Data · technology · limitations
How METER Measures Audiences
What data goes into the calculation, where observation ends and modeling begins, and which limitations must accompany the result.
No single source measures the entire audience
Each layer answers only one part of the question. For every source, it is therefore important to show not only its value, but also the limits of what it can prove.
Mobility data
- What it provides
- The time, coordinates, accuracy, and sequence of anonymized signals. Depending on the market, SDK/GPS data and available aggregates may be used.
- How it is used
- Route reconstruction, direction, speed, dwell time, and the creation of contact sessions.
- What it does not prove
- A signal does not prove a view. One device does not always equal one person and cannot represent the population without a panel and weighting.
Advertising assets
- What it provides
- The coordinates, side, direction, format, size, height, type, and operating mode of each asset.
- How it is used
- An inventory profile, an asset model, and individual visibility geometry.
- What it does not prove
- Metadata does not confirm that an ad was actually played or that an audience was present nearby.
Road network and maps
- What it provides
- Road geometry, pedestrian zones, buildings, transport context, and available information about obstructions.
- How it is used
- To check the physical plausibility of a route, direction of travel, and viewing conditions.
- What it does not prove
- A map does not prove the actual route and may not reflect roadworks or temporary changes in the environment.
Points of interest
- What it provides
- Shops, offices, transport hubs, restaurants, and other places that describe the context of an area.
- How it is used
- To analyze the purpose of locations, visits, and available subsequent actions.
- What it does not prove
- A POI visit alone does not prove exposure to an ad or a causal campaign effect.
Campaign parameters
- What it provides
- Period, schedule, SOV, ad loop, creative duration, operating hours, and uptime.
- How it is used
- To move from the overall potential of an asset to the share of audience available to a specific brand.
- What it does not prove
- Planned parameters do not confirm actual delivery and may differ from the placement that took place.
Proof of Play and monitoring
- What it provides
- Playback logs, player statuses, photos, videos, and other verifiable evidence of delivery.
- How it is used
- To align a campaign with time and verify that the required creative was available on the media asset.
- What it does not prove
- PoP does not measure a person’s presence, gaze direction, unique Reach, or advertising effect.
Statistics and external sources
- What it provides
- Population data, traffic flows, sensors, surveys, and available external observations.
- How it is used
- Universe definition, statistical weighting, calibration, and additional validation of results.
- What it does not prove
- Source definitions, periods, and geographies may not match. They do not replace the logic of an advertising contact.
First-party outcomes
- What it provides
- CRM, POS, visits, QR, website, app, loyalty, and other outcomes when they are lawfully available.
- How it is used
- To assess subsequent actions, attribution, and, with an appropriate study design, incrementality.
- What it does not prove
- A match after contact does not prove causality. A valid window, control, and lawful linkage are required.
11 steps from input data to reporting
Raw location signals are not treated as an audience until contact qualification, deduplication, weighting, and projection have been completed.
Data and panel
-
01 Observed fact
Quality control
Invalid coordinates, duplicates, anomalous movements, and suspicious signal concentrations are removed.
-
02 Model
Movement reconstruction
Signals are connected into time sequences and plausible routes, taking roads, speed, and direction into account.
-
03 Model
Research panel
Devices with a sufficiently long and stable history are selected for analysis; available attributes are assessed in aggregate.
Inventory and contact
-
04 Observed fact + calculation
Inventory model
Coordinates, side, direction, format, size, height, operating mode, and road geometry are taken into account.
-
05 Calculation + model
Visibility zone
Individual viewing geometry is created for each asset instead of applying the same circular radius to all assets.
-
06 Calculation
Contact sessions
A route is qualified by distance, angle, direction, and time; nearby signals are grouped into a single session.
Metrics and projection
-
07 Calculation + model
OTS and LTS
OTS counts potential contacts; LTS refines them using visibility factors and available campaign parameters.
-
08 Calculation
Deduplication
Repeat contacts are consolidated at the level of the observed participant, advertising assets, and campaign period.
-
09 Model
Weights and Universe
Statistical weights translate panel observations into an estimate of the selected target population.
Campaign and results
-
10 Observed fact + calculation
Campaign layer
Period, operating hours, SOV, loop, uptime, and available PoP narrow the potential down to a specific brand.
-
11 Result
Reporting and diagnostics
OTS/LTS, Reach 1+/N+, Frequency, GRP/TRP, asset contribution, maps, and methodology disclosures are produced.
On the web page, the nine stages of the original methodology are expanded into eleven: the inventory model and visibility zone, as well as deduplication and projection, are shown separately. The calculation sequence remains unchanged.
What is observed, what is calculated, and what is modeled
Professional measurement shows how a result is derived. It is not a claim that “we see everyone,” but a reproducible chain from observations to estimates.
| Observed fact | Calculation | Model-based estimate |
|---|---|---|
| Location signal with coordinates and time | Distance to the asset | Route between sparse signals |
| PoP record or player status | Route intersection with the zone | Probability of seeing the message, LTS |
| Asset coordinates and metadata | Time spent in the zone | Behavior-based audience profile |
| Contacts of an observed panel participant | Sessions and unique IDs within the panel | Statistical weight and projection to the Universe |
| Observed POI visit in available data | Reach N+ within the panel | Incremental effect under incomplete observation |
Uncertainty cannot be eliminated with a single accuracy figure
It can be managed, tested, and disclosed at every level of the calculation.
Input data
Inaccurate coordinates, asset azimuths and dimensions, and gaps in signals or logs directly affect the calculation.
Panel bias
Observed participants may differ from the population in geography, behavior, and devices used.
Device and person
One device does not always equal one person, and one person may use multiple devices.
Environmental changes
Vegetation, barriers, vehicles, and other temporary obstructions may not be fully captured.
Temporal variability
Seasonality, events, weather, and roadworks change mobility and contact conditions.
Model uncertainty
Profiles, LTS, projection, and attribution depend on data, assumptions, and available controls.
Six mandatory checks
-
01
Period, territory, Universe, and target filters
-
02
Metric basis: OTS, LTS, or Audience
-
03
Data sources and the period for which they are current
-
04
Rules for visibility, sessions, deduplication, and projection
-
05
Actual campaign parameters and PoP availability
-
06
Excluded periods and indications of incomplete data
Aggregate measurement, not individual-level reporting
Identifiers and location signals are processed in accordance with applicable legal bases, supplier agreements, and the requirements of the relevant jurisdiction.
Operational identifiers must not become names or individual profiles in client-facing results.
Audience metrics and characteristics are delivered at the level of groups, campaigns, and the selected Universe.
The available dataset and procedures depend on legal conditions, consent, and supplier requirements.
The term “anonymized data” does not eliminate the need for access controls and an assessment of re-identification risk.
Short answers to questions about data and methodology
Definitions link to the glossary, campaign questions link to the pre-buy and post-buy material, and technical answers are explained on this page.
About metrics
How is Reach different from OTS?+
OTS counts all qualified contacts, including repeats. Reach counts each person once within the specified period.
Learn more about OTS and Reach →Can OTS be treated as views?+
No. OTS means an opportunity to see the message. A confirmed view requires separate attention measurement.
OTS definition →Can GRP exceed 100?+
Yes. GRP includes repeat contacts. For example, 50% Reach at an average Frequency of 3 produces 150 GRP.
How GRP is calculated →What is LTS?+
LTS refines OTS using visibility and movement factors. It is a model-based estimate, not a confirmed view.
LTS definition →About data
Does one phone equal one person?+
Not always. This is why the rules for panel construction, deduplication, and conversion from devices to people estimates must be disclosed separately.
View limitations →How accurate is mobile data?+
It depends on the source, market, signal density, device, and environment. Reliability comes from filtering, reconstruction, calibration, and disclosure of uncertainty—not from a single universal figure.
About validation →What sources does METER use?+
The mix depends on the market and legal conditions. It typically combines mobility, inventory, maps, POI, campaign parameters, PoP, statistics, and available first-party outcomes.
All sources →Can audiences be measured in real time?+
Delivery and screen operation can often be monitored in near real time. Audience metrics are updated after the delay required for data availability and quality control.
How the result is produced →About campaigns
What does Proof of Play confirm?+
PoP confirms playback or placement at a specific time. It does not prove a person’s presence, a view, unique Reach, or an effect.
The role of Proof of Play →Can visits and sales be measured?+
Subsequent actions can be assessed when suitable first-party or location data is available. Causal lift requires a control group, match-market approach, or another incrementality design.
First-party outcomes →Why can’t Reach figures for different screens be added together?+
One person may come into contact with several screens. The overlap increases Frequency but does not create a new person in campaign Reach.
Deduplication example →How should pre-buy and post-buy be compared?+
Use the same territory, period, Universe, target filters, and contact basis, then account for actual uptime, SOV, substitutions, and available PoP.
Campaign stages →About vendor selection
What should you request from a measurement provider?+
The visibility-zone methodology; OTS/LTS/Audience basis; data sources and periods; Universe; rules for the panel, deduplication, weights, PoP, and validation; and a list of limitations.
Report checklist →How can you tell whether a report is high quality?+
A high-quality report lets you trace the path from input data to every metric and see where observation, calculation, and modeling were used.
Observed facts, calculations, and models →Which limitations should be disclosed?+
At a minimum: input-data quality, panel composition, device-to-person conversion, environmental currency, temporal changes, and modeling assumptions.
All limitation categories →Can OOH be compared with digital?+
Yes, if the definitions of impressions, Reach, Frequency, period, and Universe are aligned. A direct comparison of OTS with served impressions without qualification is incorrect.
Metric context →Sources
- MRC Out-of-Home Measurement Standards, Phase 1 & 2 Combined Final, December 2025.
- WOO Global OOH Audience Measurement Guidelines 2.0, June 2026.
- IAB Digital Out-Of-Home Measurement Guide, July 2025.
- METER Data Specification and materials on OTS, LTS, Reach, and the panel.
- METER materials on advertising assets, visibility zones, and route reconstruction.
This material explains the methodology and does not constitute a claim of MRC accreditation, WOO certification, or an independent audit of METER. The specific set of data and rules is defined for each market, period, and study.