The WhatsApp chats between Republic TV editor Arnab Goswami and Partho Dasgupta, former CEO of Broadcast Audience Research Council, reveal how the television ratings agency was struggling to justify the data that it was putting out and tinkering with it constantly. This was all being done to push the narrative of “All iz Well” and “we know what we are doing”.

This is the story of how a couple of men in positions of power helped set the narrative for what is “news”. A BARC insider we spoke to revealed that Dasgupta and ex-COO Romil Ramgarhia were part of a system where they were manipulating data to decide how channels should be ranked.

“A ‘ratings’ meeting would be held every Wednesday at BARC,” said the insider. “There a team would gather to remove data points which they thought didn’t make sense. Then they would come up with stories on how to justify the outlier data, and measure the rankings.”

Thursday was the day BARC put out these weekly rankings. Since we have the chats, we found multiple instances of this being done. What’s worrying is the fact that all of this seems like everyday business for Partho and Romil. But in reality, they were changing national narratives across news channels and deciding where advertisement money flows.

The word “outlier” is key in this whole enterprise. A conversation from 16 April 2019 between the two gives a hint regarding this. Ramgarhia told Dasgupta, “ About 90 percent of control for Bharat is outliers and 10 percent is landing.”

“Outliers = Tampering,” Dasgupta responded. “And the policy applies.”

Ramgarhia told him that he will provide a “draft” tomorrow on these lines.

In September 2019, a two-member committee of BARC found the outlier policy setup to be “accurate and unbiased”. The committee noted that there was “no manual intervention in the process for treating outliers”.

But just a few months later, in July 2020, a forensic investigation conducted by Acquisory Risk Consulting Pvt Ltd gave an important observation:

“We have seen evidence of multiple re-runs to get the desired outcome. The CAC, outliers were removed and added back to change the ratings.” CAC in the report refers to channel audience control.

To add to the forensic report, the Arnab-Partho chat leaks reveal instances of this as well. That there was a fair amount of manual interference happening to adjust data for and decide the ratings.

But first, we need to understand what these ‘outliers’ are and what ‘controlling’ means.

What are ‘outliers’?

BARC, an industry body owned by the country’s top broadcasters and advertisers, is tasked with measuring ratings of TV channels in each genre. The ratings give advertisers a sense of who is watching which channel, helping them decide where to spend their ad money.

This viewership data is collected from a fairly spread out sample of 44,000 households where BARC has installed “bar-o-meters" on TVs. The data collected from these “bar-o-meters”, as they are known in industry jargon, is used to approximate the viewership numbers for the rest of the population.

The data shows how much time the viewer has spent watching a channel, and ratings are derived accordingly. This is what the rankings look like:

But, here’s the catch. Among this sample size of 44,000-odd households, not everyone watches news. They watch entertainment channels, movies, sports and much more. Even in the news genre, more people watch Hindi news than English. So the actual sample size for the news genre is tiny.

Let’s say you are a person living in Delhi and have one of these meters. One day, you fall asleep while watching TV. Now the data shows that you spent 8 hours watching Republic TV, although I’m surprised you were able to fall asleep watching that channel.

Snark aside, now the data would show a sudden spike in ratings, literally driven by that. Now the people at BARC will look at the average view time spent by other people, which is usually just minutes as opposed to hours, and see your data. That is an “outlier” data point.

Here’s the thing, the sampling can get mighty bonkers at times as evidenced from an email in the chat dump from Vijay Subramanian, the head of analytics at BARC, to Ramgarhia and Dasgupta.

Think about it, only one person was watching CNBC TV18 from Delhi for 20 minutes. They removed the only viewer because they assumed it’s an ‘outlier’. Subramanian further pointed out the fact that if the same viewer comes back later to watch the channel for 3 minutes, it would be difficult to explain the ratings drop.

“We understand the business pressures, we would request you to understand the tightrope walk which is our function with regard to news,” says Subramanian’s email.

If you go back to that “outlier = tampering” screenshot above, you’ll notice that the BARC officials were talking about controlling data to identify these outliers. In the case of TV channels, there is another issue with landing pages. When you turn on your set-top box, the first channel you see becomes your landing page. If a particular channel shows up first, they get viewership for it by default. Instances of that happening were also treated as an “outlier”, sometimes.

Controlling data

The BARC insider told us about how raw data coming in was “controlled” by the news team, at the direction of Dasgupta and Ramgarhia, to decide rankings. What this means is that they would adjust the numbers to bump up certain channels while punishing others. There were some instances we found where the outlier policy was used as an excuse for this.

There are instances of this found in the forensic report. According to that, the ratings of Times Now were being disproportionately “controlled” to favor Republic TV. The report lists out multiple emails allegedly showing drastic reduction in the ratings of Times Now. To give you one instance (there are a lot of them), the emails they shared, looked like this:

Here, the email said, “Please find numbers on English news. As required, Times Now numbers were changed, while Republic TV is kept the same.” In the key updates, one line says “Republic is number 1 at all India level.”

But other than Times Now, there are other instances as well involving other channels. On August 8, 2019, Ramgarhia told Dasgupta, “Even this week we will wrapped up for News18. They are off for landing everywhere. However since we already control the ratings they might not see a fall. Specifically reach. (sic)”

On August 11, 2019, where the two discussed CNN-News18 landing pages. “They have taken multiple landing pages,” Ramgarhia said. “We are rationalising this.”

Dasgupta asked whether their reach was down, if the landing pages were gone. “Apparently some prasad guy is guiding these calls and he is saying give Landing to colors.(sic)” Colors is also a channel owned by Viacom18, the parent company of CNN-News18.

“Hmm, yes,” Ramgarhia responded. “They have given up News18 Hindi, cnnnews in favor of regional news, regional GEC and colors.(sic)”

Dasgupta remarked how English news won’t give them much money by being higher on the ranks, but Colors will give them disproportionately more. Ramgarhia said, “Yes. But we will control the news ones?”

“Till the board decides otherwise,” Dasgupta remarked.

Here, you can see the two officials selectively ‘controlling’ whether to allow the data of landing pages to be a part of the final ratings. The statement, “Till the board decides otherwise” is curious, but there is a convoluted background to that as well.

Landing pages

As we were doing the Arnab-Partho chats investigation, we got an email from Times Now titled Times Network Statement on bonafide usage of Landing Page. This letter is in defence of the usage of landing pages and it offers some perspective on what is happening here.

The statement read:

This referred to the same period of the forensic audit and the chat conversations. According to Times Now, since TRAI had already allowed the usage of landing pages, the effort to ‘control’ data by BARC was spurious.

As for the outlier policy, the letter stated:

This allegation is confirmed with an instance from the report which dates back to 2016. Here, Dasgupta asked Ramgarhia, “Can’t we get ABP to no 3? We are encouraging India News by being no 3. (sic)”

Ramgarhia pointed out, “In urban ABP is no 3.... The outlier removal is having a major impact on Abp. (sic)”

This one conversation here is the real clincher in this entire saga. It shows how “outlier data” was being used to control rankings of channels. The whole comment about ‘encouraging India news’ points to the fact that these rankings were used to promote narratives on certain channels by raising or downgrading their ranks.

Newslaundry reached out to BARC but they refused to comment on the issue. “As the matter is a subject of an ongoing investigation by the various law enforcement agencies, we are constrained to respond to your enquiries,” a spokesperson said.

Incentivised news narratives

“I ceased my association with BARC because of how toxic the environment had become,” the insider told us. “We were all aware that Dasgupta and Ramgarhia were indulging in such activities. I wouldn’t say these chats are surprising or shocking.”

There should have been solid logic applied to deal with outliers across the board, as is the norm. Instead, what we found here was that they cherry picked data and applied their own standards and narratives to justify their actions.

The insider pointed to this shady game of weekly manipulation. Think about it this way, by manipulating the data and consequently the rankings, these BARC officials were also indirectly affecting the flow of ad money to channels. Naturally, if a channel is ranked at three, the ones at four and below will try to do the same story that it is doing to raise their own rankings.

If channel at one is focusing on, say a suicide involving a celebrity and presenting the case as a murder, everyone further down the rankings will literally fall in line to do the same story, sensationalise it and get more ads. It’s the same logic used by companies who have fast consumer products, like shampoos or cold-drinks. The distributors will make sure the products are available in all general stores, even if two or more shops are right beside each other, to increase sales.

If these TV rankings were indeed manipulated, as evidenced here, then these men were actually deciding which stories will be covered by the rest of the media. They were controlling the flow of ad money, which was the incentive for channels to present certain kinds of “content”.

As we dig deeper into the Arnab-Partho chats, we are uncovering a very dirty system of ratings which was essentially being run like a mafia.

This is the sixth report in a series on the Whatsapp chats that purportedly show former BARC chief Partho Dasgupta's collusion with Republic TV head Arnab Goswami to manipulate TRPs, political lobbying and influence peddling, and more. Read the other reports in this series.