group: Flagged · Adjacent: Short-dated food closeouts — the reagent/device version

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 1 · 1 · 3 · 3 · 2 · 3 · 3 · 3 → structural 81.5, gates Y/Y/Y/P, actionable 51.3
• Reject reasons: Time · Scale · Complexity
• Access: Sell-side open (XS Supply: offers within 1 business day); buy-side via LabX / Bimedis / Salvex; reselling Rx-only devices into CA needs a Board of Pharmacy wholesaler licence; other states vary. Pharma-specific venues: PharmaTradz, RxOneShop, GSA medical lots; start with OTC/devices/GSA — Rx needs a VAWD-accredited pharmacy/wholesaler partner.
• Scale: Unverified — fragmented catalogue-style resellers; XS Supply '$2M+ recovered for 500+ facilities' (cumulative).
• Sales tape: Buy → cold-chain → sale to licensed buyer before expiry → expired write-off (training/research only)
• Consumption sink: Consumed or expires — hard deadline.
• Primary risk / boundary: Scale unverified; expiry makes abstention decisive; licensing per state; expired product cannot go to clinical use.
• First experiment: Surplus reseller's ledger (offers received, accepted, expiry, sale date); test expiry-aware sizing.
• Confidence: low
• Sources: https://xs-supply.com/pages/sell · https://simasgovlaw.com/dont-get-caught-without-a-california-pharmacy-wholesaler-license/ · https://www.labx.com/categories/reagents-and-synthesis

Digital infrastructure · proposed group: Tier 3 · Adjacent: Wholesale bandwidth and transit; Expiring domain drops — address blocks are new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 0 · 2 · 3 · 2 · 2 · 2 · 1 · 2 · 1 → structural 55.6, gates Y/Y/Y/P, actionable 50.0
• Reject reasons: Attention · Prediction
• Access: IPXO marketplace (lessor/lessee; LOA/RPKI/abuse handled) or brokers (IPv4.Global/Hilco Streambank); sale via RIR transfer (ARIN needs-justification); needs an ASN + upstream to announce; no licence.
• Scale: June 2026: avg $21.49/IP sale price (−24% YoY), ~580k IPs traded in the month (IPv4Center headline); /16 below $20/IP since June 2025; leases ~$0.40/IP/month (IPXO 2025), APNIC region >$0.60.
• Sales tape: Lease → announce → abuse/blocklist events → renewals / default → return
• Consumption sink: Addresses return at lease end — not consumed (score 1).
• Primary risk / boundary: Thinner than freight, no logistics, drifting exchange-like as transparency rises; blocklist and geolocation lag hidden; RIR policy variance.
• First experiment: Research-only: replay a lessor's 24 months of listings, offers declined and realised utilisation/abuse.
• Confidence: medium-high (prices)
• Sources: https://circleid.com/posts/2025-ipv4-price-trends-2026-predictions · https://www.ipxo.com/blog/ipv4-price-history/

Consumables — regulated · proposed group: Tier 2 · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 1 · 3 · 3 · 2 · 2 · 2 · 2 → structural 70.4, gates Y/Y/Y/Y, actionable 49.3
• Reject reasons: Prediction · Attention · Complexity
• Access: Liv-ex B2B only: Data membership then Trading membership at Exchange Committee discretion, joining + annual fee, weekly GBP settlement; WineBid open to consumers (52 auctions/yr). US entity needs TTB permit + state wholesaler licences (three-tier).
• Scale: Liv-ex 500–620 merchants in 42–47 countries; 16,000+ tradable wines; ~£100M live bids/offers (2018); LWIN 200,000+ wines & spirits; WineBid 'thousands of new bottles each week'.
• Sales tape: Buy in bond → storage → sale to trade/collector over weeks–months → condition disputes
• Consumption sink: Eventually drunk, but a large store-of-value component (score 2).
• Primary risk / boundary: Membership at committee discretion; alcohol licensing; provenance/counterfeit only at physical check-in; cask market rife with fraud (flag).
• First experiment: Use a Liv-ex data membership + WineBid archive to build 24 months of menus incl. lots not won; shadow-score a merchant's buys; exclude casks entirely.
• Confidence: medium-high
• Sources: https://www.liv-ex.com/liv-ex-membership-terms/ · https://www.liv-ex.com/lwin/ · https://www.winebid.com/Help/WhyWineBid · https://www.thedrinksbusiness.com/2025/03/calls-mount-to-regulate-sham-scammer-whisky-cask-investment-industry/

Real-estate ops — regulated · proposed group: Flagged · Adjacent: Peer storage and parking leasing; Equipment and vehicle rental flipping

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 1 · 3 · 3 · 3 · 3 · 1 · 3 → structural 77.8, gates Y/Y/Y/P, actionable 49.0
• Reject reasons: Prediction · Complexity · Capacity
• Access: Landlord consent + city STR permit + platform permit enforcement; effectively barred for non-resident operators in NYC/SF/LA; workable in Nashville, Houston, Phoenix/Scottsdale, several FL cities.
• Scale: AirDNA tracks 12,074 US markets; candidate menu = rental listings × market (Zillow/Apartments.com).
• Sales tape: Lease → furnish → nightly bookings → occupancy taxes → damage → regulatory change → lease exit
• Consumption sink: Room-nights perish (score 3).
• Primary risk / boundary: City-by-city rules change under a 12-month lease (constraints not stable); 12-month cadence; HyperC would flag like reservation arbitrage.
• First experiment: Only after counsel; operator with 50+ units' lease decisions + AirDNA + realised P&L.
• Confidence: medium
• Sources: https://www.rakidzich.com/articles/is-airbnb-rental-arbitrage-legal-state-by-state-guide-2026 · https://www.airdna.co/vacation-rental-data/app/us · https://www.bnbcalc.com/blog/short-term-rental-regulation/new-york-city-guide

Agri distribution — regulated · proposed group: Core · Adjacent: Procurement; Trade credit and payment terms — the ag-retail menu is new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 1 · 3 · 3 · 3 · 3 · 1 · 3 → structural 77.8, gates Y/Y/Y/P, actionable 49.0
• Reject reasons: Prediction · Complexity · Capacity
• Access: State pesticide-dealer licence for RUPs (FIFRA), state fertilizer registration/tonnage reporting; manufacturer return/carryover and rebate programs are contractual and non-public.
• Scale: CropLife 100 (2021): top 10 retailers = 3,288 facilities (68%); Nutrien ~1,000 outlets; total US outlets often cited 6,000–7,000 (unverified).
• Sales tape: Order → season sales on credit → carryover/returns → rebates → write-downs; collections
• Consumption sink: Applied to fields — consumed (score 3).
• Primary risk / boundary: Seasonal regime shifts (one menu per season); licences and EPA rules; dealer cost hidden in rebates; farm credit risk; months-scale loop.
• First experiment: Enterprise lane with a multi-outlet retailer: 3 seasons of buys, carryover and sell-through; shadow-score prebuy sizing.
• Confidence: medium
• Sources: https://www.croplife.com/croplife-top-100/top-10-ag-retailers-with-the-most-locations/ · https://www.nass.usda.gov/Charts_and_Maps/Agricultural_Prices/prod1.php

Equipment — regulated · proposed group: Flagged · New (regulated commerce)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 2 · 1 · 3 · 3 · 3 · 1 · 1 · 3 → structural 74.1, gates Y/Y/Y/P, actionable 46.7
• Reject reasons: Complexity · Capacity · Prediction
• Access: DOTmed free registration; Centurion auctions: $2,000 deposit, 20% premium, pickup 7 business days; Rx-device resale restricted to licensed practitioners/facilities (21 CFR 801.109); state device-distributor licences vary.
• Scale: DOTmed: 375k+ registered users, 'thousands of new listings/month'; Centurion 100+ auctions/yr.
• Sales tape: Win → de-install (radiation/magnet) → refurb → sale to licensed buyer over weeks–months
• Consumption sink: Durable (score 1).
• Primary risk / boundary: Regulatory: FDA refurbisher vs remanufacturer, HIPAA data, state radiation registration; months-long cycle; tube/detector life hidden.
• First experiment: Enterprise-only with a refurbisher; counsel review of resale licensing first.
• Confidence: medium
• Sources: https://www.dotmed.com/features/stats.html · https://centurionservice.com/buying/medical-auction-bidding · https://relinkonline.com/blogs/homepage/navigating-regulatory-compliance-when-selling-used-medical-devices-what-vendors-need-to-know-in-2026

Commerce — IP-flagged · proposed group: Flagged · Adjacent: Discontinued cosmetics and fragrance; Distributor excess inventory — grey import angle is new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 2 · 2 · 3 · 3 · 3 · 3 · 2 · 3 → structural 85.2, gates Y/Y/Y/P, actionable 46.0
• Reject reasons: Complexity · Attention · Prediction
• Access: Open commercially (importer of record, customs broker); sourcing via diverters/closeout wholesalers; Amazon gating needs authorised invoices (≥10 units, entity match) — diverter invoices usually fail.
• Scale: No aggregate volume data; categories: fragrance, cosmetics, small electronics, luxury; pharma excluded.
• Sales tape: Import → listing → sale → IP takedown / stranded inventory → returns
• Consumption sink: Consumed.
• Primary risk / boundary: Legal boundary: Tariff Act §526 / Lever rule (materially different goods excluded), Lanham Act material-differences test, EU regional exhaustion; marketplace gating — HyperC should flag under API Terms.
• First experiment: Counsel review first; then restrict to non-materially-different, non-gated categories; test on eBay-channel ledger only.
• Confidence: medium (law high; scale low)
• Sources: https://www.mayerbrown.com/en/insights/publications/2025/10/navigating-the-gray-market-landscape · https://goaura.com/blog/amazon-top-restricted-brands-in-2025

Consumables — regulated · proposed group: Industrial · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 1 · 1 · 3 · 2 · 3 · 3 · 1 · 3 → structural 70.4, gates Y/Y/Y/P, actionable 44.4
• Reject reasons: Complexity · Capacity · Prediction
• Access: Untaxpaid bulk wine only to TTB-bonded premises (winery or bonded wine cellar; bond min $1,000) + state ABC; brokers (Ciatti, Turrentine) intermediate; custom-crush is the asset-light route.
• Scale: 29.2M gal actively for sale statewide (Turrentine, Mar 2026); 2025 crush 2.62M t (lowest since 1999); lot counts unpublished.
• Sales tape: Buy in bond → transfer → blend/bottle → distribution sale over months → unsold write-down
• Consumption sink: Drunk; oversupply is being destroyed (vine removals).
• Primary risk / boundary: Bonded-premises gate; months-long realisation; buyer's market until ~2027–28 = strong drift regime; declined lots only via broker snapshots.
• First experiment: Enterprise lane with a négociant: 3 years of bulk buys + passes + bottled sell-through; test P34 abstention under the 2024–26 oversupply drift.
• Confidence: medium-high (supply/prices) / low (margins)
• Sources: https://www.turrentinebrokerage.com/wp-content/uploads/2026/03/Turrentine-Market-Update-3-26.pdf · https://www.ecfr.gov/current/title-27/chapter-I/subchapter-A/part-24/subpart-D · https://www.pressdemocrat.com/2026/01/23/wine-business-forecast-grapes-turrentine-moramarco-vinoshipper/

Environmental commodities · proposed group: Core · New

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 0 · 2 · 2 · 3 · 3 · 3 · 3 · 2 · 1 → structural 70.4, gates Y/Y/Y/P, actionable 44.4
• Reject reasons: Prediction · Attention · Complexity
• Access: Verra general account: $750 opening + $750/yr + KYC (no project needed); Xpansiv CBL onboarding unpublished (1,100+ participants); Gold Standard/ACR/CAR terms unverified.
• Scale: 168M credits retired in 2025 (the consumption sink); 22,000+ projects × vintages × methodologies; CBL 300M t cumulative; $12.3B forward offtakes announced 2025 at ~$180/t avg.
• Sales tape: Buy → hold → resale or retirement → rating reassessment / methodology downgrade write-down
• Consumption sink: Retirement literally destroys the credit — purest consumption sink in the list.
• Primary risk / boundary: Price dispersion $2 to >$50 within one type; integrity events reprice held inventory; CFTC anti-fraud authority; commodity-regulated (CFTC anti-fraud), not securities — treat as a gated launch; no logistics/ops moat.
• First experiment: Research-only: replay registry issuance/retirement + rating changes 2022–25 against a trader's book; test abstention before integrity downgrades.
• Confidence: medium
• Sources: https://verra.org/new-verra-program-fee-schedule-faqs/ · https://www.sylvera.com/blog/sylvera-state-of-carbon-credits-2025-market-shifts-from-volume-to-value · https://www.xpansiv.com/trading-platforms/cbl · https://www.morganlewis.com/pubs/2024/09/cftc-issues-final-guidance-on-listing-voluntary-carbon-credit-derivative-contracts

Industrial recovery — regulated · proposed group: Industrial · Adjacent: Scrap metal and e-waste recovery — assay-uncertainty market is new

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 1 · 1 · 3 · 2 · 3 · 2 · 1 · 3 → structural 70.4, gates Y/Y/Y/P, actionable 44.4
• Reject reasons: Complexity · Capacity · Prediction
• Access: Toll refiners (Sabin, Heraeus, Ecotrade, Red Fox): lot sampled 100%, split samples for owner/refiner/umpire, assay by ICP/XRF/fire assay; not open retail.
• Scale: Lot sizes and volumes unverified.
• Sales tape: Purchase → shipment (RCRA manifest) → sampling → assay → settlement 60–180 days (per brief; unverified) → treatment charges
• Consumption sink: Catalyst destroyed, metal recovered (score 2).
• Primary risk / boundary: Assay uncertainty is the market; RCRA K171/K172 listings and 40 CFR 266 Subpart F; months of capital lock; feedback slow.
• First experiment: Enterprise lane with a refiner or aggregator: lots received/declined, estimated vs assayed content, settlement.
• Confidence: medium (mechanics) / low (terms)
• Sources: https://www.digitalrefining.com/article/1000660/recoup-money-from-your-spent-catalyst · https://www.govinfo.gov/content/pkg/CFR-2024-title40-vol29/pdf/CFR-2024-title40-vol29-sec266-70.pdf

Equipment · proposed group: Tier 2 · Adjacent: Office clearance lots; Hotel FF&E liquidations

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 1 · 3 · 3 · 2 · 2 · 1 · 2 · 3 → structural 74.1, gates Y/P/Y/N, actionable 40.0
• Reject reasons: Time · Capacity
• Access: UsedBooths.com: sellers list free, platform brokers Q&A/transaction and offers transport; ExhibitTrader comparable; show-close teardown creates timed supply.
• Scale: 'Thousands' of booths, counters, kiosks listed; no exact counts; no price index.
• Sales tape: Buy → haul at show close → refit → resale or scrap branded skins
• Consumption sink: Durable hardware (score 1).
• Primary risk / boundary: Sporadic, event-driven supply; declined lots rarely recorded; condition and completeness unstandardised.
• First experiment: Exhibit house's 3 years of teardown buys; likely too sparse for floors — blueprint only.
• Confidence: medium (access) / low (pricing)
• Sources: https://usedbooths.com/how-it-works/

Collectibles / furniture · proposed group: Tier 2 · Adjacent: Small regional estate auctions; Office clearance lots

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 1 · 2 · 3 · 3 · 2 · 1 · 2 · 3 → structural 70.4, gates Y/P/Y/N, actionable 38.0
• Reject reasons: Attention · Capacity
• Access: Propstore auctions (26% buyer's premium, consignment rate undisclosed, seller paid 35 days post-auction); studio set sales and prop-house liquidations ad hoc; WB Property Dept is rental, not liquidation.
• Scale: No sales totals; wrap-driven supply; no price index (Propstore/Heritage results are the comps).
• Sales tape: Win → pickup at lot → resale (collectible or furniture channel) → provenance disputes
• Consumption sink: Resold furniture/collectibles (score 1).
• Primary risk / boundary: Provenance is the hidden variable (no accepted third-party authentication); sporadic supply; declined lots rarely recorded.
• First experiment: Prop dealer's 3 years of buys; blueprint only.
• Confidence: medium (Propstore) / low (studio supply)
• Sources: https://propstore.com/general-faqs/

Equipment · proposed group: Tier 4 · Adjacent: Office clearance lots

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 3 · 1 · 3 · 2 · 2 · 2 · 1 · 1 · 3 → structural 66.7, gates Y/P/Y/N, actionable 36.0
• Reject reasons: Capacity · Operational fit
• Access: Negotiated buyout/consignment with owner, GC or brand PIP team; no licence; crews, trucks, warehouse required.
• Scale: 2025 US hotel transactions 1,528 hotels / 186,925 rooms; 237 flagged for renovation; renovation cycles 5–7 yrs soft goods, 10–12 full.
• Sales tape: Phased removal → warehouse → showroom / bulk sale over months
• Consumption sink: Durable (score 1).
• Primary risk / boundary: Sporadic supply; hotel nets only $200–300/room ('not going to get rich'); declined projects rarely recorded (gate N).
• First experiment: Liquidator's project log for 3 years; probably too sparse for floors — treat as blueprint only.
• Confidence: medium-low
• Sources: https://www.costar.com/article/66442661/how-to-handle-a-hotel-liquidation-sale · https://hotelbusiness.com/u-s-hotel-transaction-volume-surges-in-2025-fueling-a-wave-of-renovation-and-conversion-activity/

Digital — platform-flagged · proposed group: Flagged · Flagged class (cf. Aged social accounts)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 0 · 3 · 3 · 2 · 3 · 3 · 2 · 3 · 1 → structural 74.1, gates Y/Y/P/P, actionable 36.0
• Reject reasons: Time · Attention
• Access: Merchant onboarding with KYC and proof of key source; commissions ~10% + fees; 30-day review hold on payouts (Kinguin).
• Scale: G2A: 30M customers, 40,000+ sellers, 75,000+ digital items.
• Sales tape: Buy → list → sale → redemption → revocation / chargeback (irreversible) → payout after hold
• Consumption sink: Keys are redeemed — consumed (score 2).
• Primary risk / boundary: Publisher EULAs, documented fraud (tinyBuild, Factorio), revocations; paper boundary: telemetry manipulated by strategic participants.
• First experiment: None — flag under API Terms.
• Confidence: medium
• Sources: https://en.wikipedia.org/wiki/G2A · https://kasepo.com/become-merhant-sell-on-kinguin/ · https://www.osborneclarke.com/insights/paris-court-rules-users-may-resell-games-purchased-steam

Printing equipment · proposed group: Industrial · Adjacent: Spare-parts stocking; Industrial surplus equipment

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 2 · 1 · 3 · 2 · 2 · 2 · 1 · 1 · 2 → structural 59.3, gates Y/P/Y/N, actionable 32.0
• Reject reasons: Scale · Attention
• Access: No dedicated marketplace; Machinio lists presses with rolls incidental; a handful of eBay listings with specs in titles (e.g., 600 LPI / 6.05 BCM); equipment dealers.
• Scale: Very thin; no counts; no pricing reference.
• Sales tape: Buy → cell inspection → resale to a matching press owner
• Consumption sink: Durable (score 1).
• Primary risk / boundary: Effectively no market — press-specific fit fragments demand; cell wear unknowable from a listing.
• First experiment: None; keep as an exploratory note.
• Confidence: medium (that it is thin)
• Sources: https://www.machinio.com/cat/anilox-rolls

Venture — securities · proposed group: Flagged · Separate perimeter class (like Alt-coin hedge fund / Thin public equities)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 0 · 2 · 2 · 3 · 3 · 3 · 3 · 1 · 1 → structural 66.7, gates Y/Y/Y/Y, actionable 26.7
• Reject reasons: Prediction · Attention
• Access: Reg CF open to all investors with income/net-worth caps; AngelList syndicates accredited-only (SPV min raise $80k, setup ~$10k, carry ≤20%, close in 2–4 weeks).
• Scale: Reg CF 2025: $378.3M raised, 1,006 new offerings (−29%), 67.4% hit target, median equity raise $194k, avg 285 investors; AngelList claims 25,000 syndicates & funds.
• Sales tape: Invest → 12-month resale restriction → follow-on rounds / failure over 5–10 years
• Consumption sink: The user's own example class: startups die (score 3).
• Primary risk / boundary: Securities perimeter (excluded from membership/profit-share); feedback arrives years after regime change (paper boundary); low ops leverage.
• First experiment: Research-only: backtest KingsCrowd/EDGAR 2016–2020 cohorts for follow-on/failure; do not operate under membership.
• Confidence: high (Reg CF) / medium (AngelList)
• Sources: https://kingscrowd.com/2025-investment-crowdfunding-annual-report/ · https://www.angellist.com/syndicates · https://www.joinarc.com/guides/angel-syndicate

IP finance · proposed group: Tier 3 · Adjacent: Kindle and audiobook backlist rights (Tier 3)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 0 · 1 · 3 · 3 · 2 · 3 · 1 · 1 · 1 → structural 55.6, gates Y/Y/Y/Y, actionable 22.2
• Reject reasons: Prediction · Attention
• Access: Royalty Exchange: identity verification only, non-US OK, timed auctions <1 week, ACH/wire in 2 days; SongVest = Reg A+ securities; Spotter = bilateral advances (not an open menu).
• Scale: Royalty Exchange cumulative $200M / 2,000+ transactions, 30k investors (listings/month ~20–40, unverified); SongVest 185 offerings.
• Sales tape: Assignment (~1 month) → quarterly/biannual distributions → decay
• Consumption sink: Catalogs decay slowly (score 1).
• Primary risk / boundary: Securities status ambiguous (SongVest explicitly securities); thin flow (≈1/day); quarterly feedback.
• First experiment: Research-only backtest of Royalty Exchange archive vs subsequent distributions.
• Confidence: medium-high
• Sources: https://royaltyexchange.com/faq · https://altstreet.investments/platforms/reviews/songvest · https://www.spotter.com/creator-capital

Travel capacity · proposed group: Flagged · Adjacent: Reservation arbitrage (flagged)

• Scores (logistics · density · access · long-tail · breadth · partial obs. · consumption · cadence · ops): 0 · 3 · 0 · 2 · 3 · 2 · 3 · 3 · 1 → structural 63.0, gates Y/Y/P/P, actionable 5.1
• Reject reasons: Time · Attention
• Access: Strictly B2B: Hotelbeds API needs signed commercial agreement + certification; Expedia TAAP needs IATA/ARC/CLIA; wholesale contracts prohibit public resale of net rates — no evidence small operators can buy allotments to resell.
• Scale: Hotelbeds 180k hotels; WebBeds 365k+; hotel × date × room type = millions of cells.
• Sales tape: Book → stay → cancellations/no-shows/chargebacks → monthly commission
• Consumption sink: Room-nights perish (score 3).
• Primary risk / boundary: Access effectively closed (contract bans) → boundary example: perfect menu, inaccessible.
• First experiment: None for membership; only as an enterprise lane with an accredited agency.
• Confidence: medium
• Sources: https://hoteltechreport.com/news/hotel-bed-banks · https://developer.hotelbeds.com/documentation/activities/knowledge-base/certification/