Motionworks Panel and Sampling Methodology

A Privacy-Conscious Approach to Scalable Mobility Insights

Motionworks combines Census-scale coverage with advanced machine learning, multi-source validation, and privacy-by-design architecture to deliver population insights.

Our methodology provides coverage of 13.5% of U.S. adults for the Geopath 2025 Forecast Period—surpassing government-grade sample density while providing the temporal resolution and behavioral depth that traditional surveys cannot achieve.

📘

Why it matters

Motionworks provides real-world population behavior insights—at massive scale and with unmatched privacy safeguards—so businesses can make smarter, faster decisions without relying on guesswork or outdated assumptions.

Our Data Foundation: Three Complementary Sources, One Unified Signal

Motionworks' robustness begins with our multi-source approach. We integrate three distinct commercial data feeds, each optimized for different strengths, processing data from over 150 million devices weekly:

Maximum Coverage
Maximizes geographic and device coverage across diverse populations and regions

Maximum Consistency
Prioritizes longitudinal continuity with devices present week over week for trend analysis

Maximum Observations
Opt-in feed with the highest granularity from privacy-consented data sources

Our proprietary reconciliation process merges these feeds at the device level, filtering unreliable locations and cataloging the best-known locations throughout each day. This creates a unified dataset that maximizes both coverage and accuracy—a fundamental advantage over single-source methodologies.

Vehicle Telematics Layer
For traffic volume estimation, we have an exclusive partnership with Streetlight Data to supplement our core mobility feeds with their specialized Aggregated GPS (AGPS) data that has been optimized for road segment coverage. This additional layer provides:

• Navigation-grade precision from turn-by-turn routing applications
• Comprehensive road network coverage including rural and low-volume segments
• Validated penetration rates averaging 27% of actual of vehicles in use
• Machine learning enhancements through integration with permanent counter ground truth data

This AGPS integration enables Pathcast products to deliver traffic volume insights across the complete U.S. road network.

📘

Why it matters

To get an accurate picture of where people go—and when, you can't use a single data source. We bring together the best available to capture mobility patterns with unmatched breadth and reliability while reducing blind spots and noise.

Algorithmic Sophistication

Machine Learning That Learns as People Move

We use Bayesian machine learning to dynamically update hourly location profiles per device. This allows our models to:

• Separate signal from noise
• Adjust in real-time to behavior shifts
• Detect events or anomalies as they unfold

Bayesian Machine Learning at Scale: Our algorithms dynamically evolve hourly behavioral profiles for each location based on continuous observations. This approach allows us to distinguish signal from noise while adapting to changing patterns and identifying events—something static sampling methodologies cannot achieve.

📘

Why it matters

Static panels can't keep up with a dynamic world. Ours does. This means planners can optimize advertising campaigns mid-flight, data analysts can understand trends as they happen, and brands can move from reactive to predictive.

Panel Architecture and Products

Motionworks operates multiple specialized panels, each designed for specific analytical needs:

Longitudinal Consistency Drives Measurement Confidence

The true power of our panels lies in their temporal persistence, enabling accurate reach and frequency measurement:

• BroadSelect: While maximizing reach, 10% of devices are observed for 4+ weeks monthly, and 7% for 10+ weeks annually
• PrimeSelect: Optimized for measurement consistency, 34% of devices are observed for 4+ weeks monthly, and 33% for 10+ weeks annually
• BroadRoute: Optimized for seasonal and weather-sensitive algorithms for accurate year-round modeling to provide consistent historical analysis from 2019 to today

This longitudinal consistency—particularly in PrimeSelect—enables true observed frequency measurement and deduplicated reach estimation across multi-week campaigns, providing the statistical foundation for OOH television ratings and cross-channel attribution.

📘

Why it matters

Marketers use BroadSelect for reach modeling and audience profiling.

PrimeSelect powers deduplicated reach and frequency metrics for attribution.

BroadRoute delivers comprehensive movement data at the highest resolution possible.

Our opt-in panel validates results and augments with qualitative insights.

With this persistence, we don't model frequency—we observe it. That's the backbone of our measurement data products.

Sampling Power in Context

To understand Motionworks' statistical power, it's essential to benchmark against established standards:

Quality Assurance and Bias Mitigation

Ensuring our panel data is stable and useful.

Inclusion Criteria

We only include devices that meet strict behavioral thresholds to ensure statistical reliability. For a device to be included in our panel it must:

• Have a valid household type weighted for representation
• Be observable for 8+ hours per day, on average
• Be active for 7 days a week
• Have valid trip behavior patterns (i.e., not stationary, spoofed, or erratic)

Bias Mitigation Strategies

To maintain fairness and accuracy, Motionworks aims to reduce the impact of any potential inherent biases in our data by looking at:

• Edge-case exclusions: We remove data from long-haul truckers, gig workers, and nomadic patterns that would otherwise distort results.

• Peer group validation: Behavioral norms are identified per place-type—so for example, a nightclub and a sports bar are analyzed through different behavioral lenses.

• Post-stratification weighting: Weekly adjustments ensure the panel matches U.S. demographics at the county level using the Motionworks Digital Population updated annually using Claritas PopFacts.

• Cross-validation: Our results are continuously running through stratified validation processes ensuring model reliability across place types, road volume categories, and market segmnets where ground truth data sources where available at scale, including Roadway Traffic Counts, Public Transit and Airport place types, among partnerships with other sensor data companies.

📘

Why it matters

Big data is only powerful if it's trustworthy. Our panels aren't just large—they're rigorously validated for representativeness and accuracy to make sure we're reflecting real people's real lives.

Temporal Scalability and Statistical Power

Motionworks' longitudinal panel design enables multi-resolution insights:

Data Resolution Capabilities:

• Daily/Weekly: Variable coverage for event detection and traffic and travel metrics
• Monthly: 2.43% sample coverage, enabling campaign measurement and trend analysis
• Annual: 13.5% coverage of U.S. adults for strategic planning and market analysis

The power of our approach is cumulative: while any individual week provides valuable directional data, the longitudinal structure builds into massive population-level understanding over months and years.

📘

Why it matters

The more time we observe, the stronger the insights become. Whether you need daily event detection or multi-month planning inputs, Motionworks provides the resolution you need—with statistical power that builds over time to deliver the strategic intelligence our customers need.

Privacy by Design- Not an Afterthought

Motionworks does not receive any personally identifying information in its data feeds, and does not associate demographic data at the device level. Motionworks expects our data providers to be complying with relevant regulations before the data reaches us. Going above and beyond, Motionworks also has the following privacy focused methods:

Aggregation Thresholds
Preventing small-group identification through minimum sample requirements

Temporal Obfuscation
Strategic data generalization in reporting to prevent the possibility of individual tracking

No Persistent Tracking
Individual devices are never analyzed or tracked across campaigns or client applications. All reporting is at a full population level that cannot be reverse engineered.

Regulatory Compliance
Continuous adaptation to evolving privacy regulations and best practices

📘

Why it matters

Consumer trust is non-negotiable. Our products are designed to deliver insights without compromising individual privacy. Our customers get population-level intelligence with zero risk of personal targeting or misuse.

What Makes Motionworks Different

What sets Motionworks apart from traditional sampling methodologies:

1. Census-scale reach + Machine Learning accuracy + Complete road network coverage
2. Three-source validation for unmatched robustness
3. Behavioral depth from actual observed frequency
4. Daily-to-annual reporting flexibility
5. No compromise privacy architecture
6. Models that improve automatically with time

📘

Why it matters

Traditional sampling is static, slow, and siloed. Ours is dynamic, privacy-safe, and built for today's decision-makers.

Methodology Note: All sampling rates are calculated based on deduplicated device counts mapped to person-equivalent samples using industry-standard device-to-person ratios. Panel composition is continuously monitored and adjusted to maintain demographic representativeness.