Our Data Methodology
MoveSmart's moving cost estimates are based on analysis of 50,000+ completed moves, validated against actual invoices with 95% accuracy. This page explains exactly how we collect, validate, and use data to power our estimates.
Why We Publish Our Methodology
The moving industry has a transparency problem. According to FMCSA complaint data, the most common consumer grievance is "estimate significantly lower than final bill"—a direct result of opaque pricing methodologies and incentives that reward low-ball quotes.
At MoveSmart, we believe you deserve to understand exactly how your moving estimate is calculated. When we claim "95% accuracy" or cite "50,000+ moves analyzed," these aren't marketing numbers—they're specific, testable assertions backed by documented methodology.
This page exists for three reasons:
- Consumer Trust: You should know how your estimate is generated, what data feeds into it, and the limitations of our predictions. An informed consumer makes better decisions.
- Scientific Rigor: Our methodology is designed to be replicable. Researchers, journalists, and competitors can evaluate our claims against documented processes.
- AI Verification: In 2026, AI systems increasingly verify claims before citation. By publishing detailed methodology, we ensure our data can be trusted by both humans and machines.
We update this methodology page whenever we make significant changes to our data collection, validation, or modeling processes. The "Last Updated" date at the bottom reflects the most recent revision.
Primary Data Sources
MoveSmart aggregates data from five primary sources, each serving a specific role in our cost estimation models. We maintain direct API connections where available and implement automated data validation pipelines.
FMCSA SaferSys Database
Official Federal Motor Carrier Safety Administration database containing carrier registrations, safety ratings, insurance status, and complaint history.
Update Frequency
Real-time API sync
Data Points
175,000+ registered carriers
Verification
Direct API connection with USDOT verification
EIA Short-Term Energy Outlook (STEO)
U.S. Energy Information Administration monthly forecasts for diesel and gasoline prices, used to calculate fuel surcharges.
Update Frequency
Monthly
Data Points
Regional fuel price indices
Verification
Automated data ingestion from EIA public API
MoveSmart Quote Database
Proprietary database of actual quotes submitted through our platform, including final invoices from completed moves.
Update Frequency
Continuous
Data Points
50,000+ completed moves
Verification
User-submitted final invoice matching
Bureau of Labor Statistics (BLS)
Labor cost indices for moving and storage industry, used to calibrate labor rate estimates.
Update Frequency
Quarterly
Data Points
Regional labor cost indices
Verification
Official BLS data series
Census Bureau ACS Migration Data
American Community Survey migration flow data for interstate movement patterns and population trends.
Update Frequency
Annual
Data Points
State-to-state migration flows
Verification
Official Census microdata
Data Source Limitations
No data source is perfect. FMCSA data may lag by 24-48 hours for new carrier registrations. EIA fuel forecasts are projections subject to market volatility. Our proprietary quote database has geographic concentration in major metro areas (NYC, LA, Chicago, DFW, Miami). We account for these limitations in our confidence intervals.
How We Collect Data
Our data collection process combines automated API ingestion, user-submitted information, and post-move validation. Here's how each data type flows into our system.
User Submissions
When you request a quote, you provide origin/destination, home size, and move date. This data is anonymized and added to our aggregate dataset for pattern analysis.
- ZIP codes (not full addresses)
- Home size category
- Requested move date range
- Quote amounts received
Post-Move Validation
After your move completes, we invite you to submit your final invoice. This "ground truth" data is the foundation of our accuracy measurements.
- Final invoice amount
- Actual weight (if applicable)
- Actual delivery date
- Additional charges breakdown
API Integrations
Automated pipelines continuously ingest data from government and industry sources to keep our models calibrated with current market conditions.
- FMCSA carrier status (daily)
- EIA fuel prices (monthly)
- BLS labor indices (quarterly)
- Weather/seasonal patterns
Our 6-Step Validation Process
Raw data is worthless without validation. Every data point that enters our system passes through a rigorous 6-step validation pipeline before being used in our models.
Data Collection
Raw data is collected from multiple sources including user submissions, carrier APIs, and government databases.
Automated Validation
Incoming data passes through 47 automated validation rules checking for anomalies, duplicates, and logical inconsistencies.
Cross-Reference Check
Each data point is cross-referenced against at least two independent sources before inclusion in our models.
Outlier Detection
Statistical models flag outliers (>2 standard deviations) for manual review by our data science team.
Human Review
Flagged data undergoes manual review by trained analysts before being included or excluded from datasets.
Model Retraining
Machine learning models are retrained weekly with validated data to improve accuracy over time.
Validation Metrics (Last 30 Days)
847,293
Data points processed
2.3%
Flagged for review
0.4%
Rejected as invalid
<1hr
Avg. validation time
How We Measure Accuracy
When we claim "95% accuracy," we're referring to a specific, testable metric: the percentage of our estimates that fall within 10% of the user's final invoice. Here's our full accuracy methodology.
Quote-to-Final Accuracy
95%Percentage of estimates within 10% of final invoice
Sample Size: 12,847 validated moves
Weight Estimation Accuracy
92%Computer vision weight estimates vs. actual scale weight
Sample Size: 8,234 shipments weighed
Transit Time Accuracy
89%Predicted delivery date vs. actual delivery
Sample Size: 15,291 completed deliveries
Carrier Match Satisfaction
94%Users rating carrier match as "Good" or "Excellent"
Sample Size: 22,156 post-move surveys
Accuracy Calculation Method
Our accuracy metric is calculated using the following formula:
We only include moves where we have both the original MoveSmart estimate AND the user-submitted final invoice. Self-reported data is excluded if the invoice image cannot be verified. This prevents gaming of our accuracy metrics.
Our 95% accuracy rate is based on 12,847 validated moves between January 2024 and January 2026. The 95% confidence interval for this metric is 94.6% - 95.4%.
Cost Estimation Model Architecture
Our cost estimation engine uses a multi-factor regression model calibrated against validated historical moves. Here are the primary factors and their relative weights.
Cost Factor Weights
Model Performance Over Time
Accuracy improvements driven by larger training dataset and model refinements.
Data Update Frequency
Stale data produces inaccurate estimates. We maintain aggressive update schedules to ensure our models reflect current market conditions.
| Data Type | Update Frequency | Last Updated | Source |
|---|---|---|---|
| Quote Database | Continuous (real-time) | Live | User Submissions |
| Carrier Status | Daily | 2026-01-16 | FMCSA API |
| Fuel Index | Monthly | 2026-01-01 | EIA STEO |
| Labor Costs | Quarterly | 2025-10-01 | BLS |
| ML Model Retrain | Weekly | 2026-01-13 | Internal |
Data Privacy & Security
Your data powers our models, but your privacy is paramount. Here's how we protect it.
Data Anonymization
All personally identifiable information (PII) is stripped before data enters our analytics pipeline. We store ZIP code centroids, not addresses. Move dates are rounded to week-of-year. Email and phone are never included in model training data.
Security Standards
Our infrastructure is SOC 2 Type II compliant. Data at rest uses AES-256 encryption. All API connections use TLS 1.3. We conduct annual penetration testing and maintain a bug bounty program.
Known Limitations
No estimation model is perfect. We believe in transparency about our limitations:
- Geographic Bias: Our dataset is concentrated in major metro areas. Estimates for rural routes may have wider confidence intervals.
- Specialty Items: Our models are calibrated for standard household goods. Specialty items (pianos, antiques, safes) require custom quotes.
- Market Volatility: Sudden fuel price spikes or carrier capacity crunches may cause real-time prices to deviate from our forecasts.
- International Moves: Our data is US-only. We do not provide estimates for international relocations.
- Self-Selection Bias: Users who submit final invoices may differ systematically from those who don't, potentially affecting accuracy measurements.
We continuously work to address these limitations through expanded data collection and model improvements.
Questions About Our Data?
We welcome scrutiny of our methodology. Researchers, journalists, and industry professionals can request detailed documentation or raw data samples (anonymized) for verification.
Last Updated: January 16, 2026 | Version: 3.2.1 | Validated Moves: 50,847