3.9M
Manhattan daytime population
Census residential count: 1.6M
The problem
13,000 people work at LaGuardia every day.
The census sees zero.
Census data tells you where people sleep. But the economy doesn't happen at 3 AM. Manhattan's daytime population is 3.9 million - the residential count says 1.6 million. Airports, office districts, retail corridors - the places that matter most are invisible.
13,000
LaGuardia daily workers
Census residential count: near zero
~60%
of economic activity
Happens outside residential zones
The proof
We predicted outcomes no one thought possible -
without observing a single one.
No sales data. No fare cards. No incident reports. No foot traffic. Just the structural physics of who is where, when, and why - derived from government records and first principles.
92%
We predicted a restaurant's daily revenue
Without ever seeing a single transaction. No sales history. No POS. No receipts. No foot traffic.
96%
We predicted subway ridership
Without a single fare card swipe, turnstile count, or rider survey.
82%
We predicted traffic congestion
Who is stuck in traffic, where, and why - without a single GPS trace.
3 hrs
We see tonight's dinner rush before it happens
Highway traffic 40 km away signals restaurant demand three hours ahead.
Paradigm shift
Observation-based intelligence decays.
Structural intelligence compounds.
Privacy pressure keeps eroding panel-based systems. Our model improves as more geographies are added.
Legacy model
Observation stack
- Depends on device IDs and panel coverage.
- Degrades as ATT/GDPR/CCPA tighten.
- Needs constant re-calibration.
Ether model
Structural stack
- Built from government records and first principles.
- Gets stronger as privacy restrictions increase.
- Compounds with every new city and validation cycle.
The category was built on tracking people. We proved tracking was never required.
How it works
The structural physics
of economic geography.
Federal employment records tell us where people work. Time-use physics tells us when. Spatial gravity tells us how places pull. No phones required - just the math of how economies move.
Every hex, every hour
Daytime Population
Where people actually are during business hours. Not where they sleep - where they work, eat, shop, and commute.
The physics of foot traffic
Economic Gravity
Which places pull people in - and from where. Hub scores, commuter flows, and economic adjacency that shapes local markets.
Industry × time × place
Workforce Composition
What industries are present in every location, by hour. Finance at lunch, healthcare at shift change, retail on weekends.
Privacy isn't our constraint.
It's our advantage.
Every privacy regulation - ATT, GDPR, CCPA, whatever comes next - makes mobile panels less accurate. It makes us more valuable. We never needed device IDs. The signal was always in the structure.
Zero device IDs
Zero SDKs
Government-sourced
Immune to ATT
Immune to GDPR
Gets stronger as privacy tightens
Who this is for
Five industries. One blind spot.
We fix it for all of them.
01
Out-of-Home Media
See which screens face finance workers at lunch
“You're buying billboards by ZIP code.”
DOOH at $8-15 CPM can outperform $150 LinkedIn if you can prove the audience is there.
02
Commercial Real Estate
See the workforce before you sign the lease
“$50M lease decisions on 5-year-old Census data.”
Sub-block worker density by industry and commuter pull gives site selection real structural context.
03
Marketing Science & MMM
Stabilize your marketing mix model
“Your geo-prior is unstable. ROAS changes every run.”
Government-sourced population layers anchor experiments with a stable structural prior.
04
Healthcare Network Planning
Plan for where people actually need care
“Access models measure from where patients sleep.”
Daytime population reveals worker-heavy geographies invisible to residential-only models.
05
Campaign Intelligence
Reverse any campaign back to real cohorts
“You have impressions by geo and hour. But who was there?”
Turn impression logs into audience profiles by industry, income, and commute pattern without device IDs.
Proof points
Built on measured outcomes, not claims.
Validated metric
92%
Revenue prediction accuracy
Predicted without direct transaction feeds.
Validated metric
99%
Daytime population accuracy
Validated against known high-density zones.
Validated metric
1,000+
Signals per location
Multi-domain structural features in every cell.
Validated metric
1 query
Plug into warehouse instantly
SQL-first activation in BigQuery.
Start with Manhattan. It's free.
Manhattan at H3 resolution 8 - 3,360 cells, free in BigQuery, no signup required. The full structural layer - daytime population, economic gravity, and NAICS hourly composition - is available on request.