Where the data comes from and what we do with it.
Trust-building starts with transparency. This page describes every data source GrowthFactor uses, how each is processed, and what the model does and does not claim about its outputs.
Data Sources
What feeds the model
Three distinct data categories, each with different update cadences and known coverage limitations.
Foot Traffic Panel
Aggregated mobility data from opted-in consumer devices via two licensed data partners. Panel is normalized for demographic bias and daypart seasonality before ingestion. Coverage: 98% of US census tracts. Updated weekly.
Spending Category Data
Credit and debit transaction aggregates by NAICS retail subcategory, licensed at zip-code level from a commercial financial data provider. Refreshed monthly. Used for demand-gap analysis within defined trade areas.
Census and ACS Base Layer
US Census Bureau data (5-year ACS estimates) for demographic and income segmentation at the tract level. Used as a base layer for trade area characterization, not as a primary signal.
Demand Modeling
Trade areas and demand gaps
Population density does not equal demand. GrowthFactor delineates trade areas using drive-time and walk-time isochrones rather than fixed-radius circles.
Trade Area Delineation
Five-minute, ten-minute, and twenty-minute drive-time polygons computed from routing data, adjusted for urban street networks and transit patterns.
Category Affinity Scoring
Spending category data is matched to your retail type. Affinity indices surface whether the population in a given trade area over-indexes on your category relative to the national average.
Demand Gap Analysis
Category supply (existing retailer count and estimated capacity) subtracted from category demand (spending volume) reveals trade areas where demand is unmet by current supply.
Cannibalization Model
How we model sales transfer
Cannibalization is inevitable when a chain opens near an existing location. The question is whether the math still works. GrowthFactor quantifies both the probability and the expected magnitude.
Customer Gravity Model
Based on Huff's gravity model, calibrated on observed customer travel patterns from the foot traffic panel. Attraction is a function of store size proxy and travel time resistance.
Overlap Weighting
Trade area overlap between proposed and existing locations is weighted by actual visit origin density. Dense overlap with a top-quartile existing store triggers a cannibalization flag regardless of aggregate score.
Conservative and Aggressive Scenarios
Each report includes a base-case cannibalization estimate alongside a conservative (higher transfer) scenario. The overall GF Score uses the base case; the conservative scenario is disclosed in the breakdown.
Calibration and Honesty
What the model does and does not claim
A GF Score of 82 does not mean a store will succeed. It means that location outperforms 82 percent of evaluated candidates on the signals we measure. This is a materially different claim, and we keep it precise on purpose.
Score Calibration
Scores are calibrated against first-year revenue outcomes from early-access pilot partners. We disclose the calibration sample size and the margin of each pilot's outcome relative to the score prediction.
Data Freshness Disclosure
Every report states the data vintage for each input. If foot traffic data for a specific census tract is more than 90 days old, the report flags it and adjusts the signal confidence accordingly.
Coverage Gaps
Panel coverage drops below 90% in some rural census tracts. We do not issue scores for locations where foot traffic panel coverage falls below our minimum threshold. We tell you why a score is unavailable rather than issue a low-confidence number.
Ask us about data coverage in your target markets.
Coverage varies by region and trade area density. We will tell you upfront what the score confidence looks like for your specific candidate markets.
Request a Demo