DOOH Audience Measurement

Detailed FAQ

Every question a client, new team member, or curious colleague might ask about how we measure OOH & DOOH audiences — answered in plain language, with no data background assumed.

Internal Reference Document | Operations Team | Companion to the User Guide

Q1. What exactly is an "impression" in OOH advertising?

A: An impression is our estimate of one person having the chance to see an ad at a specific location — it's not a guarantee they looked, just that they were physically in a position to. It's the foundation everything else (Reach, Frequency, GRP) is built on.

Q2. Why can't we just count impressions the way websites count clicks?

A: A click is a direct, recorded action — someone tapped something. A billboard has no such action to record. Instead, we estimate audience exposure indirectly, using traffic counts, anonymised mobile device movement, camera counts, telco data, and manual observation, then combine these into a single estimate.

Q3. What is OTS (Opportunity To See), and why is it the default metric?

A: OTS counts everyone who had a chance to see the ad, regardless of whether they actually looked. It's the broadest, most inclusive metric, and it's used as the default because it gives the most consistent, comparable baseline across every kind of venue — roadside, indoor, transit.

Q4. What is OTC (Opportunity To Contact) and when would we use it instead of OTS?

A: OTC is a narrower cut of OTS — it only counts people who not only could see the ad, but were also positioned to meaningfully engage with it (for example, close enough and at the right angle). Use OTC when a client wants a more conservative, "quality over quantity" read on their audience, rather than the broadest possible count.

Q5. What is VAC (Visually Adjusted Contacts) and how is it different from OTS and OTC?

A: VAC is the most refined of the three — it attempts to estimate how many people actually looked at the ad, whether on purpose or not. Think of it as a funnel: OTS = walked past, OTC = walked past and could clearly see it, VAC = actually glanced at it. Each step down the funnel is a smaller, more conservative number.

Q6. What is "Reach"?

A: Reach is the number of unique people exposed to an ad at least once during a campaign — every person is only counted once, no matter how many times they actually saw it. It answers "how many different people did we reach?" rather than "how many total views did we get?"

Q7. What is "Frequency" and how is it different from Reach?

A: Frequency is how many times, on average, the same person is exposed to the ad during the campaign. Reach tells you the size of the audience; Frequency tells you how repeatedly that audience saw the message. A campaign can have high Reach but low Frequency (lots of different people, each seeing it once) or lower Reach but high Frequency (fewer people, each seeing it many times).

Q8. What is GRP (Gross Rating Point) and why does it matter?

A: GRP combines Reach and Frequency into a single number representing the overall "weight" of a campaign (GRP = Reach % × Frequency). It's a standard media-planning metric used to compare the overall impact of different campaigns or media plans, even across different formats.

Q9. Is a higher Frequency always better?

A: Not necessarily — it depends on the campaign objective. High frequency is good for brand recall and awareness campaigns, since repeated exposure reinforces the message. But beyond a certain point, additional frequency has diminishing returns, and a media planner may prefer to trade some frequency for broader Reach instead.

Q10. What is "Dwell Time" and why does it matter for measurement?

A: Dwell Time is how long, on average, a person stays within view of the billboard or screen. Someone waiting at a bus stop has a much longer dwell time than someone driving past at speed. Dwell Time directly affects both the Impression Multiplier (Section D) and Spot Frequency (Section E) calculations, because more time in view generally means more (or more repeated) exposure.

Q11. What is "Loop Length" and why does it only apply to digital screens?

A: Loop Length is how long it takes for a full rotation of ads on a digital screen to repeat (e.g. six 60-second ads = a 360-second loop). Static billboards show one ad continuously, so there's no loop to account for — this concept only matters for shared digital inventory.

Q12. Does every billboard use the same data sources?

A: No — the mix of data sources depends on the venue type. A roadside billboard leans heavily on traffic and vehicle data; an indoor mall billboard leans on footfall and IoT/beacon data; a bus or train screen leans on mobile SDK and transit-specific data. Section 4 of the User Guide has the full breakdown by venue type.

Q13. What inputs do we need for an outdoor roadside/highway billboard?

A: Two key inputs: the billboard's own Lat-Long, and the Lat-Long of a clear 200m stretch of road from which the billboard is visible (only traffic heading toward the billboard counts). Data then comes from cameras, manual counts, traffic APIs, mobile SDK, published traffic reports, and vehicle-mix data (cars/buses/two-wheelers/others).

Q14. What's different about measuring an indoor billboard (e.g. inside a mall)?

A: Indoor measurement relies more on foot-traffic flow than road traffic. Inputs are the billboard's Lat-Long and the audience traffic flow around it, drawn from cameras, manual counts, IoT/beacon devices, and mobile SDK data.

Q15. How do we measure a screen on a bus?

A: We need the type and placement of the screen or wrap, the bus stops along the route, and the bus's capacity. Data comes from mobile SDK, traffic data APIs, camera counts, and other supporting datasets — this combination lets us estimate both onboard passengers and passersby who see the bus from outside.

Q16. How is a taxi ad measured differently from a bus ad?

A: Taxi measurement is lighter-weight — mainly the type and placement of the screen or wrap, measured via mobile SDK, traffic data API, and IoT data. Taxis don't have fixed stops or routes like buses, so the audience estimate leans more on general traffic exposure along wherever the taxi travels.

Q17. What about train advertising — what makes that unique?

A: Trains need the billboard's Lat-Long, the audience traffic flow, and specific visibility conditions inside or around the train (e.g. how visible the ad is inside a carriage or on a platform). Mobile SDK and beacon devices are the primary data sources here.

Q18. Why do some venues rely on cameras and others don't?

A: Cameras are used wherever direct visual counting adds value and is practical — indoor spaces, transit hubs, and roadside locations. Some venues (like a taxi in transit) aren't practical to fit with a dedicated camera, so those lean more on mobile SDK and IoT data instead.

Q19. Which data sources are strongest for understanding WHO the audience is (demographics, interests) rather than just HOW MANY people?

A: Mobile SDK and Telco data are the strongest sources for Reach, Frequency, Demographics, Audience Segment, and Attribution. Vehicle traffic and IoT data are mainly useful for raw "Potential Views" — i.e. volume, not identity or interest data.

Q20. How are roads classified for traffic-based measurement?

A: Roads fall into six categories: national highways/interstates/toll roads with 5+ lanes, arterial roads (3–5 lanes), other roads (2–3 lanes), roundabouts, crossroads, and pedestrian-only areas. Each category is treated slightly differently because traffic behaviour (speed, volume, dwell) varies a lot between them.

Q21. What other factors get considered in traffic-based audience estimation, besides lane count?

A: Day-type (weekday, weekend, public holiday, or school holiday), direction of traffic, orientation (which side of the road vehicles drive on), and special features such as dedicated bus lanes are all factored into the estimate.

Q22. What happens at a roundabout or crossroad — is that measured differently from a straight road?

A: Yes. Because a billboard at a roundabout or crossroad can be seen from several different approaching roads, multiple road stretches are measured and combined into one overall audience estimate for that location, rather than relying on a single road's traffic count.

Q23. What is Google Distance Matrix (GDM), in one sentence?

A: It's a Google Maps API service that calculates travel distances, times, and real-time traffic conditions between points on a map — we use it as one of our traffic data sources for outdoor billboards.

Q24. How does GDM data become an audience number?

A: Our algorithm converts GDM's raw traffic data into audience numbers by combining it with the vehicle occupancy rate (average people per vehicle) and country/city-specific factors, giving an estimate of passersby density throughout the day.

Q25. Why is 200 metres used as the measurement distance?

A: 200 metres is treated as the ideal distance from which a billboard is realistically visible to approaching traffic. It's measured only in the direction facing the billboard, so we don't count vehicles moving away from it, which wouldn't have a real opportunity to see the ad.

Q26. Walk me through how a street billboard is actually measured, step by step.

A: Four steps: (1) the billboard's coordinates are plotted in Google Maps, (2) Street View is used to check lane count, traffic direction, road type, and location type, (3) the 200m visibility distance is measured and used as the start point for GDM traffic tracking, (4) the detected traffic is combined with vehicle occupancy rate and local factors to produce the potential-views estimate.

Q27. How is a highway billboard measured if it's visible from more than one point (e.g. below and on a bridge)?

A: Each vantage point — every lane or stretch of road from which the billboard is visible — is measured and analysed separately with its own 200m zone. The resulting impressions from each road are then added together to get the billboard's total impressions.

Q28. Is a roundabout billboard measured once, or once per road feeding into the roundabout?

A: Once per road. Every road bringing traffic toward the billboard — including lanes that encircle the roundabout — gets its own measurement and traffic analysis, and all the individual totals are summed for the final number.

Q29. Same question for crossroad billboards — how does that work?

A: The same logic applies: every road from which the billboard is visible is measured and analysed individually, then the impressions from each road are summed into one overall total for that billboard.

Q30. Is GDM the only traffic data source we use, or one of several?

A: One of several. It sits alongside camera counts, manual counts, mobile SDK data, and published traffic reports as inputs into the outdoor billboard measurement process (see Section B).

Q31. Does GDM account for how many people are actually inside each vehicle, not just the number of vehicles?

A: Yes — vehicle occupancy rate is applied on top of the raw traffic count, along with country/city-specific factors, to convert "number of vehicles" into a more realistic "number of people" estimate.

Q32. Why doesn't 1,000 ad plays equal 1,000 impressions?

A: Because a digital screen is shared — a single play can be seen by anywhere from a couple of people to hundreds, depending on the location, time of day, and how many people are typically in view. Unlike a web or app impression (one view = one impression), a DOOH "play" needs to be converted into a realistic viewer estimate.

Q33. What exactly is the Impression Multiplier?

A: It's a screen-specific factor, calculated individually for every screen, that converts a single ad play into the number of impressions it likely generated: Impressions = Ad Play × Impression Multiplier.

Q34. Why can't the multiplier be the same across every publisher or screen?

A: Because audience size varies hugely by location, format, time of day, and even season. A multiplier that worked for a busy mall entrance would badly overstate a quiet suburban indoor screen. That's why it's implemented on a screen-by-screen basis rather than as one blanket formula.

Q35. How does the Impression Multiplier connect to CPM (Cost Per Mille)?

A: CPM pricing charges advertisers per 1,000 impressions delivered. Without the multiplier, media owners would need to constantly change their CPM throughout the day to reflect changing audience sizes (e.g. rush hour vs. midnight), which would make pricing unstable — either overvaluing or undervaluing the same inventory at different times. The multiplier absorbs those swings so CPM pricing can stay consistent and fair.

Q36. What are the actual steps used to calculate the Impression Multiplier?

A: Four steps: (1) calculate a Dwell Time Discount factor based on loop length and dwell time — longer dwell time allows multiple exposures, shorter dwell time is floored at a minimum; (2) calculate a Share of Time factor based on the number of spots and spot duration within the hour; (3) multiply the average-minute impressions by both factors to estimate the hourly Spot Potential Views (Spot PV); (4) divide the Hourly Spot PV by the number of spots in that hour to get the per-spot impression multiplier.

Q37. What is the "Dwell Time Discount factor" actually discounting?

A: It adjusts for how much genuine exposure time a person has. A longer dwell time (e.g. someone seated at a café facing the screen) allows for multiple exposures within one visit, so it isn't penalised. A very short dwell time (e.g. someone walking briskly past) has a floor applied so it isn't undercounted to an unrealistic degree.

Q38. What does the "Share of Time" factor account for?

A: It reflects that a single advertiser's spot only occupies part of the loop — if there are multiple spots and advertisers sharing the same hour, the Share of Time factor apportions the available audience fairly across the number of spots and their individual durations.

Q39. What if a media owner already has their own footfall or impression data — do we override it?

A: No — we use it as a benchmark. For example, if a media owner reports 10,000 monthly footfall, that figure becomes the "truth" benchmark used to calibrate our hourly impression multipliers. Our final platform output will be close to, but not always identical to, their number — typically within a 5–10% difference.

Q40. Can a media owner connect a third-party measurement tool directly instead of giving us a manual number?

A: Yes — if they use a third-party measurement provider with an API (e.g. camera analytics or an industry data platform), our SSP platform can connect to that source directly, for both pre-campaign impression estimates and post-campaign measurement.

Q41. Why is the Impression Multiplier described as important for Programmatic DOOH specifically?

A: Programmatic buying is built around the impression-based purchasing model used in digital and app advertising. The Impression Multiplier lets DOOH — a fundamentally different, shared-screen medium — translate its audience delivery into that same impression-based language, without losing its core strength of reaching large public audiences.

Q42. In simple terms, how do we estimate Frequency for a location?

A: We collect 14 days of anonymised mobile device data for that location, clean it to remove one-off appearances, then fit a straight-line (linear regression) trend to the data. That trend line gives us two numbers — a coefficient and an intercept — which we plug into a simple formula to estimate frequency for any campaign length.

Q43. Why 14 days specifically, and not a week or a month?

A: 14 days is considered the optimal window — long enough to see a reliable pattern, but short enough to minimise the chance that a meaningful number of users' device IDs will have reset or changed in the meantime (which happens periodically on phones and would distort the data).

Q44. Why do we remove Ad IDs (devices) that only show up once in the 14-day window?

A: Frequency is fundamentally about repeat exposure — "how many times did the same person come back?" A device seen only once contributes no information about repetition, so including it would understate or distort the true frequency pattern. Removing single-appearance devices keeps the model focused on genuine repeat visitors.

Q45. What kind of model is used, and what does the formula look like?

A: A linear regression model, in the classic form y = mx + c, translated for our purpose as: Frequency = (Coefficient × No. of Days) + Intercept.

Q46. Can you explain "coefficient" and "intercept" without the maths jargon?

A: Picture a straight trend line drawn through 14 days of frequency data. The coefficient is the steepness of that line — how quickly frequency climbs as more days pass (a bigger coefficient = a steeper climb). The intercept is where that line would sit at day zero — the baseline starting point before any time has passed.

Q47. Can you walk through a real worked example?

A: Using 14 days of data for a location, the model produces a coefficient of 0.3289102049 and an intercept of 2.742958928. To estimate frequency for a 21-day campaign: Frequency = (0.3289102049 × 21) + 2.742958928 = 9.65. So on average, someone in that location's audience would see the ad about 9.65 times over 21 days.

Q48. Why store just a coefficient and intercept instead of the full raw dataset?

Q49. How often is the model refreshed?

A: Every quarter, using the most recent 14 days of available data, so the frequency estimate reflects current audience behaviour rather than becoming outdated.

Q50. What is "Spot Frequency" and why do digital screens need a separate calculation?

A: Digital screens rotate through multiple advertisers' spots in a loop, so a viewer's actual exposure to any ONE spot is only a fraction of their total time in view of the screen. Spot Frequency adjusts the "non-spot" (overall) frequency down to reflect just that one advertiser's share, based on the ratio of Dwell Time to Loop Length.

Q51. What are the exact rules used to calculate Spot Frequency?

Q52. Can you show a worked Spot Frequency example?

A: Non-spot frequency = 9.65, Dwell Time = 45 seconds, Loop Length = 360 seconds. Since 45 seconds is less than 25% of 360 seconds (i.e. less than 90 seconds), the first rule applies: Spot Frequency = ⅓ × 9.65 = 3.22.

Q53. What happens to the Loop Length if more than one spot is booked in the same loop?

A: The loop length used in the calculation is adjusted downward: Loop Length = Default Loop Length ÷ Number of Spots booked. This reflects that each individual spot now represents a smaller share of the full loop.

Q54. Does Frequency Logic use the same data as the Index Score?

A: Yes — both draw on the same underlying anonymised mobile location data supplied by Quadrant (see Section G), though they process it differently: Frequency Logic looks at repeat appearances over time, while Index Score looks at where those devices go (POI matching) to build audience segments.

Q55. What question does the Index Score actually answer?

A: It answers: "How good is this specific location for reaching a specific target audience, compared to the market average?" It's a relative, benchmarked score, not a raw audience count.

Q56. What does a score of exactly 100 mean?

A: 100 represents the market average for that specific audience segment. A location scoring 100 for "Students" attracts students at exactly the rate you'd expect from an average location in that market — no better, no worse.

Q57. What does a score above or below 100 mean in practice?

A: Above 100 means the location over-indexes — it attracts that audience segment more than the market average (e.g. 150 = 50% more than average). Below 100 means it under-indexes — it attracts that segment less than average (e.g. 70 = 30% less than average).

Q58. What three things do we need before we can even start calculating an Index Score?

Q59. What is Geohash, and why is it used instead of raw latitude/longitude?

A: Geohash is a short alphanumeric code representing a geographic area. Comparing two geohash strings to check if a device and a POI are in the same area is much faster and simpler than comparing raw decimal latitude/longitude coordinates directly, especially at scale.

Q60. How does a device get labelled with an audience segment like "Commuters" or "Students"?

A: The device's Geohash is matched against the POI Geohash to determine what category of place it visited (e.g. an educational institution). We count how many times that device appeared at that category of place within the month. If the count meets or exceeds the threshold set in the Segment Rulebook for that category, the device is labelled with the matching segment.

Q61. What is the Segment Rulebook actually deciding?

A: It sets the minimum number of monthly visits to a POI category required before a device qualifies for the associated segment. For example, a category like "Adult Education" might require 10 visits a month to qualify a device as "cat_students", while a category like "Acupuncture" might only require 1 visit to qualify as "cat_healthcare_seekers." Different segments have different natural visit frequencies, so thresholds vary by category.

Q62. What happens in Phase 2 ("segment summary") of the Index Score process?

A: For each billboard location, we calculate a Total Count (all visits) and a Unique Count (distinct devices) per segment. We then calculate the proportion between a segment's raw count and the location's total device count, and apply that proportion to the monthly unique and total counts to get the final "multiplier-applied" figures stored for that segment.

Q63. How is the actual Index Score number calculated in Phase 3?

A: We compare the location's segment share of monthly unique visitors against the market average share for that same segment. The resulting ratio, scaled so 100 = the market average, is the Index Score: Segment Index Score = (Location's segment share of monthly unique visitors) ÷ (Market average share for that segment).

Q64. Can you give a real example of Index Scores for one location?

A: For a sample billboard location: Auto Lovers — 174.6 (well above average), Students — 84.7 (below average), Food Lovers — 89.5 (slightly below average), Gadget Enthusiasts — 380.1 (very strong), Commuters — 97.1 (close to average). A media planner targeting gadget enthusiasts would see this as a strong-performing location for that audience.

Q65. Does the Index Score use personally identifiable information about the people visiting a location?

A: No — it's built entirely from anonymised device IDs and location pings, matched against categories of places (not personal identities). No names, phone numbers, or personal accounts are used.

Q66. Can an Index Score change over time?

A: Yes — because it's recalculated from ongoing movement and POI data, an Index Score reflects a rolling, current view of who's actually visiting an area. If a neighbourhood's businesses or foot traffic patterns shift, the score shifts with it.

Q67. How would I explain Index Score to a client who has never heard the term?

A: A simple framing: "Think of 100 as par for the course. If a location scores 150 for your target audience, it's over-delivering on exactly the people you want to reach — 50% more than the average site in this market. If it scores 70, it's under-delivering for that specific audience, even if the raw traffic numbers look fine."

Q68. Who is Quadrant?

A: Quadrant is a mobile location and Point-of-Interest (POI) data company, part of Appen, operating globally with a large network of contributors and a field-verification process (called "Geolancer") used to keep their POI data accurate.

Q69. What exactly does Quadrant supply to Moving Walls?

A: Two main things: (1) a daily feed of anonymised mobile device location pings, used in Frequency Logic (Section E) and Index Score (Section F); and (2) verified POI data describing what exists at a given location (restaurants, malls, schools, etc.), used to build audience segments.

Q70. How does the mobile location data actually arrive?

A: Quadrant pushes a file daily to Moving Walls' AWS S3 bucket, in one of three formats: CSV, JSON, or Parquet. Each row in the file is a single anonymised location ping from one mobile device.

Q71. What information is in each row of that data feed?

Q72. Does the data include people who have opted out of ad tracking?

A: No — records will not include users who have limited ad tracking or otherwise opted out. Every record also carries a consent flag confirming the user consented to data collection.

Q73. What is POI data, exactly, and how is it different from the movement data?

A: Movement data tells us WHERE a device has been (coordinates and time). POI data tells us WHAT is at that location — its name, category (e.g. "restaurant", "school"), and other details. We need both together: movement data alone can't tell us anything meaningful about audience interests without knowing what kind of place someone visited.

Q74. How accurate is Quadrant's POI data?

A: Quadrant combines automated collection with human field verification to keep POI data accurate, generally in the high-90s percentile range depending on the market and data category — verified through their Geolancer field-collection process.

Q75. Is Quadrant the only data partner Moving Walls uses?

A: No — Quadrant is one important partner supplying mobile location and POI data, but Moving Walls also draws on telco data, IoT/beacon devices, traffic APIs (including Google Distance Matrix), camera counts, and manual observations, depending on the venue type (see Section B).

Q76. What does "Geolancer" mean?

A: It's Quadrant's name for their field-verification workforce — real people who physically visit and confirm details about points of interest, which helps keep the automatically-collected POI data accurate and up to date.

Q77. How accurate are these numbers, honestly?

A: No single sensor or data source is perfectly accurate on its own — that's exactly why Moving Walls blends multiple sources (traffic data, mobile SDK, telco, cameras, IoT) rather than relying on just one. Wherever a primary data source isn't considered highly reliable, we validate the resulting numbers against manual observation or a stronger data source at least once a quarter.

Q78. How do we validate the numbers if we're not fully confident in a data source?

A: Through manual observation (physically counting or checking a sample of traffic/footfall) or by cross-checking against a more accurate available data source, done at least quarterly for any location where the primary source's reliability is in question.

Q79. What happens if a media owner's own footfall figures don't match our platform numbers?

A: We use their figures as a benchmark to calibrate our hourly impression multipliers (see Section D), so our output tracks closely — typically within a 5–10% difference — rather than ignoring or overriding either number.

Q80. How can a brand-new location with almost no sensors still show audience numbers on day one?

A: Moving Walls' measurement system is trained on data from comparable, well-sensored locations already in our network. That trained model, combined with whatever limited sensor data the new site does have, is used to predict its likely audience — this predictive approach is the mechanism described in Moving Walls' own viewership-measurement patent (US 10,587,922 B2).

Q81. Does measurement accuracy improve over time for a given location?

A: Yes — as more real sensor data accumulates for a location, the model's predictions typically become more precise, since it's iteratively regenerated and refined against actual observed numbers rather than relying purely on comparisons to other sites.

Q82. If two different data sources disagree for the same billboard, which one wins?

A: Rather than picking a single "winner," Moving Walls combines sources using a weighted approach — each source is assigned a weight based on the volume and reliability of the data it provides, and the sources are consolidated together (using techniques like isotonic regression) rather than one source overriding another outright.

Q83. Is any of this tracking individual, identifiable people?

A: No. The mobile data used is built on anonymised advertising identifiers (like IDFA or ADID) — not names, phone numbers, or personal accounts. These IDs are tied to a device, not a verified personal identity.

Q84. What if someone has opted out of ad tracking on their phone?

A: Their device is excluded. Quadrant's data feed explicitly does not include users who have limited ad tracking or otherwise opted out, and every record carries a consent flag confirming the user has consented to data collection.

Q85. Can this data be traced back to a specific named individual?

A: No — the identifiers used (device ID / advertising ID) are anonymised and are not linked to personal identity information like names or contact details within this measurement process.

Q86. Why does the data include a "consent" field?

A: It's built into the feed as a compliance safeguard, confirming that the specific record reflects a user who has consented to having their (anonymised) data collected — this keeps privacy compliance built into the raw data itself, not just applied afterward.

Q87. A client asks: "Why does the price of this screen change throughout the day?" What do I tell them?

A: Explain that pricing is based on CPM (cost per 1,000 impressions), and the Impression Multiplier adjusts for real audience size changes throughout the day — busier periods naturally deliver more impressions per play, and quieter periods deliver fewer, so the price reflects what's actually being delivered rather than staying artificially flat.

Q88. A client asks: "How do I know this location actually reaches MY target audience?" What's the simplest answer?

A: Point them to the Index Score for their specific target segment (e.g. Students, Commuters). Explain that 100 is the market average for that audience, and anything meaningfully above 100 shows the location over-delivers for exactly the audience they care about.

Q89. A client asks: "Will the same person see my ad too many times?" How do I frame Frequency for them?

A: Explain that Frequency estimates how many times, on average, the same person is likely to see the ad over the campaign period — and that we can also break this down to Spot Frequency for digital screens, showing exposure to their specific spot rather than the whole screen's rotation.

Q90. A client compares our impressions to a competitor's much higher number for a similar site. What should I check before responding?

A: Check which metric is being compared (OTS vs. OTC vs. VAC — OTS will always be the largest of the three), whether the competitor is using play-count instead of an impression-multiplier-adjusted figure, and whether validation against manual/observed data has been done recently for that location.

Q91. A client asks whether our measurement is independently verified or "made up."

A: Explain that we combine multiple independent, named data sources (Google Distance Matrix traffic data, Quadrant mobile location data, telco data, camera and manual counts), validate the outputs periodically against manual observation, and that the overall predictive methodology is protected under Moving Walls' own US patent (10,587,922 B2) — this isn't a single unverifiable black-box number.

Q92. Where should I send a client or colleague who wants the full technical detail behind any of these answers?

A: Refer them to the companion User Guide ("DOOH Audience Measurement — User Guide & FAQ"), which covers the full methodology, worked examples, and formulas behind every metric referenced in this FAQ.