Benchmarks · asset-heavy operators · North America

How does your operation stack up against operators your shape?

Allometry is named for comparative biology — how parts scale against the whole. Labs is that idea applied to a market we counted one operator at a time: 10,053 North-American asset-heavy operators, enriched and scored individually. Market-structure benchmarks are live below. Operating-performance benchmarks publish vertical-by-vertical as the attested cohort matures — and we state plainly which is which.

10,053
operators counted & ICP-scored, one by one
$1.05T
combined customer revenue on the counted list
$70M
median operator revenue
100
operator interviews behind the metric set
01 · Choose your vertical

Built from a counted market, not a Gartner percentage.

Every card below is a named cohort inside the 10,053. Market structure is published today; performance benchmarks open per vertical once ten or more operators have twelve months of attested ledger behind them.

Pallets & Packaging
Programs, repair loops, per-unit margin against materials that move monthly.
Market structure →
HVAC & Mechanical
Install + service mix, maintenance entitlements, seasonal demand shape.
Market structure →
Solar & Renewables
Deployment economics per address, O&M contracts, long-dated service revenue.
Market structure →
Smart Lockers & Kiosks
Installed-address density, per-unit revenue, network expansion economics.
Market structure →
Elevators & Building Systems
Service portfolios, modernization pipelines, entitlement drift.
Market structure →
Waste & Recycling
Route density, container economics, contract-vs-invoice reconciliation.
Market structure →
EV Charging
Per-address utilisation, deployment payback, network occupancy.
Market structure →
3PL & Logistics
Per-lane and per-account margin, capacity commitments, demand shape.
Market structure →
Manufacturing — industrial mid-market
SKU-level morphology, quote discipline, oversold exposure.
Market structure →
Field Services & Multi-Site
Dispatch economics, SLA performance, revenue per served address.
Market structure →
02 · The counted market

What the list looks like today.

Enriched from a curated export and tiered by fit; the same filters on a second data universe return roughly 2× the count, so treat the list as the floor of the reachable market, not its ceiling.

TierWhat it meansOperatorsProvenance
VelocityFast-moving fit — the volume motion7,839counted
CoreThe center of the ICP — $20–200M asset-heavy operators1,919counted
StrategicLarge, multi-entity, portfolio-shaped295counted
Top 500Prioritized cohort — where outbound runs first500counted
03 · What we benchmark

The morphology metrics — shape, not vanity.

Marketing benchmarks measure spend. These measure the operation itself — and because they derive from an append-only, attested ledger, a benchmark here is a number a lender can also price.

MetricWhat it tells youStatus today
Allometric coefficient · per SKU classHow each revenue line scales against the whole book — the claw that outgrows the crab.live at n=1
Realised price vs listWhat discounting actually costs, per line, per customer.live at n=1
Floor-breach rateShare of accepted quotes below their own SKU floor — margin given away silently.sample workspace
Oversold exposureQuantity committed beyond available stock, inside the lead time.live at n=1
DSO & AR healthCash trapped past terms — the predicate a lender verifies without opening the books.sample workspace
Entitlement leakageContracted-but-never-invoiced service — the maintenance revenue that quietly vanishes.sample workspace

The publishing rule — stated in advance

Market-structure benchmarks (counts, tiers, revenue distributions) are published now: they come from a list we enriched one operator at a time, and they are labeled counted.

Operating-performance benchmarks publish per vertical only when ten or more operators have twelve or more months of attested ledger — enough for a cohort statistic that survives diligence. Today the attested cohort is one design partner (its result: $300K of unflagged margin loss surfaced, $100K captured — measured, n=1 and stated as such) plus a generated sample workspace whose figures are computed by the live engines (computed).

We would rather show you an honest n than a confident average of nothing. That is the same doctrine the product runs on: a number with missing inputs returns null, never a flattering estimate.

Want your number, not the cohort’s?