TRP scam: BARC's TV audience measurement system is rotten

On October 15, the information and broadcasting ministry was summoned before a parliamentary standing committee where one topic of discussion was television audience measurement in India, the system that measures “what India watches”. That is, incidentally, the tagline of the Broadcast Audience Research Council, or BARC, an industry body owned by the country’s major broadcasters.

Six years ago, BARC was accredited by the Indian government to measure TV audiences. Its competitor, TAM Media Research, had faced flak for being consistently inaccurate and was duly outmoded. In an industry where Rs 32,000 crore in advertising revenue is contingent upon TV ratings, BARC was welcomed with much anticipation.

But its star wasn’t shining so brightly this October.

A day after the parliamentary committee convened, Hindustan Times reported that its members had described the BARC system as “easily manipulated”, “not very scientific and accurate”, and “not in tune with current technologies”. A News18 report said all sides had agreed that "TRP numbers are not a true reflection of Indian viewership" — undercutting the very purpose of BARC.

After the latest “TRP scam”, BARC announced a three-month suspension of ratings of news channels to “review its already stringent protocols and further augment them”. On November 4, the information and broadcasting ministry constituted a committee to review guidelines on television rating agencies in India.

A presentation made by the I&B ministry before the standing committee said BARC had filed 11 FIRs through its vendors into cases of ratings tampering across Andhra Pradesh, Karnataka, Kerala, Madhya Pradesh, Maharashtra and Assam. These claims were included in slides titled “Submissions by BARC” in the ministry’s presentation. The slides pointed out that the body had an “automated system” to flag abnormal behavior in ratings, a code of conduct to redress grievances by subscribers, and a “graded escalation/penalty matrix”. BARC’s disciplinary committee, it added, had heard 31 cases of tampering and penalised 19 broadcasters.

BARC’s claims were angled to assert that it has a robust mechanism to detect tampering and serve swift justice. However, it betrayed a fact that the agency might be reluctant to admit: the BARC system allows for easy manipulation of ratings. It is gameable by design.

The English news industry in India understands this well. The genre occupies a tiny and extremely fragmented space in the TV landscape, a dozen channels competing for one percent of the national viewership. That makes this segment particularly prone to manipulation.

For an agency that influences where thousands of crores of advertising money is spent on TV, a three-month pause on news ratings is far from a fix.

Some numbers and nuance

BARC was founded in 2010 by the Indian Broadcasting Foundation, the Indian Society of Advertisers, and the Advertising Agencies Association of India. While the ISA and the AAAI have 20 percent representation on BARC’s board, the IBF has 60 percent representation and owns 15,000 of its 15,001 shares. Uday Shankar, the chairman of Star India, owns one share.

The IBF claims to manage “350 plus channels and about 90 percent of television viewership” across India. It counts some of the biggest Indian and international broadcasters as members – Disney, BBC, Discovery, Star, Sony. TV networks like Times Television, TV Today, Network18, ABP, and Asianet are also part of it.

So, BARC essentially measures ratings for channels which constitute the body that owns and operates it.

BARC gathers data from 44,000 households across India – which is known as the “sample size” – to rate the viewership of over 600 channels. The households are segmented by class, language and geography. India has 862 million TV viewers.

According to the Telecom Regulatory Authority of India, China, with a TV audience of 1.2 billion viewers, has a sample size of just over 10,000 households. The US has a sample size of 43,000 for 300 million viewers. Hong Kong and Chile, each with an audience of nearly seven million, have sample sizes of 645 and 790 households, respectively.

The unit of these ratings are called “impressions” and they are calculated using bar-o-meters — little machines whose location is purportedly a closely guarded secret and which can drastically alter ratings if “fixed” smartly enough.

Sixty two percent of bar-o-meters in India are installed in urban households, and 38 percent in rural ones. In the urban segment, 27 percent of the meters are deployed in megacities. Mumbai alone has about 2,000 of them.

SY Qureshi, the former director general of Doordarshan, said until about a decade ago these meters — then called people meters — were connected to TV sets. The data was stored in a cartridge, which the operator would manually collect twice a week. Contemporary meters have done away with much of this hassle. They are wired directly to each household’s set top box and include GSM sims. Sometime between 2 am and 4 am every day, the sims transmit the viewership data to the relevant agency.

According to BARC, its ratings not only reflect the number of people watching a certain channel but also the time they spend on it. So, by this metric, Channel X, which is watched for 10 minutes each by four people, may rank below Channel Y, which is only watched by two people but for half an hour each.

“For any channel, the objective is to lengthen the time a viewer is spending on that channel, the stickiness factor," explained Aunindyo Chakravarty, former managing editor of NDTV India and NDTV Profit. "Getting viewers is not the most important concern for a channel. It's about how long you can keep them watching."

The weekly impressions released by BARC are extrapolated from “clock minutes” measured by the baro-o-meter in 30-minute cycles. These minutes are registered using the audio watermarking technology, which enables the bar-o-meter to detect which channel is running from its distinct audio component.

If a channel plays for more than 30 seconds, consecutively or not, it registers a clock minute. If, say, Channel X and Channel Y play for 25 seconds each, and Y spills over to the next minute, it will be awarded the clock minute. If neither spill over, an algorithm randomly allots the minute to one of the two channels.

This can become cumbersome in a metered household. The bar-o-meter comes with a remote. If you are over two years old, you have a button assigned on it. A family member has to press this button once they sit down to watch TV and again once they stop. This way, BARC claims to identify the age group and gender of the viewer.

A poor record

Since the “TRP scam” hit the headlines in early October, implicating Republic TV and two Marathi news channels, bar-o-meters have come under intense scrutiny.

The Times of India pointed out that Mumbai, with 2,000 meters, recorded a mere 0.4 percent reach for English TV news channels in week 39 this year. The weekly average of the time spent watching this genre was 15 minutes. This means only eight metered households in the entire city account for the entire English news viewership.

So, an English news channel wanting to rig these ratings has to influence only a couple of households to watch their channels for, say, an hour each. This would shoot up the genre’s average watching time in Mumbai by 73 percent and bring it to 26 minutes. If this is the impact on the English TV news genre as a whole, the effect on the channel’s ratings will be astronomical.

Such manipulation is harder to do in bigger genres with higher reach, like Hindi news or general entertainment channels, known as GEC.

Not that the record of Hindi news channels has been spotless. In July this year, the News Broadcasters Association, a private body that represents TV news channels, approached BARC with a complaint accusing TV9 Bharatvarsh, a Hindi channel that figured regularly in BARC’s top five ratings chart, of TRP manipulation.

Rajat Sharma, the association’s president and chairman of Hindi news channel India TV, alleged that the TV9 group had a history of “manipulating” viewership data. He did not spare BARC either. “These are corrupt practices, which are being done with complete connivance with BARC and the broadcaster,” he complained, referring to TV9 Bharatvarsh.

The NBA’s rival, the News Broadcasters Federation, also drafted a comprehensive admonishment. It said that it had “scathing evidence to indicate manipulation and tampering of BARC meters by the TV9 group”. The federation questioned BARC’s integrity and claimed the rating agency’s closely guarded information on metered homes was “completely compromised”. It demanded an investigation into “complicity at a higher managerial level” at the TV9 Network.

A source in the NBF told Newslaundry that TV9 Bharatvarsh had jacked up its TRPs by bribing people in metered homes, one of them in Chandigarh. An individual was purportedly stung in the act and the video was sent to BARC. The ratings agency told BestMediaInfo that the alleged sting was “fake”, and those who figured in it were “paid to say what they did”.

The TV9 Network rubbished the allegations. It suggested that the NBA’s main grouse was the network’s decision to not renew its membership with the body, which had expired in March this year. Barun Das, the CEO of TV9 Network, added: “The fundamental reason behind our success is a concerted strategy of content, distribution and promotion aided by astronomical rise in viewership during the early weeks of the lockdown. This has been discussed in relevant media forums openly, repeatedly and live.”

In a fragmented market competing for a percent of the viewership pie, television channels strike deals with cable operators for one-upmanship. A seasoned media professional said over the years a small section of the ratings industry had identified households with bar-o-meters. They could be local operators, or ex-employees of a BARC vendor concerned with meter installation. Usually called “consultants”, they approach TV channels with a promise of better ratings. Once recruited, they induce metered households to let certain channels run for hours.

Another media professional who has observed the workings of BARC as well as its predecessor, TAM, from close quarters explained how events like floods, riots or power outages give away areas with high meter density. “You can’t have a flood across all of India, or even a power outage. Such events are geographically distributed. So if a few bar-o-meters stopped working in one part of the country because of a flood or a riot, a news channel’s ratings would be seriously hit. And when the channel studied the data, the sample would reveal itself.”

Knowledge of areas with high meter density, he added, also helps TV channels and their cable operator associates to ensure rival channels have weaker connectivity in those areas.

The TV9 matter hasn’t been put to rest. An executive in the news industry claimed that BARC has not taken any discernible action against the network. The source at NBF alleged that two members of the federation who are part of BARC’s technical committee were not invited when the committee convened to discuss the manipulation of TV9 Bharatvarsh’s impressions. “BARC’s response to the allegations was simply an obfuscation,” the source said. “The reality is that tampering is happening.”

But BARC has acknowledged meter manipulation in the past. In a statement in September 2017, the body admitted there were “unscrupulous elements” who wanted to “infiltrate security of BARC India’s sample (panel homes), and unfairly influence their viewership habits.”

The statement added: “Their goal (and business) is to skew final viewership data in favour of some channels, using unfair means that BARC India defines as ‘viewership malpractice’.”

Other ways to rig ratings

Tampering bar-o-meters is one of several ways of rigging ratings. Two other methods that channels have at their disposal is the landing page feature and acquiring multiple logical channel numbers, or LCNs.

The landing page feature allows a channel to be the first channel to appear on screen when a television set is turned on. It enhances both reach and average time spent of a channel, more so if it is a niche genre.

Acquiring multiple LCNs has a similar effect. It allows a channel to appear more than once while the viewer is flipping through channels. If Channel X appears on channel numbers 203, 303 and 403, for instance, and a surfing viewer spends 12 seconds on each, Channel X will clock 36 seconds and will be awarded the minute.

In December 2018, TRAI directed distributors and broadcasters “to restrain from placing any registered satellite television channel, whose TV rating is released by TV rating agency, on the landing LCN or landing channel or the boot-up screen with immediate effect”. In June 2019, the Telecom Disputes Settlement and Appellate Tribunal set aside this order, stating that TRAI didn’t have the powers to issue such a direction.

BARC and TRAI have a thorny history. One of the most critical assessments of the BARC system came from an 80-page TRAI consultation paper put out in December 2018, titled “Review of Television Audience Measurement and Rating System in India”. The regulator pulled up the ratings agency over its operations, structure, transparency, and accountability.

TRAI set two months to receive comments on the review. News broadcasters, entertainment networks, data companies, and BARC itself replied to pointed questions raised by the regulator. In its response, BARC admitted that empirically “it is observed that placing channels on the landing page can influence the ratings specially for niche and smaller viewership channels”.

TRAI’s detailed recommendations were published in April this year.

In September, BARC announced it would introduce algorithms to mitigate the effect of the landing page feature on viewership data across all genres. A few weeks later, Times Now chief executive officer MK Anand slammed the decision, arguing that “the reach that’s coming out of landing pages is not spurious reach”. BARC and Times Now are now fighting a legal battle over the algorithms in the Delhi High Court.

BARC’s position on multiple LCNs is rather convenient. In 2017, it said acquiring more than one LCN was a common practice for channels, especially in the news genre. “BARC India neither monitors channel placements across the various DTH platforms/cable head-ends in the country, nor does it have the mandate to do so...BARC India is not the regulatory body for resolving issues concerning multiplicity of LCNs for a channel,” it said. “Ideally these issues should be sorted among broadcasters themselves rather than dragging BARC India into these.”

Archaic and flawed?

Pankaj Krishna, cofounder and CEO of Chrome Data Analytics and Media, believes that the technology used by his firm is years ahead of BARC’s. Audio fingerprinting, for example, is like Shazam for TV channels: the Chrome app on Krishna’s phone can detect and identify channels playing on TV a few yards away within seconds.

In his eighth floor office in Noida, Krishnna, who counts several TV channels as clients, said it will take a leap of faith to believe that BARC’s 44,000 panels work flawlessly. His company’s comments on the TRAI paper pointed out that at any given time “there are over 30 channels which are actively involved in tampering the people metered homes”.

It added that the addition of BARC remotes to bar-o-meters was a drawback as well. “Current demographic profiling is incorrect as it is an unrealistic expectation to get sample to manually select the appropriate button on the remote to identify themselves and that too each time they watch TV – which in turn leads to inaccurate information. Respondents would conveniently press the all button or would not press any button at all, leading to wrong demographic extrapolation.”

This is a common concern in the industry. The viewership lifestyle in metered households – with technology tracking every second of viewership and family members registering on the remote whenever they start watching a channel – leaves only a certain type of household to volunteer for bar-o-meters.

“The BARC meters end up being installed in areas where people are needy,” the source at NBF told Newslaundry. “The data reflects the choices of these groups.”

Audio fingerprinting technology uses a technique called spectrogram to produce a signature for a piece of audio and then finds an identifier – a song, a tune, an ad, and even a TV channel. Krishna demonstrated that the data on viewership of TV channels using audio fingerprinting is produced in real time. This, he added, makes it quicker than the system used by BARC, which takes around five days to process primary viewership data.

“Everyone is doing it for the heck of it,” said Krishna, about BARC’s monopoly over TV audience measurement in India. “This makes no sense. Assuming the secrecy of 44,000 homes in a country like India is preposterous; releasing data after five days in 2020 seems archaic; and why use meters when you have audio fingerprinting?”

Mahira Kaul, a staffer at the US-based media company Desis.Live, which aggregates content for North American audiences, is “suspicious” of the BARC ratings system. The body’s methodology and ownership, she believes, puts a question mark on the veracity of its data.

Kaul thinks that one of the big drawbacks of the BARC system is that it does not account for viewership on digital platforms. “There was a 55 percent rise in overall streaming earlier this year. Amazon saw a hike of 83 percent and Netflix 202 percent. Much of the audience, especially millennials, consume digital content. But BARC does not provide any numbers on it,” she said.

According to a report by the Internet and Mobile Association of India, the country now has 504 million active internet users. In 2019, ad spend in digital media in the country grew by 26 percent to reach Rs 13,683 crore.

When Kaul wrote to BARC asking them about viewership measurement in the digital realm, the body told her that it was “working with multiple partners to introduce multi-screen measurement”.

In the US, TV audience is measured by Nielsen Media Research, a sister body to Nielsen NetRatings, which measures internet and digital media audiences. Kaul was surprised when she found out that BARC, unlike Neilsen, is owned and run by broadcasters. “Why are stakeholders rating themselves?” she asked. “This isn’t politics. This should be run by a data science company.”

BARC’s predecessor, TAM, was a joint venture between Nielsen and Kantar Media, owned by WPP Plc, a London-based advertising company.

In its comments on TRAI’s paper, BARC pointed out that its ownership pattern reflected standards set in the UK, France, Canada, South Africa, and Australia, where rating agencies are purportedly governed by broadcasters and advertisers. The Canadian body, it added, is “operationally run by an independent executive team”, an element missing at BARC.

Desis.Live, unlike BARC, measures demand for South Asian content on the internet through multiple parameters. While viewership and demand are different metrics, they tend to be correlated. Kaul’s suspicion about tampering at BARC was strengthened when she studied the body’s data on popular Indian shows.

For week 41, for which ratings came out on October 22, Kundali Bhagya on Zee TV ranked at the top of BARC’s list of most-watched shows in India. Saath Nibhana Saathiya on Star Utsav ranked second and Anupama on Star Plus third. But in the demand numbers extracted by Kaul, Kundali Bhagya ranked 16th, Saathiya 13th, and Anupama 424th.

In other words, the rating agency’s data did not reflect dips in demand for particular shows.

“We have real time data to show that a popular show just had a bad episode. The demand falls on search engines and websites, but BARC’s numbers would not budge,” she said. “That is when I began to doubt the veracity of the data BARC collects.”

Coda

A top executive at an ad buying company told Newslaundry that TV news channels have given BARC a bad name. “It is a perfect system to produce viewership data for GEC and sports categories, but not for news. The recent controversies are a result of a battle of egos in English news media, which controls just one percent of the market,” the executive said.

He added that advertisers in India don’t solely rely on BARC to map ratings and determine sellers. “We look at digital activity, follow the demand, and look at a channel’s perception. Republic Bharat has recently overtaken Aaj Tak, as BARC data shows. But that does not mean it now gets more ads than Aaj Tak, which has legacy credentials.”

The manipulation of ratings — tampered meters, landing pages, multiple LCNs, leaked locations of metered homes — casts a shadow on BARC, which was ushered in with much hope in 2015. But now, like TAM a decade ago, BARC is fast becoming unpopular. The fix it needs would have to be much more than halting ratings for a few weeks.

In its 2018-19 annual report, the NBA lamented that it was “unfortunate” that its concerns of “flawed ratings and mechanism” found no takers at BARC. “If there is no resolution on the issues raised by NBA, members would be left with no other option but to move out of BARC, which situation, NBA sincerely hopes would not arise,” it said.

In a private communication after the TV9 fracas, the NBF told BARC that it should not only provide unprocessed data to broadcasters but also share its audit reports. It added that BARC needs to be “self-vigilant”and not shrink away from investigating “its own”.

Since the coronavirus lockdown this year, TV news channels have been anxious about ad monies. Their viewership increased but revenue tanked. The source at NBF told Newslaundry that the dip is as much as 40 percent. He added that since the BARC ratings for news are now paused for the next 12 weeks, advertisers are negotiating for lower rates. All this is happening around Diwali, a season most fertile for earning ad bucks.

“We are helpless,” said the NBF source. “It’s like the broadcasters and BARC are married. If we are unhappy, there is little we can do. BARC is all we have.”

Sunil Lulla, the CEO at BARC India, declined to comment for this story. Newslaundry also sent a set of questions to BARC. The body responded: “Many of your questions have been reported and responded in the past which are available in the public domain. Furthermore, many of your questions are private and confidential between BARC and its subscribers. We won’t be able to comment on any of this.”

***

Independent media has a critical role in sustaining democracy, telling stories that need to be told and asking questions that require answers. Support independent journalists by paying to keep news free. Subscribe to Newslaundry today.