Dr Derrick Gray
Chief Research & Operations Officer, Numeris
As an executive leader with almost 25 years of building, guiding, adapting and governing audience information systems, I have had the unique perspective of working on two completely different continents leading the audience information systems for two world class Joint Industry Committees (JICs), Numeris and BARC India. This experience has led me to often step back, recognizing the uniqueness of both the Canadian and Indian video advertising eco-systems, and think about the best practices of each. While the environments, needs, regulatory environments, and actors all differ, it is quite apparent to me, that despite these differences, there are common objectives and goals, and as such, each organization can learn from one another.
While this article focuses on Canada and India, my hope is that this framework can be applied to multiple countries where global JICs can learn from each other. While their systems will remain unique to their countries, sharing of lessons, best practices, and experiences will only strengthen audience measurement globally, and will support an industry that is moving towards a global and local structure.
Within this framework, I present a unique perspective of eight lessons which could be shared between Canada and India.
When it comes to the population, Canada is a diverse market. We have a highly dispersed population, spread across the second largest country by area in the world, a highly multi-cultural society, grown through immigration over many decades, media distributed across many languages, and two distinct societies and media ecosystems (i.e., French Canada, English Canada).
But India takes heterogeneity to another level.
At a national level, two official languages, and 21 additional languages under the Eighth Schedule to the Constitution of India, with each state or Union Territory having their own set of official languages. The 2011 Indian Census identified 121 languages and 270 mother tongues spoken by at least 10,000 people, while nearly 69% of the population lived in rural areas. These are not simply demographic statistics. They represent fundamentally different media environments, content preferences, economic circumstances, and patterns of technology adoption.
Economic diversity in India is illustrated by India’s 2023-24 Household Consumption Expenditure Survey where urban residents spend approximately 70% more per person than rural residents. Within urban India itself, the highest-consuming 5% spend, on average, more than eight times as much as the lowest-consuming 5%.
BARC's panel design explicitly accounts for geography, town class, socioeconomic characteristics and other control variables. Its methodology also oversamples certain areas to improve the usefulness of estimates; for example, its published design shows different sample proportions for megacities, rest of urban India and rural India.
The lesson here is not that Canada needs a bigger panel, but rather, that representation matters more as audiences fragment.
Typically, measurement systems were designed around relatively broad demographics (e.g., Adults 25-54, Adults 18-49) and geographic classifications. However, fragmentation starts to create smaller and smaller audience segments.
The Canadian advertising eco-system increasingly needs to understand not simply Canadians, French Canadians, English Canadians, but rather, Canadians whose media behaviour and consumption patterns vary according to their language(s), geography, age, cultural background, connectivity, platform adoption and device ownership.
Audience fragmentation isn't just fragmentation between platforms. It is fragmentation between people.
India has always had to confront that reality. Canada increasingly does too.
In the early 2010s, Numeris made a deliberative move to stop referring to “Television” measurement, in favour of “Video” measurement. This was an important move as it shaped the future direction that the company would take. In 2025, Numeris launched National VAM putting linear television and streaming consumption into a common dataset and providing duplicated and unduplicated reach across platforms. More than 95% of video platforms in Canada are now covered in National VAM.
This represents a change in philosophy.
Numeris is no longer asking, “who watched television?” and is now increasingly asking “how did people consume video?”. These are very different questions.
BARC India has deployed an enormous and highly sophisticated measurement system in India. However, this system is fundamentally anchored on television viewing using watermarks and BARC’s proprietary people meters to capture viewing. But India is not a TV only ecosystem. In fact, according to BARC India, only 210 million household in India have a television. On the other hand, 86.3% of Indian households have internet access at home, which includes mobile connectivity.
This means that streaming, connected television, mobile video, YouTube and other digital platforms are increasingly competing for the same consumer time and advertiser budgets.
Canada can therefore offer an interesting lesson: Don’t allow the medium that created the currency to define the future boundaries of the currency.
BARC India measures television at a scale not seen in any other country.
As of late September 2026, the BARC panel consists of 77,000 metered households with plans to reach 80,000 homes by November. India’s Television Ratings Policy 2026 issued by the Ministry of Information & Broadcasting stipulates an eventual target of 120,000 homes. With that sort of scale, research operations now become industrial operations.
Recruitment, installation, panel maintenance, compliance, servicing, logistics, quality control, and data processing all must operate systematically at scale. It moves to a model where the methodology only works if the operations work.
BARC India recognized this early on and quickly moved to a complex panel and operations management system leveraging best in class technology and software. Because the best-in-class sample design is meaningless if the meters don’t function, weighting algorithms can’t compensate for issues in maintaining households, and the best data processing systems cannot help when the data pipeline itself is unreliable.
This is an excellent illustration that measurement science and operational excellence are interlinked and cannot be separated. A best-in-class measurement organization must not only excel in measurement but most importantly in operations. The role of the COO is just as important, if not more, than the role of the CRO.
"The critical lesson is that measurement at scale is not only about understanding audience behaviour; it is equally about managing a diverse and dynamic panel. In a market as large and varied as India, representative measurement depends on continuously monitoring engagement, compliance and panel health to maintain data quality over time.
As panel sizes grow, success increasingly depends on operational excellence. Strong governance, rigorous quality controls, and advanced analytics help identify issues early, support efficient interventions, and ensure the long-term integrity of the measurement system."
Sumit Singh
CEO, Meteorology Data Pvt. Ltd.
Panels used to be the core of measurement by directly providing signals which could then be extrapolated to the population to produce audience estimates. While panels are still very important to measurement, their role is changing.
Rather than simply expanding panels to address fragmentation, Numeris has been moving towards hybrid measurement for the media they measure. This is demonstrated through Numeris’ Enhanced TAM product, Numeris’ National VAM product, and more recently Numeris’ move from Online Radio Diary to Radio Local+ to measure local radio.
In both Enhanced TAM and National VAM, the panel information is coupled with large-scale data such as return-path data from set-top boxes, tag data, or first party data integrations, where the panel is used to calibrate, attribute and interpret the larger dataset.
This is an important conceptual evolution.
Traditionally audience measurement was:
Panel Audience Estimate
Under this new model it becomes:
Panel + Big Data + Identity + Modelling Audience Estimate
This is an entirely new role for the panel.
Instead of asking how large the panel needs to be to address fragmentation, the better question becomes “What information must come from people, and what information can come from machines?”.
While the massive panel deployed by BARC India provides incredible visibility into the viewing habits of the market, Canada provides an example of a different route. Panels provide demographic, behaviour, and calibration and Big Data provides scale and stability.
Due to the high economic stakes, it is natural that certain actors may be incentivized to find ways to manipulate ratings systems. Manipulation can have dire consequences on the reputation of not only the measurement company, but also the accuracy and validity of the audience estimates which form the currency.
BARC India recognized this risk early on and established many protocols to thwart any malicious attacks. In the April 24, 2024, version of BARC India’s Code of Conduct for Redressing Viewership Malpractices, BARC India defines its Vigilance Team as being responsible for investigating incidents involving viewership malpractices. It also defines the Measurement Science and Market & Business Analytics Teams as being responsible for statistical analysis, data validation, and methodology oversight.
There is a lesson to learn here that measurement integrity requires more than good sample design and sophisticated statistical methodology.
India’s experience also makes clear that methodological rigour alone is not enough to protect an audience measurement currency. BARC India publicly documents dedicated vigilance operations, statistical monitoring, field investigations and formal disciplinary procedures to detect and address attempted panel infiltration and ratings manipulation. These safeguards reflect a fundamental reality of currency measurement: when audience estimates carry significant commercial value, the measurement system itself becomes a potential target.
In the digital era, this has become even more important. The identity of panel households being revealed is no longer the only risk, but now, digital systems can be attacked. This can include bots, invalid traffic, identity manipulation, device farms, data contamination, and potentially even synthetic behaviour.
Given this, measurement integrity needs to be designed into the audience measurement system from the very beginning. It cannot be treated simply as an audit function after the fact. Systems require sophisticated safeguards, quality controls, and monitoring to ensure that the published currency data remains clean, and a voice of truth.
A campaign generated 3 million television viewers, 2 million streaming viewers and 1 million digital viewers. How many Canadians did it reach? Six million? Maybe. But probably not. This question illustrates the importance of something which seems simple, but is very complex, de-duplication.
Numeris’ Single Source Panel, in which in addition to household members carrying a portable meter, the households have a router meter installed, capturing television, radio, and online-video consumption from the same participating individuals and households. This provides something extraordinarily valuable, observed deduplication in a deterministic way.
Numeris can see that Person A watched linear television and Prime Video and Crave rather than trying to calculate probabilities to assess if three devices or datasets represent the same person.
For India, where digital consumption is massive and heavily consumed through mobile devices, cross-platform identity is particularly challenging and important to understand.
This doesn’t suggest that India should copy Canada’s technology, but rather, the underlying principle: The closer measurement gets to observing people across media rather than independently measuring devices, the more valuable it becomes.
I-JIC, the International Association of Joint Industry Committees for Media Research, defines a JIC as a nonprofit-oriented, neutral organization in which all key players of an advertising market (i.e., normally advertisers, agencies and media owners) are represented.
Numeris is a Joint Industry Committee (JIC) which was formed on May 11, 1944, by the Canadian Association of Broadcasters. It employs a governance structure that has representation from Advertisers, Agencies and Media Owners on its Research Advisory Committee (RAC), Audio and Video Executive Committees (AEC, VEC), as well as its Board of Directors. Similarly, BARC India, which rolled out its measurement system in April 2015, is a joint-industry body owned by the Indian Broadcasting and Digital Foundation (IBDF), Indian Society of Advertisers (ISA) and the Advertising Agency Association of India (AAAI). It has a Technical Committee and a Board of Directors with representation from Advertisers, Agencies and Media Owners.
These structures have become even more important in the digital era with many competing versions of reality: broadcaster data, streaming-platform data, ad-server data, agency data, advertiser data, connected TV data, set-top box data, and measurement-company data.
All these various forms of data may be legitimate. But what is important is that these data may not all agree. Therefore, the purpose of a JIC is no longer simply to produce ratings. It is to establish the rules by which competing datasets become trusted measurement. The role of governance provided by a JIC becomes more important, cementing the need and value that a JIC can bring to the market.
India should not seek to build a “Canadian” measurement system.
Nor should Canada seek to build an “Indian” one.
The economics, populations, media structures, technology environments, and industry requirements of the two markets are fundamentally different. The opportunity, therefore, is not to replicate systems, but to borrow the principles that make them effective.
Design for diversity.
National averages should not be allowed to conceal increasingly heterogeneous patterns of media behaviour.
Think operationally at scale.
Measurement science and measurement operations are inseparable; the methodology only succeeds when the operations do.
Protect the currency.
Integrity and resilience must be designed into the measurement system, not added after the fact.
Remember local media.
Fragmentation is not only about national audiences; regional, linguistic, and local media remain essential parts of the ecosystem.
Move beyond television measurement towards audience measurement.
The currency must evolve to reflect how people consume video across linear television, streaming, mobile, and other digital platforms.
Let big data provide scale and panels provide the truth needed to interpret it.
Large datasets can deliver breadth and stability, but panels remain essential for calibration, attribution and understanding who is behind the behaviour.
Build cross-platform measurement around people, rather than devices.
Devices should generate signals, but the market ultimately needs to understand the individuals and audiences moving between them.
Make duplication foundational, rather than adding it after separate media currencies have been built.
A common approach to identify overlap must be established at the outset if the system is to produce credible unduplicated reach across platforms.
Allow the JIC to evolve from simply producing ratings to serving as the steward of standards, data interoperability, and industry trust.
As measurement becomes more fragmented and data sources multiply, the JIC must provide the neutral governance needed to align standards, reconcile competing datasets, and maintain confidence in the currency.
Canada and India sit almost at opposite ends of the audience-measurement spectrum. One is a relatively small, highly connected market confronting rapid media fragmentation. The other is one of the world's largest and most diverse media markets, where measurement must operate on an extraordinary scale.
Yet the direction of travel is remarkably similar.
Both markets must reconcile panels with big data. Both must understand audiences across platforms. Both must distinguish devices from people. Both must maintain industry trust while methodologies become increasingly complex.
And both ultimately face the same question: How do we build a measurement system that reflects not how media used to be consumed, but how people actually consume it today?
Perhaps the most important lesson Canada and India can teach each other is that the future of audience measurement won't be built by choosing between panels and big data, television and digital, or established currencies and new technologies.
It will be built by connecting them.
For more on why people remain at the heart of trusted audience measurement, read these perspectives from Alicia Olson-Keating, President & CEO, Numeris:
People always matter – Why independent people panels are essential to calibrate, validate and keep big data honest.
From Recall to Reality – Why audience measurement must be grounded in observed behaviour, not just what people remember.